Ladder Anxiety is a term from the Starcraft community referring to distress caused by playing ranked/ladder games leading to the inclination to avoid them. This is usually just emotional, but can also include physical symptoms such as cold extremities, quick breathing, fatigue, etc. Though the advice given in response to this situation alternates between folksy relaxation techniques and the advice to "man up", it is still a common response and a response that is the opposite of what a game developer should want. Not only have I seen this discussion come up in the League of Legends community as well, suggesting that it's not something isolated to Starcraft, but I suspect the plague of general poor sportsmanship that infects online gaming may share a contributing factor with ladder anxiety -- the negative emotional response that manifests in some people as ladder anxiety could manifest in others as a desire to insult their opponents, make excluses, gloat, or otherwise behave in a "toxic" manner.
A feature in a game which motivates a person to avoid playing is an error of design. The common community response to this complaint makes the assumption that something is wrong with the player, not the game. I disagree. I think that, with some small changes that take into account human nature, game developers can create a much larger population of satisfied, competitive gamers. They'd have a better ladder system, too.
What is a rating system
Put simply, a rating system is a method to create a dynamic and ongoing hierarchy of a group of competitors. Based on previous performance, players are given a numerical rating which represents their skill level and is continiously updated after each match based on the result and the skill level of their opponent. The primary benefit this system has to a competitive online game is in matchmaking -- it's little fun for novices to play against experts or vice versa, and being able to play an evenly skilled opponent with the click of the button adds greatly to the value and longevity of a competitive game.
But this benefit is largely unappreciated by the mass of gamers who value rating systems solely for quantifying skill. Instead of being used to find appropriate opponents or to help the learning process, the rating functions like an appraisal of self worth. Players lose interest in having fun or learning the game and focus entirely on increasing their rating -- not their skill level, but their rating. It is exceedingly common for people to take rating drops that come as a result of misfortune (such as technical troubles) as a personal slight. This even extends past the realm of actual misfortune and into taking offense at perfectly valid behavior such as their opponent employing certain tactics they take to be "unfair". In truth, being rated below your skill level is little more than an inconvenience -- you'll be matched with easier opponents until your performance raises your rating to an appropriate level.
A skill rating is not a posterboard of gold stars and the goal of a competitive hobby is not to accumulate rating points like coins in Super Mario Brothers. Regardless of how precise the now famous "10,000 hour rule" is, acquiring skill in a difficult activity is a long, difficult road. The unfortunate fact is, an improving player in a truly accurate, honest rating system can expect his rating to increase at roughly the rate people actually improve: slow.
Primer on Probability
Even though it's perhaps a little too elementary to point this out to the people who would be interested in reading this kind of article, I'm going to do so anyway just to ensure everyone is on the same page. Let's take a coinflip as an example -- a random event between two equally likely outcomes (50/50). Even though there are only two outcomes and they should happen equally often, flipping a coin four times will not necessarily create a string of alternating results (heads, then tails, then heads, then tails). A person flipping a coin could easily see a result of heads all four times and from that experience mistakenly draw a conclusion that the coin is flawed or has a picture of a head on both sides. Despite the simplicity of a coinflip, it would take a large number of trials to be very confident of obtaining a result that gives an accurate reflection of the truth. Not only is human perception and memory is simply unable to handle that amount of data, our brains are designed for prediction. This makes humans really bad at making sense of a series of events where the outcome is subject to randomness. At least when they must rely on their subjective experience, of course. We fare far better when we rely on Excel.
Ratings are meant to reflect the probability of one player defeating another: two players with the same rating should win an equal number of times, while a higher rated player should win more often. More often, not always. Regardless, this means that losing regulary is expected. Not only are individual losses unavoidable, but streaks will be as well, bouncing your rating up and down around where it's meant to be. This is called variance. While chance will come close enough to evening out over the long run, a person who is experiencing these results in real time can't help but misinterpret them. A streak of victories is taken to represent a surge in ability; a streak of defeats is taken as an unfortunate injustice or personal failure.
Neither is accurate. If you were to wake up one day better, marginally, than your current rating, your performance would not result in win after win in a direct path towards your new, true rating. Instead, you would win slightly more often until your rating reflected your newfound skill.
Loss Aversion
People are irrational. This shouldn't be news to anyone. One of these universal irrational tendencies is a psychological principle called loss aversion. Put simply, humans put more value in avoiding losing something than they would on acquiring that something to begin with. For example, a typical person will be more upset by losing $20 than they would be happy upon finding $20. This preference has an effect on our behavior and is so powerful it can lead to decisionmaking which is quite ludicrous when analyzed objectively.
What does this have to do with rating systems?
Let's look at the player experience when playing a rated game in either Starcraft II or League of Legends. While both use different systems, each give players a rating which increases after each win and decreases after each loss. Not only is the rating adjusted but for the purpose of clarity, the difference is displayed quite prominently to each player alongside their new rating in the statistics screen that follows each match. This feedback is intended to be motivational: the joy of gaining points synergizes with the natural joy of winning, increasing one's desire to play. Upon losing, however, the player has the exact opposite response. Because of loss aversion, these two possibilities do not cancel each other out -- losing is the more significant factor.
In an accurate rating system, hitting the "Play" button matches you up with a person who is approximately your skill level. This means someone against whom your chances of winning should be as close to 50-50 as possible. If you win, you will be rewarded with points. If you lose, you will be punished and have points taken away from you. Even though you will win and lose about the same number of points in the long run (assuming your skill remains static), psychologically you put more value on the points you currently have than any points you may win. Losing ten points hurts more, a lot more in fact, than winning ten points feels good. From an emotional perspective, this is a losing proposition. Unless your victories will substantially outnumber your losses (which should not happen except for the few very best players), participation is a recipe for misery. If you're in the business of cooking up fun, this isn't something you want in the oven.
Ironically, instant feedback is misleading
The Elo rating system (the most popular and widespread -- though always with adaptations) was implemented by the USCF in 1960 as an improvement over a previous rating system already being used. This was long before personal computers were everywhere, and one of the strengths of Elo is that it relied on relatively simple arithmetic. This was important since rating adjustments were done by an actual person (perhaps with a calculator) thumbing through sheets of chess results after a tournament. Logistically, this meant that rating adjustments were neither instantaneous nor were they for individual matches. Waiting for the results to be processed blunts the negative response since the material loss, the decrease in points, is not immediately attached to the emotional impact of losing. Handling the entire tournament in one adjustment meant that the change in rating reflected not a single game but the handful played over the entire tournament. Processing the games in batches makes for less variance in each adjustment -- ratings will not bounce all over the place, and each change carries more significance.
This is where irony comes in. Many people would assume that the more immediate and clear the feedback, the more reliable and helpful that feedback will be. With ratings, this isn't the case. Showing the adjustments of each individual game is a lie. Once your rating has been established in an accurate system and you are paired against similar opponents, you are meant to lose from time to time. Often, in fact! Around as much as you win! The outcome of one or two games carries little, if any, statistical significance. However, this is lost on someone without a solid understanding of both the rating system and statistics, which is almost everyone. All they see is a mean robot taking away gold stars everytime they lose a game.
What's the solution?
Simple: batch adjustments. Not only would ratings (once established) work better changing on, say, a weekly basis, rather than game to game, it would also be best to update less often for players who do not play many games.
This idea is actually built into the Glicko rating system and called the "rating period". Any matches that take place during one rating period are to be considered simultaneous, so changes are made based on chunks of games rather than one by one. The official paper outlining the Glicko system recommends a rating period which has, on average, 10-15 games per player.
While removing the instant adjustments is an improvement in a rating implementation, it isn't necessarily an improvement in overall game design. Instantaneous feedback is an important part of game design. The problem is that losses are not just unavoidable -- a good matchmaking system guarantees that they will happen about half the time. Providing negative reinforcement (the loss of points) for an unpreventable event is akin to alternating randomly between smacking a puppy on the nose and giving it a treat -- the natural response is fear and anxiety. The idea that a player just needs to perform better to obtain his reward (and avoid punishment) isn't valid, since better performance will only result in more difficult challenges.
Instead, developers need to find other ways to integrate feedback and rewards into the competitive gaming process that bear in mind the reality of participating in a competitive hobby. Even for the best of us, losses happen and are obviously a suboptimal result. No one likes to lose, even if they've grown to accept it. Rubbing it in by taking points away does not accomplish anything. The satisfaction derived from this kind of hobby comes from ongoing self improvement and testing yourself, and the design should encourage this aspect of the experience.
I previously blogged about MLG's unpopular rule Extended Series. Rather than simply criticize, my goal with the post was to explore why MLG might have put ES into practice (it's not common), why they would stubbornly stick with it despite consistent community outcry, and the idea that the concepts behind a quality tournament structure are not universal to all games. This got lost in the haze of my fuzzy ideas and the debate of my impressions of other esport games. For this and also because actually writing the article (as writing often does) helped evolve and clarify my views on the subject (i.e. even though I wrote the article it's still on my mind), I thought I would revisit the topic.
I will never be in favor of structures or rules that give one player in a match a dramatic advantage over his opponent.
This is regardless of whether or not the advantage is warranted based on the overall structure or doled out at random. The most common example is a standard double elimination bracket, which forces one player in the grand finals to win twice. Some might dispute my premise that ES constitutes an advantage since the match is simply "continued" from an earlier meeting. Since the rematch is rare rather than inevitable, I don't consider this a valid defense. "Continuing" the series from where the previous match left off is no different, to me, than incorporating results from any previous encounter between the two.
This is not to say I am against any advantage (such as seeding or earned byes) in any situation. However, I do not like the idea of an isolated match between two players beginning on an uneven playing field. If one player has truly performed so much better than his prospective opponent that to not award him a large advantage would be unfair, then they shouldn't even be playing.
From the perspective of a spectator, it's not suspenseful to watch the player who is already an underdog futilely attempt to make a near-impossible come back. It's exciting and memorable when it actually happens, but won't happen often enough to stick around to see many of those exciting and memorable moments.
For me, though, it's more than that. It's something I have difficulty putting into words -- a matter of taste, if you will. Other forms of advantages may change who a player faces and when or skip ahead a stage in a tournament, but these advantages do not disrupt any individual match between two players, so I do not find these rules to be disagreeable. To me, each match has an identity outside the tournament of which it is a part. A single bout between two entities. An advantage for one is like penalizing the other, and the match loses significance. A bye, on the other hand, doesn't feel so personal. Even though I'm well aware of the value a bye can have (as far as increasing a players' probability of winning), the penalty is shared amongst all the players who don't receive byes and the integrity of individual match results are preserved.
Ultimately, I guess the idea doesn't mesh with my idea of sportsmanship or sportsmanlike conduct -- a true warrior should prefer a clean victory that began on even footing and welcome the opportunity to prove himself. Clamoring for an advantage no matter how deserved strikes me as... well, petty.
I do not agree that ES is illogical from the perspective of tournament accuracy.
Any tournament is basically a tool to estimate the relative skills of a group of players. In elimination style tournaments such as double elimination, each batch of eliminations constitutes a new, more specific guess -- the players left standing are estimated to be better than the players that are eliminated. The tournament will continue to refine its estimations of the players still in contention, but once a player is eliminated, the decision is final.
The losers bracket in double elimination adds a third category: players who are considered to be worse than those in the winners bracket, but better than those who were eliminated. It also recognizes that, unlike eliminated players, the tournament is still not entirely sure where these players stand. The structure thinks that they are worse and will be eliminated soon, but it has an open mind.
When two players rematch in the losers bracket, a traditional double elimination bracket has estimated these players to be in parallel. Even though there is already a result between the two players, the structure does not take that into account -- the subsequent performance of the two players since their match is what matters and that subsequent performance has put them in the same group. Having a rematch does not bother the tournament, since if the loser of the previous match wins the new match, it suggests that the players are more or less equal in skill. If two players are roughly equal, then it doesn't matter which one of them continues since it will make little difference in the average skill of players in the next round. Should the previous winner win once again, the tournament feels no sympathy for the loser since it kind of figured he was on his way out anyway. After all, the guy that beat him in the first place has lost once by now.
The tournament has the same impression of every player grouped together in a round and losing to one is no different than losing to another. The tournament doesn't worry about the chance that a better player might lose, because its entire method for estimation is built on the premise that the better player will win. Maybe a few outliers slip through the cracks and get a drastically incorrect ranking? "No big deal," says the tournament, "I'm just guessing, bro". Algorithms are laid back dudes, they don't get stressed out by the occasional boo boo.
Simply from the perspective of tournament accuracy, adding ES into the mix is perfectly reasonable. The players face off because the winning player is assumed to be better -- if there's already a result between them, why not take it into account? In fact, it would be also reasonable to not even play the rematch and just carry over the results from the first meeting, or to create a super extended series based on all results between the two players over the last month. The tournament just wants to know which of the two players is better so it can refine its estimations. Sure, this sucks for the player with the short end of the stick. "Too bad", says the tournament. "You hate, I estimate'"
If the question is "which of these two players is better?", previous matches are perfectly relevant. They certainly aren't random or arbitrary. You could criticize previous results for not being current. Even over a single weekend tournament, the second match could reflect factors such as fatigue. You could also criticize a single match for not being statistically significant. Ultimately, when it comes to the math ES as applied by MLG doesn't make much of a difference either way. Arguing about it is a waste of breath and only serves to cloud much more important criticisms. If you think it makes the tournament less enjoyable to watch or play, well then I'll just have to go ahead and have to agree with you. Deal with it.
I will only insist that ES is truly unjust in one particular situation.
This situation occurs when a player is eliminated after only losing (twice) to a single opponent and there were other viable opponents in the round of their rematch. In MLG Summer Arena, Oz and Stephano faced off in an extended series at the end of Bracket 1. There were no alternative opponents for them to face, so while it's not my preference, I don't consider ES in this instance to be unjust. It's true that, if instead of what happened, Oz had beaten Alicia, then Stephano would be in a better position for that round (starting off 0-0 rather than 2-1), but this isn't unreasonable or illogical. If Oz beats Stephano but loses to Alicia, it is estimated by the structure that Alicia and Oz are both better than Stephano. If Oz beats both players, then the structure has no basis to consider either Alicia or Stephano to be better than the other -- they have the same number of losses (1) and to the same player (Oz). Therefore, Alicia and Stephano start off even. When Stephano and Oz met again, the tournament has a reason to think Oz is better, which is reflected in the handicap. To change the tournament's mind, Stephano must be convincing.
Similarly, I don't feel the ES rematch between Oz and Alicia was unjust either. Everyone else in their bracket has been eliminated, and ES is used in place of the traditional win twice system. While it's uncomfortable to choose between two things one doesn't like, I actually prefer this to the common system. While superficial, it seems so silly when, after several games, the score resets to 0-0.
In Bracket 2 on the other hand, there was a rematch between aLive and Daisy. At this point of the bracket, GanZi and First were also viable opponents. Forcing Daisy to play at aLive from a disadvantage rather than playing GanZi or First is unjust. The tournament structure typically only eliminates players after they lose to two different opponents, except when they advance far enough that a second unique, parallel opponent is not available. If Daisy loses twice to aLive, all the tournament really knows is that he is worse than aLive -- this isn't quite as accurate as when the two losses are to two different players. If a rematch occurs, the only way to avoid this problem is if the results are different the second time around. Extended series exacerbates this problem (just a little) since it makes a repeated result more likely.
The same thing happened to Dream at the previous MLG Arena.
ES in MLG's Championship format is a little more complicated.
MLG uses a strange hybrid of pool play and elimination structures for their championship events. The tournament is only double elimination to players from the open bracket who lose their first game prior to qualifying for the pool. For these players, the previous point is applicable.
Pool players are in a different boat. Finishing first place in the group will grant a player two "lives" in the elimination stage of the tournament along with the other three group winners. Whether or not the player went undefeated to win his group is not relevant. All other pool players (along with the few remaining players from the open bracket) are given one "life" in the elimination stage regardless of how many losses they accrued during the group stage -- the rewards for fewer losses are metered out with byes rather than points toward elimination.
This massive advantage given to players in the group stage amid a needlessly and ridiculously complicated format makes it hard to sympathize when these situations arise since, after all, being in the pool is still a big advantage over the majority of players. In the 2012 Winter Championship, for example, Socke had a rematch against HuK starting down 0-2 when there were several other viable opponents. But, Socke had already lost four matches (including the one to HuK) so far in the tournament.
Of course, it will still be better if rematches were avoided when there are available alternatives. But I hesitate to call the above situation (or ones like it) truly unjust.
As needed randomized pairing is a simple, ad-hoc solution
In an elimination tournament, it is standard to draw up the initial brackets (whether at random, with seeding, drafting, or a combination) and then let them play out. Spectators and players can plan out and predict what might lie ahead. However, this bracket isn't mandatory. It's not written in stone. When an unnecessary rematch happens in any double elimination tournament (ES or not), it's trivial to simply regenerate the pairings at random until all rematches are avoided.
As discussed earlier, every player in parallel is considered equal as far as an elimination bracket is concerned. Remaking the pairings in the event of a rematch is not somehow unfair or biased. In fact, regenerating the pairings for unique match-ups would actually increase the accuracy of the tournament's results ever so slightly (ES or not).
Let's use the Daisy example from above. Instead of Daisy vs. aLive (the rematch) and GanZi vs. First, First vs. Daisy/aLive vs. GanZi and GanZi vs. Daisy/aLive vs. First are both unique pairings. The tournament administrator could choose one of the two at random to use instead. Not only does this avoid victimizing Daisy, it's more entertaining to watch. It's a win-win-win.
Previously I wrote a blog post about TSL's RevivaL and the circumstances surrounding his elimination from Code A. In that post, I discussed the GSL rule which dictates that the four players who finish last in their groups during the Code S RO32 are paired against four newcomers from the preliminaries in the RO48 of Code A. All players who lose in the RO48 are eliminated from GSL entirely and must once again pass through the preliminaries. Unfortunately, RevivaL is not the only player forced to start over from scratch after an unlucky trek in Code A.
The current structure was introduced at the beginning of the November 2011 season. After the initial group stage of Code S, the eight players finishing last in their groups are placed into the first round of Code A along with forty other players. They are each paired against players from the preliminaries rather than each other or players from the Up & Down matches. The likely goal was to dismiss stagnant Code S players who were no longer top tier players, but had been able to tread water in the generous structure previously used. In this respect, the rule functions well -- a player who is swept in his group should not have an easy path to make the next season's Code S, and a subsequent loss to an unknown is reasonable grounds for dismissal from Code A.
On the other hand, over half the players coming through the preliminary will not be facing Code S players. The few unfortunate enough to do so shoulder an incredible burden in comparison. Though it may be reasonable from the perspective of the Code S player, it's unreasonable that a Code A novice is expected to beat someone who simply had a bad day in Code S to earn a berth in next season's Code A.
Since this format has been put in place, there have been four seasons of Code A. Of those 30 matches, the Code S players have won 25 (two results were excluded as they involved players who received direct seeds into GSL). Here is a list.
NOVEMBER 1W 7L 13% YuGiOh(S) over MinSeOk asd(S) over GuineaPig Clide(S) over Creator Bomber(S) over RainBOw NaDa(S) over Ven SuperNoVa(S) over VINES Curious(S) over TheStC Cezanne over Ensnare(S) SEASON1* 2W 4L 33% Fin(S) over cOre YuGiOh(S) over Sculp Brown(S) over Lure JYP(S) over Life Avenge over IdrA(S) Heart over Bomber(S) SEASON2 2W 6L 25% DongRaeGu(S) over RevivaL BBoong(S) over GGanDoL Puzzle(S) over True Polt(S) over Sound GuMiho(S) over dreamertt aLive(S) over SocceR Lure over InCa(S) GhostKing over NaDa(S) SEASON3 0W 8L 0% HerO(S) over JookTo TheStC(S) over Punisher GuMiho(S) over Ready SuHoSin(S) over Crank Sniper(S) over AnNyeong Oz(S) over JKS Leenock(S) over TAiLS YuGiOh(S) over Harrier
Code S washouts are still some pretty stiff competition. Losing a single match to one of these players does not signify a player is not Code A material. In order to further illustrate my point, I collected results from the first round matches that pair up players from the preliminaries against players who entered Code A through the Up & Down matches (in November there were also Code A holdovers from the previous season -- in the current structure, any player that isn't demoted from Code A is eligible for Up & Down matches). I excluded matches where the player earned his Code A spot outside of the preliminaries (wildcards/other tournaments).
vs. UP&DOWN/CODE A November 8W 5L 62% Season 1 8W 5L 62% Season 2 5W 7L 42% Season 3 3W 9L 25% Overall 24W 26L 48% vs. CODE S November 1W 7L 13% Season 1 2W 4L 33% Season 2 2W 6L 25% Season 3 0W 8L 00% Overall 5W 25L 17%
This is a dramatic difference. Though a few dozen games is hardly enough data to draw concrete findings, it does suggest that it's a strong disadvantage to be paired against the Code S player. Since this curse is randomly cast on four of the twenty or so qualifying players each season, it's reasonable to say that it's not a fair rule. Considering how difficult it is to simply pass through the Code A preliminaries, being unjustly demoted from Code A is a bad beat indeed.
TSL Zerg player RevivaL is one of the "faceless Koreans" playing this weekend at MLG Summer Arena that hasn't gotten much screen time thus far on the premier tournament circuit, either in South Korea or abroad. We won't see him this season in GSL or in the upcoming inaugural OSL.
It's easy to assume that we haven't seen him because he isn't any good (at least not yet), but analyzing tournament results in detail can tell you an interesting story.
The bad beat in question is RevivaL's last GSL appearance, in the previous season of Code A. GSL's structure change late last year did a great job at addressing stagnancy in Code S and Code A, making it easier for a strong player to move up to Code S and doing a better job at quickly relegating Code S players who were no longer top tier material. However, in accomplishing this they've increased the burden of difficulty on players coming from the preliminaries. One unfortunate quirk specifically is relevant in this situation. Here is the explanation quoted from Liquipedia:
First round (Ro48) consists of 40 qualified players and the 8 4th place finishers in the Code S Ro32 group stage.
Losers in the first round lose their Code A status and must requalify for next season.
The second round (Ro32) consists of the 24 advancing players from the first round and the 8 3rd place finishers in the Code S Ro32 group stage.
Losers in the second round get a chance to acquire Code S by competing in the Up & Down Matches.
What this means is that eight unlucky new Code A players will be matched up against eight Code S players from that season, fresh off their losses in the initial group stage. The luckier ones will face off against those relegated in the Up & Down matches from the previous season. The first round of Code A is pivotal -- win the game, and get two chances (the remainder of your Code A bracket and then, if you lose, your Up & Down group) to move up to Code S and a guaranteed slot in Code A next year. Lose the game and be forced to go through the preliminaries another time.
I know what you're thinking. "So what? The Code S players are the ones who took 4th in their groups -- they're the worst of the bunch. No big deal." Guess who RevivaL was paired against? DongRaeGu.
That's right. Though he was the champion of the previous season, DongRaeGu lost to TheStC and TaeJa to take fourth in his group and earn the chance to ruin someone's month. To add insult to injury, the games must have been very frustrating for RevivaL as DongRaeGu won the first game with a delayed one base speedling all-in and the second with a blind 10 Pool followed up with an offensive Spine Crawler. DRG would go on to win every series he played in GSL through his Code A bracket and the next season of Code S until the final four of the following season three months later.
This wasn't the only time RevivaL's been within an arm's reach of making the next level. Last summer, he lost in the semifinals of the NASL S1 qualifiers; at the end of the year, he took second after Sound in the Korean HomeStory Cup IV qualifiers. Both times he was just a single match away from the trip abroad.
I was pretty happy to see RevivaL on the list for Summer Arena because his game against DRG broke my heart when I first watched it. Having lost in the most recent preliminaries for both OSL (Group J) and GSL (Group 4), he should have had plenty of energy to focus on this event. He also is a teammate of Symbol, fellow Zerg who took the silver medal at the previous Arena. Recently, TSL called him in during their match against NSHoSeo immediately after Symbol's first loss. RevivaL was able to take two wins and close out the series 5-3. Hopefully he can perform just as well this weekend at MLG!
MLG announced a couple days ago that they would be implementing a new structure for their Summer Arena. Here's the reddit thread.
Thus far, MLG has done three arenas. Winter Arena and Spring Arena 2 were nearly identical. Both used 32 players, both used the same, standard double elimination bracket (well, almost standard -- it wouldn't be an MLG without extended series). The only difference was how the first round pairings were set: Winter Arena went directly by seeding while Spring Arena 2 let the top half seeds draft their opponents from the bottom half. Spring Arena 1 used a much different format and only had eight players, however it seems that MLG will not continue hosting this sort of event.
Summer Arena adds another subtle change. Rather than using a standard double elimination bracket, the players will first be separated into two groups. From there, double elimination play will yield two finalists. Those two players will face off in a single championship match.
While far from ideal, I think this is a wonderful change for MLG. A key problem with using double elimination brackets is the monumental advantage held by players who have not yet lost. As the tournament should be growing more exciting near the conclusion of the event, matches taking place in the losers bracket are between players highly unlikely to ultimately win the tournament. The finals, which should be the climax, often seems predetermined since not only does one player have a large handicap, oftentimes the match is a repeat of a pairing from one or two rounds prior.
In other words, tournament does not end with a "bang". It fizzles. The winners bracket finals and semifinals are often the most significant matches. While using double elimination affords the losers of these bouts a second chance they wouldn't get in a typical knockout tournament, the losers bracket tends to play out as though they were glorified consolation matches.
Winning from the losers bracket, of course, is not impossible. But "not impossible" doesn't make an event exciting to watch. The fact is, the winners bracket player will win an awful lot, and enough tournaments have happened on the Starcraft 2 circuit for this to become apparent. For me, it actually feels worse when the handicapped player is able to put up a strong fight, but still loses (see Squirtle vs. aLive at IPL4) -- it breaks my heart to know that they could have won if the match started out on even footing.
Considering this, MLG's structure change is a good move. While last few matches before the championship will suffer the same fate as before, the integrity of the finals is preserved since two players (who will have yet to play each other) will meet on an even playing field. No advantages, no obvious result. This should make watching the finals much more entertaining.
If you watched the first day of Red Bull Battlegrounds, you'll remember a third stream was added on short notice towards the end of the first day in order for all the group stage games to be aired. Games went longer than expected, and the decision was made to start playing more games in tandem than the original plan called for and broadcast the replays after they finished.
Some might call this unavoidable. Starcraft does not have a pre-set time limit, and games can end in just a few minutes or go as long as an hour. When the tournament calls for each match to be a best-of series then not only is the game duration variable but the number of games played is as well. This makes scheduling and organizing a tournament an unpredictable nightmare. This isn't good for anyone. Fans don't know when to tune in to see their favorite player, players in the event aren't sure when their next match will be, and tournament administrators only have so much hair to pull out from stress. It's easy to chalk up a debacle like this to bad luck. But is it just bad luck, or does a bad format share some of the blame?
The Red Bull Battlegrounds tournament used a structure in the first phase of the tournament which was simply wasteful. The idea of splitting players into four man groups and playing a round robin with the top two players advancing has a history with Starcraft dating back to the OSL. However it has been mindlessly misapplied in the western SC2 tournament scene. Four man round robin is almost completely inferior to a simpler, quicker, common alternative. This is something GSL realized over a year ago when they replaced the system's use in Code S's group stage.
GSL switched their group stage to a double elimination bracket or short swiss style tournament. Two pairs of players face off, creating two winners (1-0) and two losers (0-1). The two winners play to decide who advances (2-0) and who plays the consolation match (1-1), while the two losers play to decide who else enters the consolation match (1-1) and who is eliminated in fourth place (0-2). The winner of the consolation match advances in second place (2-1), while the loser is eliminated in third (1-2). Five matches are played in this method, one less than the six required to settle a round robin with the same number of players.
Some in favor of round robin argue that the extra match offers additional games to broadcast for the fans, but this does not strike me as a benefit. We live in a time when there is an overabundance of Starcraft content. There are multiple major organizations putting out content multiple times weekly, along with live streams, minor tournaments and showmatches each day, larger tournaments each weekend, talk shows, and so forth. There is simply more Starcraft related media than any one person could reasonably expect to consume.
It stands to reason that competition organizers should not to try and squeeze every possible minute of Starcraft footage out of each tournament but instead to make every minute as relevant and exciting as possible. Since these round robin groups take place at the earlier stages of the tournament, no tension has been built and the prize is far off in the distance. Saving time by settling the earlier rounds with less games affords the organizer the possibility to increase the number of games later in the tournament when there is more at stake. The DreamHack Summer Open had 56 groups play out before entering the RO16. Playing five series instead of six would skim 56 BO3 series off the schedule. The time saved would allow DreamHack to host longer series during the climactic stages of the tournament instead of using BO3 all the way until the finals, with plenty of time to spare!
A proponent of this group play system would claim that I'm in favor of sacrificing the integrity of the tournament in order to save time, but this isn't the case. In this particular application of round robin, losing two games guarantees elimination before the third game is even played most of the time. To advance after two losses one must not only win his third game, but also hope for the right sequence of results to take place among the other players in the group to create a three way tie. Oftentimes it's clear this isn't even possible before the third match has even begun. Furthermore, since most tournaments ignorantly use criteria to break ties rather than extra games (see another article, Break Ties, Not Hearts), advancing after two losses is even less likely to be possible before the last match is played, let alone actually happen. This suggests that the double elimination method saves time with little loss in accuracy.
In a round robin group the meaning behind the matches is vague. It's clear that losses are bad and wins are good, but precisely which matches or games are the most significant are usually not apparent until the moment has already past. There is little excitement or suspense if the match seems to lack a clear and direct consequence. In the double elimination group, the first loss is recognized as a setback, but not a catastrophe. Two things are made clear: the next game or two games must be won, and neither will take place against the best player in the group.
In a round-robin group, this isn't necessarily the case. Imagine a group where two players are closely matched (B1,B2), another is better and will always win (A) and one is worse and will always lose (C). Regardless of which format, we can expect that A will advance and C will be eliminated. In a round-robin format, B1 and B2 paired in their first or second match would obscure (to each other and the viewers) the fact that the match between them is the key to continuing in the tournament. In the double-elimination format, these two players will always meet in the consolation match to decide second place, regardless of the pairings. Better still, being paired together in the first match will mean the consolation match takes place after they've already played each other and can better adapt to one another. Considering this, one could argue that double elimination is more accurate than round robin when it comes to determining the rightful second and third places.
Double elimination groups will never creates a situation where a meaningless or lopsided match can be played. Every match takes place between two players on equal footing with the same at stake. Round-robin groups can create circumstances where one player will advance and his opponent will be eliminated, regardless of the outcome of their match. Neither player has much incentive to perform at their best and make a game worth watching. Ethics and obligation could motivate the players to go out and play anyway, but why even bother? And who could expect a forlorn player, his hopes demolished only moments earlier with the knowledge that his tournament has ended, to go out and play well even if he wanted to? An even worse circumstance is when the eliminated player is matched up against a player whose chances for advancing is still up in the air -- this gives the player in contention a clear advantage that the other players in the group did not have.
These lopsided scenarios not only degrade the accuracy of a result from a round-robin group, but also provide the opportunity for nefarious players to collude. If in a match between two players there is an outcome where one gains much and the other loses nothing, the players can guarantee that outcome and fix the match. While there is no format that can prevent players from match fixing if they so choose, a round-robin format actually incentivizes it by creating opportunities for a player to enter a match knowing that he can take a dive and still be guaranteed to continue. This is in contrast to other formats, where choosing to lose a match would involve sacrifice. Perhaps even a greater problem than the potential for cheating is that creating these opportunities could cast suspicion on any player involved whether or not they engaged in foul play as long as motive could be argued. Since proving whether or not a match was thrown is virtually impossible, not only is there no way to catch prospective cheaters but also no way to clear an innocent man's name.
OGN's use of four man round robin used single games, not BO3 series, to settle matches. The likely intent behind this structure was that playing single games against multiple opponents was an alternative to a multiple game series against a single opponent. Furthermore, the fact that matches were individual games and took place amongst a highly skilled pool of players greatly increases the likelihood of a tie occuring, in which case OSL always held sudden death matches. While I would still prefer double elimination (so did MBCGame, RIP), OGN's application has merit considering its place in a league spanning several months rather than a single weekend and that a single group would play out over several weeks rather than a couple hours.
The method like the one used at DreamHack and Red Bull Battlegrounds is unnecessarily excessive to the point of madness. Whether it is stemming from a misguided attempt for an accurate tournament, a desire to offer players more competition time, or that simply the longer a stream stays live the greater the number of total views, in the end it just creates a poorer product.
Is four man group play with double elimination indisputably the optimal format? Certainly not. However, it would be trivial for any tournament used to or partial to four man groups to choose the double elimination method instead of round-robin. Doing so would do wonders for reducing the length and increasing the quality of the early stages of a competition.
While I suggested that three way ties in four man groups were relatively uncommon in a previous article on the subject, this past DreamHack was a full on tie extravaganza. This created a handful of amusing examples of injustice. It should stand to the organizers involved in choosing and applying a structure to simulate the possibilities and make design decisions that avoid these kind of mishaps. Absent that, one would at least hope that when these blunders actually occur people would recognize that something is wrong and that things should be done differently.
HuK finished 3-0 (6-1) in his first group, 3-0 (6-1) in his second group, and 2-1 (4-4) in his third group. Due to a three way tie for first place with Stephano and sLivko, he was eliminated before the RO16 based on his map score despite losing only one series. He lost to Stephano, who finished fourth.
On the flip side, nine of the players in the final sixteen had all lost two or three matches on their road through the group stages.
Because of tiebreakers, three other players were also eliminated after a single series loss, but in the second group stage rather than the third. Amusingly, one of them was HuK's teammate, JYP, despite winning his series against silver medalist DIMAGA. The other two were DeathAngel and ToD, though they didn't face off against anyone in the playoffs.
LiveZerg (out in the third group stage) and qxc (out in the second) each took a series off two players (NaNiwa and MaNa, Harstem and MaNa respectively) who would make the RO16 without actually getting there themselves. HerO and sLivko were the only ones in the playoffs who would accomplish the same feat.
Spitting in the face of extended series, DreamHack matched up Vampire and Cytoplasm in the same group in the first and the third stage. Vampire won both series 2-0 without any help from MLG, but still managed to get himself eliminated.
The third group stage had one three way tie for first in Group G (2-1, 2-1, 2-1, 0-3) and one three way tie for second in Group A (3-0, 1-2, 1-2, 1-2). Would it have been a better idea to just take away a spot from Group A so that three players could advance from Group G?
The eventual winner, MaNa, lost three series throughout his group stage games. He and Harstem were the only players in the RO16 with that weak a record. Furthermore, in the second group stage MaNa was in a three way tie for second place and a hair away from being quietly eliminated in the RO64. Fortunately, map score gave him the edge over qxc and merz.
Another thing to note about DreamHack was the dearth of Korean presence at the end of the tournament. From what I can tell, this is DreamHack's step by step plan for western domination of a Starcraft tournament.
Wait for a period showing the most lopsided performance for one race in the history of the game (Zerg).
Only let Koreans participate if they play the other two races (Protoss: HerO, JYP, Daisy, Vampire/Terran: Dragon, PuMa, Keen, TaeJa).
Use the rules against them (JYP was eliminated on a tiebreaker).
Sic as many of the rest on each other as possible (PuMa helped eliminate Daisy and Vampire in group stages, HerO eliminated TaeJa in the RO16).
Count on EG to live up to its cursed reputation (PuMa and JYP being eliminated prematurely after losing to foreigners).
Bring out the big guns (NaNiwa vs. HerO, NaNiwa vs. Keen, SaSe vs. TaeJa, Stephano vs. Keen, HuK vs. Vampire).
If there are any left, cross your fingers and hope for the best.
Normally in these tournaments we've all assumed the trophy is destined for the orient and measure foreigner success on Korean scalps as much or more than actual prize money. Even though this DreamHack had a strong European showing, it's important to realize that not only was the field relatively soft as far as South Korean talent is concerned, foreigner vs. Korean results were not particularly impressive or different from other recent tournaments. The stars managed to align just right to create some rather unexpected placings.
Here are the results, links to VODs, etc. at Liquipedia.
Break Ties, Not Hearts: Why the Only Tiebreaker to Use in a Tournament is More Games
You can also read this on my TL Blog
In my opinion, tiebreakers are tragic. There's little I find more appalling than eliminating a perfectly capable player on the spot just to save some time. In this article I go over a handful of examples of the different sorts of ties that have sprung up in tournaments over the past year, discuss interpretations of the rules, and try to make a case to ditch this practice completely and just use extra games instead.
HomeStory Cup III was a milestone tournament in the ongoing "David and Goliath struggle" between western players and South Korea when HuK and Naniwa took first and second with wins over two time GSL Champion MC just a few days after HuK had captured the DreamHack crown in a field that included Bomber, July, MC, and Moon. It was the first glimmer of hope that the "Korean invasion" may not shut western players out of their own tournaments forever. In this respect, HomeStory Cup IV in January of 2012 stood in stark contrast. With several more representatives in attendance this time around, South Korean dominance was on full display. Korean players took six of final eight slots and nearly shut the westerners out in matches for the entire weekend. The only exceptions: Polish stalwart Nerchio winning his group with consecutive victories over HerO [2-1] and viOLet [2-0], and Delphi beating MarineKing [2-1] in one of the opening matches of the tournament.
Wait, Delphi who?
My point exactly. Delphi is a German Zerg, active in the online scene in Europe but yet (then, and still) to make his mark internationally.[1] This competition had a global player pool and audience but took place in Germany, which means a big chance for a local player like Delphi who wouldn't have the chance to compete on a major stage month in and month out. A strong showing could be the kind of breakout performance that gets him invitations to future tournaments, access to sponsorships, and opportunities to compete around Europe or in North America. Unfortunately, he drew the only first round group with two Koreans, pitting him against MarineKing and viOLet along with compatriot HasuObs. Despite these odds, Delphi managed to win his match against MarineKing as well as his match against HasuObs. This outstanding performance was still insufficient, and Delphi was eliminated. Why? Look at how the group finished:
Group B MRO GSO 1. MarineKing 2-1 [5-3] 2. viOLet 2-1 [4-2] 3. Delphi 2-1 [4-3] 4. HasuObs 0-3 [0-6]
After defeating Delphi, viOLet lost his series against MarineKing, tying all three players. Since each match was a BO3, HomeStory Cup broke ties by scoring each individual game. Delphi's win over MarineKing went to a third map, while both MarineKing and viOLet swept the two series they each won. MarineKing advanced as first in the group thanks to the extra game win, viOLet in second without having an extra loss, and Delphi went home.
Realistically, Delphi's chances in a sudden death would not have been great. I doubt anyone would argue that point. He had much less experience under pressure, and no longer had the advantage of being unknown or underestimated. But he earned his chance to prove us all wrong -- that, no one can deny. Win or lose, he deserved to have a couple extra games penciled into the broadcast schedule, the asterisk on the results table, and the extra attention. Instead, his tournament was over. It ended with him on the sidelines, watching himself be eliminated in a match between two other people.[2]
Most tournaments in the international Starcraft scene rely on a round robin system in some capacity. Any format has its pros and cons, and one downside to round robin is the potential for ties. In an elimination tournament, there are no ties -- two players face off, one advances, no room for disputes. When results of multiple games against multiple opponents are taken into account, however, two or more players may finish with the same score. This becomes more likely as group sizes increase and player skill converges. Since round robin is frequently used in groups of four and always used in the earlier stage of a tournament when the skill levels are most variable, ties do not occur often. When they do occur, few notice -- less people watch the beginnings of a tournament and those that do watch the best players, the ones least likely to tie. Because of this, tie resolutions have stayed under the radar and the use of rules which award advancement to one player or another by default has quietly become standard without public outcry from players or fans. Despite no official collaboration on rules across the various tournament organizations, not a single major tournament has taken a stand against this practice and used playoff matches as a first resort.[3] Instead, tournaments rely on one, some, or all of the following criteria:
MRO (match results overall): match by match performance against other players in the group.
MRH (match results head to head): your performance against another specific player or subset of players.
GSO (game score or map score overall): since individual matches are generally comprised of a series of games, this is your game by game performance against all other players in the group.
GSH (game score head to head): your game by game performance against another specific player or subset of players.
MRO is the standard choice to determine ranking. When MRO yields a tie, tournaments will use one, two, or all three of the remaining factors in their preferred order of priority to decide which of the tied players finish in which spots. While less common, it's also possible to determine the initial ranking by GSO instead.
One can argue merits of all approaches -- putting priority on head to head criteria (MRH/GSH) means an advancing player earned his spot with a victory over another, while putting priority on overall results (MRO/GSO) means an advancing player earned his spot by demonstrating a good performance over multiple trials. Giving weight to match results (MRO/MRH) puts significance in winning matches regardless of the details, while looking to game score (GSO/GSH) makes every single map in a series count. In theory, this seems like a simple difference of opinion with little relevance on the quality of a tournament or the results of the players. In practice, this isn't the whole story. The competitors in these tournaments are real people, players struggling to get noticed or stay relevant in a scene that is both fiercely competitive and completely unrewarding of all but the crème de la crème. These players work hard to hone their abilities and deserve better than being told "sorry, not quite good enough" -- especially, but not just, when prize money hangs in the balance.
The most common round robin implementation, as mentioned above, uses groups of four with two players advancing. If two players advance and two players are eliminated, there are three different ways a tie can occur and one is not likely to cause controversy.[4] As an example, here is a group from the third round of DreamHack 2012 Stockholm Open:
Group E MRO GSO 1. MorroW 2-1 [5-3] 2. Protosser 2-1 [4-4] 3. LaLuSh 1-2 [4-4] 4. TLO 1-2 [2-4]
If the tournament uses group results to seed the next round, as DreamHack did, then Morrow's ranking above Protosser carries some value.[5] Being paired against a player with a weaker performance is certainly an advantage, but a better seed is the opposite side of the spectrum from outright elimination. If tiebreaker rulings must be used, using them like this to establish pairings is tolerable.
There are two other possible ties in four man groups. Both involve three of the four players in a rocks-paper-scissors relationship, contesting one or both of the advancing spots. Look at this group from Assembly 2012 Winter:
Group F MRO GSO 1. TaeJa 3-0 [6-1] 2. elfi 1-2 [4-5] 3. Jinro 1-2 [3-5] 4. Adelscott 1-2 [3-5]
TaeJa takes first place without contention, but we still have three equally matched players vying for the second spot. Not only did elfi, Jinro, and Adelscott split matches with one another, each match had the same result [2-1]. The deciding factor proved to be the series between elfi and TaeJa -- while TaeJa swept both Jinro and Adelscott [2-0], he lost the second game in his win [2-1] over elfi. This earned elfi the second advancing spot.
The more common possibility[6] is when the three players tie for first place rather than last such as in this group, also from the third stage of the same DreamHack event:
Group D MRO GSO 1. LiveZerg 2-1 [5-2] 2. Cytoplasm 2-1 [4-2] 3. TitaN 2-1 [4-3] 4. traveller 0-3 [0-6]
LiveZerg beat Cytoplasm, TitaN beat LiveZerg, and Cytoplasm beat TitaN. After eliminating traveller there are still three players when only two can advance, so DreamHack used map score to break the tie. The key game was in the series between TitaN and LiveZerg -- LiveZerg had managed to win the second map, giving him an extra win and first place while the extra loss eliminated TitaN.
In these situations, one or two players is simply eliminated on the spot, despite a marginal difference in performance compared to one that advances. The logic behind the decision, while known ahead of time, still leaves a bitter taste in the mouth of the player and his fans. Regardless of which method is used to break the tie, the player is eliminated based on old information -- neither the player, his opponent, or the viewers knew explicitly during the match that this would, not could, mean elimination. Sometimes1 the information is so indirect and complex that explanations and deliberations are needed to sort everything out. All the tension and excitement that make these contests intense for the players and entertaining for the spectators is thrown away so tournaments can proceed with flow charts rather than competition.
Using larger group sizes not only do we see ties more often, but they can become more complicated as well. They cease to be an occasional annoyance and almost become expected. Let's take a look at what happened in a few tournaments which used this type of format, starting with the GSL 2011 Blizzard Cup:
Group A MRO | Group B MRO 1. DongRaeGu 3-1 | 1. MMA 3-1 2. Mvp 3-1 | 2. Leenock 3-1 3. MC 2-2 | 3. Polt 3-1 4. Stephano 2-2 | 4. NesTea 1-3 5. HerO 0-4 | 5. NaNiwa 0-4
The rules stated that the top three players would advance to a single elimination bracket, with the first player receiving a bye. In Group A, this gives us two two-way ties worth consideration -- DongRaeGu and Mvp both advance, but being awarded first place carries quite a bit more value than just preferential seeding. MC and Stephano, on the other hand, simply want to stay alive. GSL deferred to the head to head results, giving DongRaeGu first place due to his win over Mvp and MC third place due to his win over Stephano. At first blush this seems perfectly reasonable, but is it? Due to the larger group, the results of the players tell more of a story. MC and Stephano both lost to Mvp while beating HerO, but Stephano won against DongRaeGu when MC couldn't. Stephano will not proceed despite earning a win over not just an advancing player, but a player granted a spot in the semifinals. MC, on the other hand, advances with only losses against two players left in the tournament and no results against the other three. Along those lines, DongRaeGu had indeed beaten Mvp but took a loss to Stephano, a player who fared poorly and whose performance was deemed too weak to continue. Should these results grant him such a valuable placing without question?
There was also a tie in Group B. While all three players would advance, first place would get the bye into the semifinals and second place would get a better seed in the bracket. MMA beat Leenock, Leenock beat Polt, and Polt beat MMA. Since each match was a single game, GSL did not have the option of using map score and instead went to a playoff. MMA beat both players for first place, and Leenock beat Polt to take second.[7]
IEM Season VI World Championship had six players per group and plenty of chaos. Out of four groups total, two had ties:
In Group A, viOLet, JYP, and Socke have a three way tie contesting the two spots to advance below MMA. This tournament used overall map score as a tiebreaker, so despite the rocks-paper-scissor relationship between the three, Socke was sent home by a razor-thin margin due to his game losses in matches he won against both Killer and Strelok. Group B gave us a four way tie and is perhaps the best demonstration of how foolish breaking ties without extra games can be. With three [2-0] wins and no shutout losses MC took first place with elfi close behind in second. HasuObs and Nerchio had an identical map score, so HasuObs's match victory from their meeting gave him third. Look at the disparity here -- despite very similar showings of 3-2, Nerchio is eliminated in 13th-16th place and wins $700 while MC is given a bye into the top 8, guaranteed $2,000, and is just three matches away from the gold.[8]
In the structure of MLG 2012 Winter Championship no group participant is eliminated outright, but each rung above sixth place until second corresponds to a one round bye in the championship bracket.[9] While being penalized a round or two is not nearly as cruel as elimination, a single bye in a competitive bracket has a significant effect on a player's potential in a tournament. In four six man groups there were five ties!
There's a three way tie for 3/4/5 in Group A, a two way tie for 2/3 and 5/6 in Group B, a three way tie for 4/5/6 in Group C, and a three way tie for 2/3/4 in Group D. Thirteen of twenty four players had their eventual placing, prize money, and ever valuable circuit points strongly affected by the nuance of the rule-set. That's more than half. At what point does it become ridiculous?
Let's shift our focus to MLG's first Spring Arena. MLG used a novel system, a single group round robin into a knockout bracket for its eight man invitational. Not only does an eight man group mean many more matches, an invitational based on merit suggests that the players are closely matched in skill level. This shows in the pool results:
MarineKing undoubtedly marked first place as his own by finishing with six wins when no other player even managed five, while HuK clearly took last place with only a single win. GanZi, viOLet, and MC all finished next in a three way tie with Heart, DongRaeGu, and PartinG in another tie below them. MLG rules allowed for six players to pass into the next round -- after giving MarineKing first place and HuK eighth, this left six players and only five spots. GanZi, viOLet, and MC were guaranteed safe passage, but MLG rules also granted a first round bye to the top two finishers, as well as choice of first round match to the third place finisher and choice of second round match to the first place finisher. This means MLG would use their tiebreaker criteria to award one player a bye (2nd), award a second player first round match choice (3rd), and eliminate a third (7th). All in a tournament with only eight players total! This is a lot of factors left up to the stroke of a pen rather than the click of a mouse.
Take a step back to realize exactly what's at stake for each group of players. Second place in the pool would receive an automatic bye into the semifinals for this tournament. This guarantees them at least $1,450 with a better chance at winning first or second prize worth $5,000 or $3,000. Furthermore, the semifinalists in the tournament receive a travel-paid invitation to participate in the next Spring Arena as well as the Spring Championship, both of which have larger prizes and more prestige. While finishing 3rd and 4th in the group stage does not preclude a player from the semifinals, he will have to win one more match against one of the best players in the world to make it there. The player finishing third place gets the added bonus of making his choice of opponent between the players finishing fifth and sixth, theoretically increasing his chances of winning the critical first round.[10]
Things are much less sunny for the trio of Heart, DongRaeGu, and PartinG. The player slotted in seventh place is automatically relegated to the qualifiers for the next Spring Arena -- a tough online tournament with significantly less exposure and very difficult to pass through.[11] While the fifth and sixth players are still in danger of relegation, they will receive one more shot -- a single match to claim the majority of the bounty at stake in the entire tournament, circumvent the qualifiers, and be within arm's reach of the championship trophy.
How did this play out? Had MLG used GSO as the primary tiebreaker, four of the six players would be resolved. They didn't. Their priority went MRH, then GSH, then GSO, and then, only then, playoff matches. So you can follow along, here is the match information:
Tie for 2/3/4 GanZi vs. vioLet 2-1 GanZi vs. MC 3-0 viOLet vs. MC 2-1 Tie for 5/6/7 Heart vs. PartinG 3-0 Heart vs. DongRaeGu 1-2 DongRaeGu vs. PartinG 1-2
GanZi's wins over both viOLet and MC secured him second place. viOLet beat MC in their game, ranking him third and MC fourth -- all according to MRH. As it happens, the rankings come out the same regardless of which of the three criterion is used first -- GSO is listed in the first set of standings and GSH adds up to GanZi [5-1], viOLet [3-3], and MC [1-5].
Our next tie was not as cut and dry, even for the referees. There was a delay in announcing the finalists while tournament admins reconsidered the standings and the rules they themselves had written -- perhaps realizing the error of their ways when faced with a real world example. PartinG had beaten DongRaeGu, Heart had beaten PartinG, and DongRaeGu had beaten Heart. Unlike the previous example, MRH settled nothing. Moving on to GSH, PartinG's loss to Heart was [0-3] while the other two matches went [1-2]. This ruling gave Heart and DongRaeGu passage to the next round and sent PartinG home. Had MLG used GSO before GSH, Heart would have broken out of the tie and taken fifth place. PartinG and DongRaeGu would remain tied with a GSO of 9-12, but with PartinG having won the match between them during the round robin, he would have taken sixth place instead.
DongRaeGu advanced on the back of Heart's performance, not his own. Both he and PartinG had been shutout by one other player in the group. DongRaeGu's [0-3] loss to MC was not factored into the tiebreaker, because MC was not part of the tie. Heart was, so PartinG's [0-3] loss in their match cost him the tournament. Some looked at this situation and cried foul -- PartinG had beaten DongRaeGu, but was eliminated, while DongRaeGu received a second chance he ultimately used to win the entire tournament. I don't find it to be unfair to PartinG so much as it is peculiar. If PartinG traded his [0-3] loss to Heart with a [1-2] loss to MC, viOLet, or MarineKing, he would have advanced while the eventual winner of the tournament finished in 7th place. DongRaeGu's exceptional performance in the final bracket is compelling evidence that he deserved to be champion, but the slightest variation in the rules would have denied him the chance to even try. On the other hand, we have no idea what PartinG was capable of. To implement rules that generates this sort of situation is unnecessary and harms the integrity of the entire tournament.
Looking at this web of results, we could cite all sorts of supposed injustice. MC holds the only win in the group over MarineKing, making him arguably the most fit to face defending champion and top seed MarineKing in a championship match, but the tiebreaker rules cost him both the bye into the semi-finals and match choice. Of the other tied players, PartinG had the only win over GanZi, second place finisher and automatic semi-finalist. While a head to head result tells us how two players lie in relation to each other, a tournament is an attempt to give us a relative ranking of many players all together. In this light, small errors further down the rankings such as transposing two players who will always finish between 5th and 7th is trivial compared to the catastrophe of eliminating a potential champion much too early. If MarineKing and GanZi finish first and second, it could be argued that a win against one of those two players should be worth more when choosing which mediocre performing player(s) should continue and which should not, since those wins suggest a higher likelihood for winning the entire tournament.[12] A tournament could even choose to move past immediate results all together and use initial seeding,[13] results from previous tournaments, ladder ranking, or simulate matches with a rating system.
Valid or not, these arguments are all just nitpicking. I bring them up not to lobby for different or more complicated tiebreaker standards, but to show that all methods are sorely lacking in integrity. The goal of a tournament for the competitors is to put their skills to the test, under pressure, for all the marbles. The goal for the spectators is to see exciting matches and cheer on their favorite players in battle. By relying on a multitude of criteria, any criteria, rather than playoff matches, tournaments rob the players of an opportunity to compete knowing what exactly is on the table, and fans are robbed of the nail-biting suspense of watching their favorite player try to rise from near elimination. How can anyone be happy when it becomes clear the deciding match in a one player's tournament happened hours earlier, without anyone realizing it? The solution to the problem isn't to argue over methods slightly more fair than flipping a coin, the solution is to default to the notion that Starcraft tournaments should settle these issues with Starcraft, not arithmetic.
The rationale for having these rules in the first place likely comes from the administrative side of the tournament. It's plenty difficult behind the scenes to keep everything from crumbling down without the added stress of falling behind schedule as you try to jam extra games into the itinerary. But considering what is at stake, is this valid? These tournaments consist of a very large number of matches -- the group stage alone for MLG Spring Arena 1 guaranteed 84 games, while the tiered format used by DreamHack in Stockholm gave us over 300 before the RO16. Is adding a few sudden death matches such a logistical problem that it must be avoided whenever possible, even when this means eliminating a player who has earned consideration? Conventional sports all implement some sort of overtime, in spite of being aired on television and operating on much stricter schedules, and any sports fan will agree that overtime games are among the most memorable. A tournament is in the business of broadcasting games people want to watch. Why are they all so quick to avoid a match with so much on the line in favor of eliminating players based on technicalities?
Notes
1. As of four months after HomeStory Cup IV, the only other international offline event Delphi has had the opportunity to enter is the DreamHack Stockholm Open. For players who aren't invited with travel paid to every event on the circuit, big chances don't come up very often.^
2. Chronologically, the last match of the group was viOLet vs. MarineKing. Going into the match, MarineKing was in a must-win situation. If he lost, Delphi would advance to the second group stage instead. viOLet would move on regardless of the outcome, but held Delphi's fate in his hands. viOLet swiftly lost two straight. When they met again in the quarterfinals of the same tournament that same weekend, viOLet won the new series [3-2].^
3. GSL may be considered an exception by showing a strong (or growing) preference for elimination formats. In July 2011, they switched Code S group stages from a round-robin system that cancelled matches between players out of contention to a dual winner double elimination system. In January 2012, the invitational Arena of Legends: King of Kongs used this same format rather than the round-robin used in their previous invitationals. GSTL was also switched to double-elimination in 2012 after one attempt at round-robin the previous year. GSL still uses round-robin with BO1 matches for the Up & Down phase. To resolve ties, they use BO1 followed by a playoff.^
4. While there was significant controversy over DreamHack not awarding first place in the example group to TLO in consideration of his awesome beard, wildcard spots based on facial hair is beyond the scope of this article.^
5. In this particular tournament, second place matched Protosser against Polt, the winner of Group C, while MorroW went on to face Group C runner up Socke.^
6. More common based on history, not probability. In a simulation, both three way ties are equally probable. It just doesn't seem to happen that way.^
7. I don't recall hearing any viewers complain about the extra games.^
8. MC went on to win the entire tournament, worth $35,000 USD. Had he not received first place in his group, he would have not only needed to win four matches rather than three, but he would have played either a slightly or entirely different sequence of players in the playoffs.^
9. As an example from the same tournament, Socke won four matches in a row after the group stage before he was HuK's first opponent in the elimination bracket. Finishing four rungs higher gave HuK four byes. First place in the group is slotted into a different (and more advantageous) bracket rather than being given one extra bye.^
10. Getting to choose the pairings is only an advantage if a player is capable of identifying their best chance. viOLet chose Heart instead of DongRaeGu. He lost, but we don't know whether he would have fared any better against DongRaeGu, his only other choice. In the next round, MarineKing also chose to face Heart over DongRaeGu. While his victory over Heart and subsequent loss to DongRaeGu suggests that MarineKing's decision meant finishing in second place rather than losing one round prior, I can't help but point out that the semifinals were BO5 and the finals were BO7. Before losing the series, MarineKing was leading the championship 3-2 -- choosing to play DongRaeGu earlier may have won him the entire tournament. The only thing we know for sure is that Heart must not spend a lot of Friday nights alone.^
11. The South Korean qualifiers for MLG Winter Arena was a veritable gauntlet of some of the strongest competition in the world. 52 players -- 24 of whom started that GSL season in Code S and 11 in Code A -- in a double elimination bracket with just eight advancing. Even a very talented player with a marked advantage over the opposition could easily stumble over so many matches against such skilled opponents.^
12. A tiebreaker score based on the finishing strength of players beaten is a popular concept for tiebreakers used in chess tournaments. However, these methods are for swiss style tournaments, not round robin, where the field of opponents is not the same between the two players. In a round robin, it could be argued that any wins over better players are cancelled out by what must be losses to worse players.^
13. Though it didn't come into practice, the rules for Spring Arena do in fact use seeding should they be unable to avoid extra games (see Round Robin Format, #3). According to these rules, had the three way tie gone to overtime, DongRaeGu (seeded #2) would have received fifth place automatically with the winner of a BO1 match between Heart (seeded #3) and PartinG (seeded #5/6) advancing in sixth and the loser eliminated.^
