thanks for that, looks like it could be worth breaking into 2 separate algos.
Hello, could you tell me how it can be only long or only short? I would like to learn, please
Simple but inelegant: duplicate the code, then delete or // anything related to short on one and anything related to long on the other. Then they can both be re-optimized.
Otherwise you can add a long/short selector code, such as the one Paul uses:
once tradetype = 1 // [1] long/short [2]long [3]short
then you have to put your entry conditions inside an If statement:
if tradetype=1 or tradetype=2 then
(long entry conditions)
endif
if tradetype=1 or tradetype=3 then
(short entry conditions)
endif
Here’s a Long only update for this one – 6 year backtest, 75/25 opt. The 18 months OOS is much better than it should be … a good problem to have but still a bit weird ???
I feel it could be improved, esp with some sort of exit conditions, but nothing I’ve tried seems to help. All ideas welcome.
This has max positions = 3, above that makes more money but less stable.
Hello @Nonetheless!
Thanks for this sharing! I don’t have quite the same results as you however. In the screenshot you’ve provided I can see that it’s written “opt” at the end of the backtest name. Do you have an optimised version?
This one looks pretty good, I see if I spot something to improve 😉
the ‘opt’ version is just my working algo with the variables in the opt box. The one I posted has the same values inserted in the code.
How do your results differ?
Ah ok! Do you have this algo in real?
I have 2.80 of gain/loss ratio. The curve looks quite close anyway 😉
I have 2.80 of gain/loss ratio
On a 6 year backtest?
I’m not running it live yet, hoping for some inspired suggestions on how to improve it…
Hi @nonetheless
I don´t get the same result as you either, any ideas?
I have no idea why you would get a different result, although I have had some v strange experiences with backtests in v11.
I ran it again just now and I got the attached print.
Thank you! Oh, I have had many strange experiences with PRT overall…
Did you run it again from your installed algo, or did you try to install your published itf again?
it’s the exact same that I published.
I have been using this algo for only a week, so I don’t have a ton of data for this observation. The trades the algo has takes so far has always been 3 at a time and it has always happened almost at the same time, just a few moments in between. I am aware of the Cumulative order setting and I see in this thread that max number of positions are 3.
Is it intentional that the cumulative orders are picked up at the “same” time however? Shouldn’t they be paced a bit more, like if the algo takes a trade the conditions for that trade is “nullified” and it will not take another trade until the conditions have happened again based on the moment of the first trade, which would likeley be some time in between.
I am total novice on interpreting the algos but I am trying to learn, and maybe the behaviour I suggest is just not possible?
It sounds like a good idea, maybe you can add a variable (“space”) and try to add something like : (not tested)
Cspace = (barindex-tradeindex) > space
IF not longonmarket and Ctime and c1 and c1a AND C3a and c3b and c3c AND C5a and c5b THEN
BUY positionsize CONTRACT AT MARKET
elsif longonmarket and Ctime and c1 and c1a and c3 and c3a and c5 AND C5a and COUNTOFLONGSHARES < MaxPositionsAllowed and Cspace then
BUY positionsize CONTRACT AT MARKET
ENDIF
And then optimize “space”. But Im not sure of the Tradeindex behaviour when accumulating positions.
