365 trading days per year - I think not!

[stocktraderbmp]

Hi small point but the analyzer divides the total number of trades by 365 to give trades per day when in fact most analysis uses an average 220 trading days per year counting only official market opening days, so NOT Sat / Sun and holidays.

I'm sure this will have been flagged before but would anybody else like to see this changed? The analyzer uses 365 trading days per year instead of 220 ! Is there a reason for this as it would seem to render any generated results very inaccurate.

It does after all render any statistic ie profit per day, loss per day, etc etc etc all useless if the 365 has been used throughout all of the calculations within the analyzer, if it hasn't been used in other calculations why use it for the trades per day ............ Hmmmmm !!

Not a massive issue to resolve but it would at least start to produce some more accurate results ...... possibly.

Maybe even a variable could be added here so those that can't trade every day could put in a figure say 100 trading days and the analyzer will select a random 20 Mondays from 50, a random 20 Tuesdays from 50 etc etc assuming the look back period was 1 year otherwise it would be a random 40 Mondays from 100 if the look back were 2 years etc etc.

Please will all who read this add their thoughts.

[NinjaTrader_Bertrand]

Thanks for the input and feedback stocktraderbmp, I can confirm it looks like 365 days are used - I will submit to product management to enhance this for the future.

[stocktraderbmp]

thanks but worrying !!

[NinjaTrader_Bertrand]

I will update you when I have new input.

[NinjaTrader_Bertrand]

stocktraderbmp, thanks again for bringing this to our attention - we're looking into matters and tracking this under # 1509 on our product development lists.

Have a great weekend,

[stocktraderbmp]

Thanks, was analyzing this today whilst doing my other job ! What else does the 365 affect ? All the trade data is non-day specific so # of trades etc is unaffected as are any stats derived from the trade data. So can only assume it affects # of trades per day and recovery time. Not such a massive issue after all - hope you have a good weekend as well and thanks for your efforts B.

Still think a variable period with random sampling for each day of the week is a good idea for those that can't trade everyday and only want to trade say Mon, Wed & Fri.
This would also provide a facility to scramble data as well for those that want to do additional back testing on say out of sample data or generate years and years of data to rigorously back test a system in all market types with out having to download 20 years of data. This was a recent discussion on just such a point although I don't think the initial query was very well thought through and at the very least confused. I still don't know what the man was expecting as there is 50+ years of back testing data if your prepared to pay for it. If he just wanted to produce some numbers then the suggestion offered to scramble data seemed like a perfectly reasonable solution but then you can't please all the people all the time.

http://www.ninjatrader.com/support/forum/showthread.php?t=46689

What ever its used for, a facility to scramble data would be a great USP, imagine advertising that you can offer hundreds of years of data for back-testing in all futures and forex markets !! No other platforms offer that as far as I am aware.

Ben.