Nice paper backtest, probably will fail in the future, why? :( | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Why I can't backtest before Nov 2015 in FxPro Live? 3 replies

Visual backtest result different from non-visual backtest 0 replies

Why Paper Trading Is Counter Productive 6 replies

Nice indicator , nice profits , nice system 8 replies

Will a backtest fail if a trade will blow my account? 10 replies

Attachments: Nice paper backtest, probably will fail in the future, why? :(
Exit Attachments

Nice paper backtest, probably will fail in the future, why? :(

  • #21
  • Mar 19, 2021 1:56pm Mar 19, 2021 1:56pm
how can anybody answer that question if we dont know your system?
  • #22
  • Mar 19, 2021 2:06pm Mar 19, 2021 2:06pm
Quoting yonnie
how can anybody answer that question if we dont know your system?
It is a general question my friend. This strategy is just an example that I show here. We all know that most strategies fail in the future, even if they have a perfect backtest in the past. Pay attention, I say most, not all
  • #23
  • Mar 19, 2021 3:46pm Mar 19, 2021 3:46pm
I would be surprised if a general answer to your general question would be valid for every method of trading.......in my opinion different systems have different ways to avoid or lessen the pitfalls of trading that system; there is no 1 solution.......
  • #24
  • Mar 19, 2021 4:16pm Mar 19, 2021 4:16pm
Quoting yonnie
I would be surprised if a general answer to your general question would be valid for every method of trading.......in my opinion different systems have different ways to avoid or lessen the pitfalls of trading that system; there is no 1 solution.......
Of course there are concrete causes why a strategie fail, but there are more common causes in all strategies, which make them fail in the future and that is what
i am searching for.

Do not be surprised if I tell u that one of the well known causes of a lot of profitable strategies in backtest and not profitable in the future is the abuse of add many filters, but this is so basic, I thnk someone should have some move advanced theory on "Why this stratetegy will be loss in the future and how to fixed?"

Anyway, yonnie, I appreciate your contributions
  • #25
  • Mar 19, 2021 4:50pm Mar 19, 2021 4:50pm
Quoting Kfx
New update. Just add to MyFxBook, so far so good {image} http://www.myfxbook.com/members/FxFr...r-five/8101202
Looks nice.

But the win of over 13% on March 17 kinda worries me. That bet on the FOMC is quite risky IMO.
  • #26
  • Mar 19, 2021 6:18pm Mar 19, 2021 6:18pm
Quoting Kfx
Hi, Long time out here and out Forex. Now I'm back, just for fun. We all know most of EA, paper or visual baktests tend to fail in the future. Now I have one of the hundreds of strategies that I have thought: 3,5 year paper Backtest, more than 2000 trades: 6 losses (20% each) 14 wins (11% each) 20 wins ( 6% each) 22 wins ( 3% each) 52 wins ( 1,5% each) 151 wins ( 0,5% each) 1875 wins ( 0,1% each) My question is, Why this stratetegy will be loss in the future and how to fixed? Thanks for read
Strategy properties:
- Martin-gale strategy.
- Risking 0.4% more or less in the first trade.
- Trailing stop also in percentage terms.
- Possible entry logic: after some determined price speed. (It will take me some time to get into a final conclussion.) Otherwise, some 1 minute logic.

I guess you are looking for someone who tells you: because of the martin-gale logic.

How to fixed it? You are kidding!
Always ready to help
  • #27
  • Mar 19, 2021 7:59pm Mar 19, 2021 7:59pm
a common cause for all strategies to fail is that market behaviour becomes different from what the system is designed for.....
if the change is becoming persistent, one might have to stop trading the system, until the favourable behaviour returns....there could be a need to adapt to the changing circumstances.......for instance for trend trading you need directional volatility and if that falls away, you will get whipsawed all the time.....so it might be time for the range trading strategy etc etc......
  • #28
  • Mar 20, 2021 5:23am Mar 20, 2021 5:23am
Quoting brian.jer
{quote} Looks nice. But the win of over 13% on March 17 kinda worries me. That bet on the FOMC is quite risky IMO.
Thanks for ur reply Brian,

It is totally automated strategy and do not have news filter. This trade start before the new and closes in the new. The win % will be aprox. the same with or without the new.

In this case the new does not affect to much the trade, just a little on my favor (0,5 pips more because the volatility)
  • #29
  • Mar 20, 2021 5:29am Mar 20, 2021 5:29am
Quoting PaulMF
{quote} Strategy properties: - Martin-gale strategy. - Risking 0.4% more or less in the first trade. - Trailing stop also in percentage terms. - Possible entry logic: after some determined price speed. (It will take me some time to get into a final conclussion.) Otherwise, some 1 minute logic. I guess you are looking for someone who tells you: because of the martin-gale logic. How to fixed it? You are kidding!
Thanks for your reply Paul!

Strategy properties:
- Martin-gale strategy. Not exactly
- Risking 0.4% more or less in the first trade. No
- Trailing stop also in percentage terms. No
- Possible entry logic: after some determined price speed. (It will take me some time to get into a final conclussion.) Otherwise, some 1 minute logic. No

I appreciate a lot u take the time to review the trades in myfxbook, I'm sure u will get better conclusions with more trades in the future, maybe, if the strategy doesn't go bankrupt hahahaha
  • #30
  • Mar 20, 2021 5:53am Mar 20, 2021 5:53am
Quoting Kfx
{quote} Thanks for your reply Paul! Strategy properties: - Martin-gale strategy. Not exactly - Risking 0.4% more or less in the first trade. No - Trailing stop also in percentage terms. No - Possible entry logic: after some determined price speed. (It will take me some time to get into a final conclussion.) Otherwise, some 1 minute logic. No I appreciate a lot u take the time to review the trades in myfxbook, I'm sure u will get better conclusions with more trades in the future, maybe, if the strategy doesn't go bankrupt hahahaha
- Martin-gale strategy. Not exactly. Of course it is. Call it whatever you want.
- Risking 0.4% more or less in the first trade. No. All first trades closed at around 0.4 unless there was a chance to trail.
- Trailing stop also in percentage terms. No. Maybe the percentage part is wrong, but you are trailing for sure.
- Possible entry logic: after some determined price speed. (It will take me some time to get into a final conclussion.) Otherwise, some 1 minute logic. No. That was very pretencious from my side hahaha Sorry, but I feel that I had to try!


You know that what will kill your startegy is how agressively it increases the trading volume every time you get a loss trade. Some call it martin-gale, if you don't want it to call it that way because there is a more advanced logic behind that, then good. But I guess you are after that problem.
Always ready to help
  • #31
  • Mar 20, 2021 6:50am Mar 20, 2021 6:50am
Martin-gale strategy. Not exactly. Of course it is. Call it whatever you want. Again, not exactly
- Risking 0.4% more or less in the first trade. No. All first trades closed at around 0.4 unless there was a chance to trail. No, the risk vary a lot at every trade and/or phase of the strategy, remember that the stratagy have 3 main phases
- Trailing stop also in percentage terms. No. Maybe the percentage part is wrong, but you are trailing for sure. No trailing stop, maybe u say that because u see different pip amounts or becouse sometimes the trade closes without SL or TP.
- Possible entry logic: after some determined price speed. (It will take me some time to get into a final conclussion.) Otherwise, some 1 minute logic. No. That was very pretencious from my side hahaha Sorry, but I feel that I had to try! There is not only one entry logic in the strategy

Quoting PaulMF
{quote}You know that what will kill your startegy is how agressively it increases the trading volume every time you get a loss trade. Some call it martin-gale, if you don't want it to call it that way because there is a more advanced logic behind that, then good. But I guess you are after that problem.
The risk still 20% in a loss, doesn't matter if I trade big or small lots
  • #32
  • Mar 20, 2021 6:57am Mar 20, 2021 6:57am
Quoting Kfx
Martin-gale strategy. Not exactly. Of course it is. Call it whatever you want. Again, not exactly - Risking 0.4% more or less in the first trade. No. All first trades closed at around 0.4 unless there was a chance to trail. No, the risk vary a lot at every trade and/or phase of the strategy, remember that the stratagy have 3 main phases - Trailing stop also in percentage terms. No. Maybe the percentage part is wrong, but you are trailing for sure. No trailing stop, maybe u say that because u see different pip amounts or becouse sometimes the trade...
Convinced in many explanations. My apologies.!

Not in the volume increasing thing. I will keep reading. Curious on how this will end.
Always ready to help
  • #33
  • Mar 20, 2021 7:05am Mar 20, 2021 7:05am
Quoting yonnie
a common cause for all strategies to fail is that market behaviour becomes different from what the system is designed for..... if the change is becoming persistent, one might have to stop trading the system, until the favourable behaviour returns....there could be a need to adapt to the changing circumstances.......for instance for trend trading you need directional volatility and if that falls away, you will get whipsawed all the time.....so it might be time for the range trading strategy etc etc......
Thats a good one, market conditions are constantly changing and a prolonged change in time can kill your system, if it is not prepared for that change.

Stupid idea hahaha:

Create a lot of strategies, like 1 every 15 or 30 days, with 6 months or 1 year BT behind and run them during 1 or 2 months. I note that a lot of strategies have succes in the first 2 or 3 months and then fails, so it would be a good idea to do that if u share the risk between all the strategies and at the end of the day u have more than 50% of ur strategies basket winning. What u think?
  • #34
  • Mar 20, 2021 2:20pm Mar 20, 2021 2:20pm
dont think you can adapt a system that much, so that it will survive in all market conditions.....assume you`ll have to stop trading it on a real account and maybe continue on a demo account, until the system is in profit again for some time......
I remember I was into trading the first 4H breakout of the new week in G/J years ago and the first 6-7 months were great and then it was not so great.....
I dont have experience in running a lot of strategies and I can only wonder if you really need 20 to be successful......wouldnt that be overload and difficult to keep on top of all that?...imo simple is better, so maybe like 5 is plenty, because you get to know them well and knowing that they perform in certain conditions...
Would definitely not share the risk between all the strategies; would only trade the profitable ones and the rest on demo.....
just some thoughts, so feel free to discard them all
  • #35
  • Mar 20, 2021 5:58pm Mar 20, 2021 5:58pm
Quoting Kfx
... “Why this stratetegy will be loss in the future and how to fixed?” ...
The only way to prevent a winning strategy from losing in the future is to identify its pre-market condition and only play it during the same condition ... (this process of identification can’t be automated, so you need to run your EA manually)
Market is not random but unpredictable
  • #36
  • Mar 21, 2021 9:52am Mar 21, 2021 9:52am
Quoting yonnie
dont think you can adapt a system that much, so that it will survive in all market conditions.....assume you`ll have to stop trading it on a real account and maybe continue on a demo account, until the system is in profit again for some time...... I remember I was into trading the first 4H breakout of the new week in G/J years ago and the first 6-7 months were great and then it was not so great..... I dont have experience in running a lot of strategies and I can only wonder if you really need 20 to be successful......
I'm working on something similar, but with a little twist. First I have to find a way to tighten up my trade criteria to take less trades but produce a higher winning percentage, or the reverse of loosening up my trade criteria to take more trades but have a lower winning percentage. Once I have identified this adjustment to my trading system, I tighten up the criteria when my system runs cold and loosen it up when my system is hot. This should carry me through the hot and cold spells which seem to be inherent in any successful trading system.

As a greatly simplified example, say you were trading the stochastic indicator successfully by selling at 75% and buying at 25%. (I'm not, but just for an example). Say you found that increasing your sell point from 75% to 80, 85 or 90% and dropping the buy point from 25% to 20, 15 or 10% produces fewer but better trades. Then you just adjust the criteria to take more trades when the system is hot and fewer trades when the system is cold. In a sense it's a bit like a reverse Martingale, but a bit more complex and technical.

Does that make sense? Anyway, it's a work in progress. Happy trading! - G
  • #37
  • Mar 21, 2021 2:03pm Mar 21, 2021 2:03pm
hey Graviton,
I never trade with indicators, only price action......but your reasoning sounds good....backtesting over a long period will confirm your theory or not.....
if you are into contra trading similar to your example of selling high/buying low, why not have a look at monthly/weekly/daily support and resistance levels......to trade from those areas are more reliable than any indicator..... have a look at "No brainer trades" thread by Billy Ray Valentines and with a little twist at "My Baby" thread by CindyXXXX....its one of those systems that will never go out of fashion......
hope you`ll find your own awesome system
  • #38
  • Mar 23, 2021 6:15am Mar 23, 2021 6:15am
Quoting 4for4
{quote} The only way to prevent a winning strategy from losing in the future is to identify its pre-market condition and only play it during the same condition ... (this process of identification can’t be automated, so you need to run your EA manually)
Yes, but this is not easy, cristal ball is needed for that, haha

You can identify a slow range market and in one second chage to a big long tren, or you can see a big trend and the market stop quickly and start slow range
  • #39
  • Edited 8:01am Mar 23, 2021 7:34am | Edited 8:01am
Quoting Kfx
{quote} Yes, but this is not easy, cristal ball is needed for that, haha You can identify a slow range market and in one second chage to a big long tren, or you can see a big trend and the market stop quickly and start slow range
It’s not easy, but it’s not impossible ...

From your example above, the slow ranging market is the one I’d be using as a pre-market condition before playing ...
Market is not random but unpredictable
  • #40
  • Mar 24, 2021 12:39am Mar 24, 2021 12:39am
no, its not impossible to find a ranging market, but it takes time to show itself.....look for instance at G/U; it took 6+ days to show the range limits between 1.4017 and 1.3779.......you could draw a line through top and bottom and trade between those lines......if the day closes beyond the line, a trend could be starting and you might need to stop trading your range trading system until price returns between the lines or a new range forms......
0 traders viewing now