I am doing a lot of backtesting at the moment.
Moving forward with my trading in 2025 I could trade with a higher win rate of around 60% with a 1:2.2 R and say 180% profit
Or
50% win rate with a 1:2.65 R and 200% profit.
Which would you go for?
For me I think I am going to pick the first one as having a higher win rate is better for my mental state and the risk of significant drawdown should be less.
Thanks
Out of those two, I strongly agree that the first is less dangerous than the second, for exactly the reasons you said.
But I honestly wouldnât trade either. Even the first is far too unrealistic for me to try it.
I canât believe any retail forex trader could possibly trade consistently with a 60% win-rate and a 1:2.2 R.
Youâd be doing better than 99% of the worldâs fund managers!
Sorry, but thatâs just not going to happen, and if itâs what your back-testing appears to show, then your back-testing isnât accurate.
Thanks for the reply.
What are you saying is unrealisticâŚthe returns or the 60% win rate at 1:2.2 R?
Backtesting has been done on the 4h timeframe and have been as realistic as possible taking every setup except the 2am ones. The strategy is very mechanical and straightforward with little room to be ambiguous.
Obviously some trades will be missed but the win rate should even itself out to be about right. I have 250 trades in the backtesting data so far over 9 pairs starting January 24.
The only thing which I feel will mess it up is the spread which I havenât been able to take into account. Shouldnât be too bad on the 4h though and the RR is high enough to absorb it I think.
Any advice about where you think I am going around would be appreciated.
The 60% win-rate at 1:2.2 R.
No retail trader has ever maintained anything like that.
Seriously, I was understating it greatly, when I said that youâd be doing better than 99% of the worldâs fund managers: youâd actually be doing far, far better than 99.9% of them!!
Sorry to sound negative, but if you look at it dispassionately and logically, you can surely appreciate that the chances of this happening, for a comparatively inexperienced retail trader, canât possibly be much higher than 0%?
Nobody can confidently tell you where your backtesting might be misleading you, without going over it all with a fine-toothed comb (not really something you can do via a forum, obviously).
The common problems are small sample sizes (250 trades over 9 pairs is very, very little), overfitting (e.g. removing the 2.00am trades because they happened to be significantly worse than others over the sample you tested), not allowing for spreads/dealing-costs (I do agree that spreads âshouldnâtâ be hugely significant on a 4-hour system), using only a âsuitableâ market (e.g. if you test a mostly-long-trades system over a period of time when the market was mostly rising/uptrend, itâs bound to look very profitable until you start using real money and suddenly the market isnât âmostly risingâ any more!), etc., but there are many others, too.
Anyway, the thing to do next is clearly to forward-test on a demo account: do 300 trades exactly as you would with real funds, and see how well your back-testing âpredicted outcomesâ do, that way?
OK. I just donât know where I can be going wrong to be honest. I understand that these figures look good but didnât think they looked that good!
I donât mind negativity, its nice to be able to talk to someone else about it - trading can be a lonely place!!
You say the chances are very small - 0%, so what sort of figures would you expect from retail traders?
I keep doing more back testing each day and hope to have a bigger sample size - those 250 trades are the equivalent of around 70 months of backtesting and the only reason I donât do the 2am trades is because I would be sleeping and wouldnât be able to take them (UK GMT).
I do understand your points and I am skeptical but just cant see what it is I am missing.
My biggest concerns are correlated pairs giving big drawdowns and spreads at certain times of day as well as my own ability to trade the system!! Will only need to check the charts every 4 hours though and will be using limit orders most of the time to set and forget.
I will start forward testing next week but on a small live account - hate trading demo accounts. 300 trades is probably about 6 months worth of trades but will see after a month how well it goes. I will update on here if either good or bad.
The performance sounds impressive but from experience a high win rate of even slightly higher than 50% tends to be partnered with a low r:r, like 1:1.
The fact that you declare r:r to 1 decimal point or even 2 decimal points simply emphasises that this is all just theory.
What would really impress me would be demo trading this strategy. Believe me though, you absolutely must run this strategy, it will be the only way to gain the full picture, and if there is even just a small possibility it will be a consistent winner, you must find out. So definitely do it, preferably in demo, but definitely report the results here.
50 trades (1 month) will tell you nothing helpful.
Even 300 trades wonât/shouldnât give you complete confidence.
Iâll tell you what I would do, myself. There will be some people who will disagree with this, and it isnât âadviceâ: itâs just what Iâd do, myself. In addition to doing what you suggest above (not as an alternative to it) I would simultaneously test on demo a second version in which I cut that 6 months down to 6 weeks by using 1-hour bars instead of 4-hour bars.
Itâs very likely to give you at least some indication of reality.
It will be less reliable than 4-hour bars, but only slightly.
The system wonât and couldnât, realistically, produce from H4 bars anything like the outcomes predicted by your backtesting without producing something at least similar on H1 bars. This will, in practice, be a much faster (and only slightly less reliable) indication for you.
But understand two things clearly. Firstly, Iâm not suggesting that you should/would actually trade this for money using the faster method: itâs just a quick, additional way to try to identify the problems, and part of âtestingâ rather than âtradingâ. Secondly, if thereâs some sort of specific market/related or timing-related reason for not doing that, of which Iâm unaware (an example might be if itâs an âopening range breakout systemâ) then ignore me!
Thanks guys. I will keep doing my backtesting to get as much data as I can. I will also trade this on the 4h with the higher win ratio setup as I think it will be more forgiving.
Trading the 1h would definitely work but means I would have to check the charts every hour and I just donât have the time and would be inconsistent as a result.
The RR is in decimals purely because thatâs what it is based on the fib entry, sl & tp.
I will update this thread with any results I get.
Iâd go with the 60% win rate and 1:2.2 R as well for the stability it offers.
First option for me as well.