[Build 112] "too many trades closing at the same bar" with M1 intrabar backtests?

I am getting a "too many trades closing at the same bar" on an H1 strategy that is backtested with "1-minute data tick simulation (slow)". Yes, there are trades closing within the same H1 bar, but all of them are lasting about 15 minutes still, and given that the strategy is backtested with intrabar M1 data (which is all correct and has no holes), the backtests should be absolutely OK and that warning should not be there. On top of that, I am trading this strategy live for a long time already without any issues. Live trades, even the ones that close within an H1 bar often, are all replicated exactly in the M1 intrabar backtest in SQX and MT4.



I *DO* agree that this warning is absolutely correct if the backtest mode would be "selected timeframe only (fastest)" because then we can indeed not know what happened inside the H1 bars - it should appear in that case, but not if we backtested with intrabar M1 data.


Thanks :-)

Attachments
No attachments
  • Votes +1
  • Project StrategyQuant X
  • Type Bug
  • Status Refused
  • Priority Normal

History

g
#1

geektrader

18.10.2018 09:04

Task created

g
#2

geektrader

18.10.2018 09:13
P.S.: I know that this can be tricky and that this warning, of course, should appear too if we backtest a strategy that TRADES on the M1 timeframe and has no further, lower resolution (tick data in that case). In this case, we wouldn´t know what happens first inside such bars where the trade closes at the same bar too. However, for example, an M5 strategy backtested with M1 intrabar data would be OK again. Or, of course, an M1 strategy that uses real tick-data, etc. But anything higher than M1 as trading timeframe AND being backtested with M1 intrabar data should never show this warning (nor any about "ambiguous bars").
g
#3

geektrader

18.10.2018 09:13
P.S.: I know that this can be tricky and that this warning, of course, should appear too if we backtest a strategy that TRADES on the M1 timeframe and has no further, lower resolution (tick data in that case). In this case, we wouldn´t know what happens first inside such bars where the trade closes at the same bar too. However, for example, an M5 strategy backtested with M1 intrabar data would be OK again. Or, of course, an M1 strategy that uses real tick-data, etc. But anything higher than M1 as trading timeframe AND being backtested with M1 intrabar data should never show this warning (nor any about "ambiguous bars").
g
#4

geektrader

18.10.2018 09:14
Voted for this task.
MF
#5

Mark Fric

23.10.2018 18:52

Status changed from New to Refused

the warning is not related to testing precision. It will be there if there are trades that begin and end in the same main bar, in your case H1. 

It is usually not recommended to use this kind of strategies, if you don't know what you are doing.


If you have this kind of strategies I suggest to turn off this filter so you don'w see this warning.

g
#6

geektrader

23.10.2018 20:36
I fully agree with what you say, however, to make this correct, this message should not be there if backtested with M1 precision, because it isn´t a problem in that case if they close within the same bar or not.
m
#7

mabi

24.10.2018 00:35
You have to go to Real tick in order to claim that Intrabar open and close calculation is correct because depending on instrument price can travel far in 1 min, I turn it of when using real tick if I have any left . I tried to build intrabar trading strategies manually for a hole year once it never worked well it worked for periods but then it always comes a short period when it does not work that becomes devastating to the strategy. Allthough I have to add one thing even if the I turn it of I still get to many strategies closing on the same bar failure. 

Votes: +1

Drop files to upload

or

choose files

Max size: 5MB

Not allowed: exe, msi, application, reg, php, js, htaccess, htpasswd, gitignore

...
Wait please