New Post has been published on TRADE MASTER TEAM
New Post has been published on https://trademasterteam.com/forex-strategies/forex-backtesting-the-fatal-error-of-curve-fitting-trading-strategies/22693
Forex Backtesting The Fatal Error of Curve Fitting Trading Strategies
welcome to the survey which is about Forex back testing I basically getting a strategy together or system working out rules by doing hypothetical testing on historical data game backwards and getting a strategy that makes money now how much use is it my own view is it's not much use at all because back testing is really fatally flawed now in terms of why is it fatally flawed well the reason is obvious you've got the historical data you knew what happened and you ferment carafe it and bend the strategy to fit the data though that is a bit similar do just going shooting blindfolded at a wool then basically going over after the event drawing a circle around each shot saying it's a bull's-eye you're doing it in hindsight I think what you find with some traders is they were obvious they really want to have a profitable trading strategy so they're putting just too many variables so for example if the first set of rules doesn't work they'll tweak the rules were adding even more rules they'll just bend the system to fit the data the more rules in a trading system is a general rule but more likely the system is not going to work in real time they're in bad testing I have seen people come up with profitable trading systems okay I can give you a single rule trading system that works that by a four week I hold the position then reverse a four week low hold the position until new four week highs hit that is called Richard dungeons for weak rule now that will actually make profits over the longer term but you have a lot of drawdown okay and basically long periods of flat trading it'll make money but not a huge amount of money but it will work because it's only one rule you cannot kerf it a one rule sister I think system rules yeah to work once to get about three it's not going to work okay and the problem is with the very simple systems that are back tested they have too much drawdown too long in terms of flat periods in my view and not enough profit no I think we live in an age where people yeah want to use computers they make people feel comfortable and you know make them feel safe and yeah back testing is a way of making a trader overconfident in the strategy most traders do not want to have responsibility for their trading rules they want to make seem automated as much as possible and it just simply will not work if you look at the best traders in the world they'll have a rule set okay they have to have rules that they know that basically they hang out the filter the trading signals they need the basic decide how much money should go on a trading signal etc etc so they do have rules but they're very flexible in their approach to the rules now filter out you know the trades that they think are the best and the idea that Forex is computerized or you can computerize it and make it game of science it's taking wrong it cannot be against science and humans are involved and humans are unpredictable and basically what you get effectively in the markets they look yeah very sort of orderly going backwards they're always subtly changes always subtle changes in the markets that will break any trading strategy has been over optimized cur fitted in a bank test that's just a reality of trading there are no other business actually that I know of where traders actually think they can come in and delegate the money-making process to a computerised set of rules and not have to do anything if it were that easy a lot more people would be making money so I think you have a bank testing yeah you can get very simple strategies that work but normally just under three rules okay they're not going to make enough money that not too much drawdown to longer flat periods okay the most more rules the more likely you are to have curve fitted it sorts the best way to make money and how do you really get a profitable trading strategy and be confident in the answer is forward testing what I mean by that is you basically find some high probability yet chart patterns or trade setups okay then work out a basic rule set to trade them and then basically filter each setup yourself okay the only way of doing it for X rating yeah it can be very simple but you cannot put yourself outside the equation it's you and your strategy and you are responsible for it and implementing sorry implementing it and you must actually be involved you must think about the trade setups you must take responsibility if you do that you can win do you think you can just back tests come up with a strategy that's going to just make money I would actually say yes forex trading would probably not be fee in forex trading suits someone who wants to be involved in decision making process takes full responsibility for their signals and their stretching that's the video for today thank you very much watching me as future you
