Backtesting - Strange things

[schlumpfine75]

Hello! once i was here in this forum and i got good answers. I try to test some indicators for my master thesis. the problem now is that i cant understand way very often the profit per month (included comission from 400$ per Trade) is better than the profit per month with included comission from 10$ per trade??
and once i got a better profit per month (included comission true) than the profit per month without comission?
Another stange thing is that sometimes the cumulated proift is negativ and the profit per month is positiv?? in my opinion they should have the same sign. Can someone help me please, befor i´m going crazy ?
thanks a lot
nice greetings from vienna Sigrid

[NinjaTrader_Jason]

Can you please ensure both backtests were using the same strategy and parameters. Please perform the exact same backtest with different commissions and check if the profit per month remains higher when using a higher commissions rate.

The cumulated profit can be negative while profit per month is positive.

[schlumpfine75]

Hi. i started my backtesting again and first i thought everything is okay. idon´t know what happend yesterday. but today i got one strange thing again. i load up two files to show you the strange things. one is with comission 10 and the other with 400. can you take a look to this files and tell me what happened there. thanks and nice greetings sigrid

[NinjaTrader_Jason]

Commissions are not reflected in the Total Net Profit.

Unfortunately, I don't see in the screenshots what is incorrect about it. Could you please tell me what is incorrect regarding the performance reports.

[schlumpfine75]

i don´t understand why the comission should not be reflected in the total net profit? If you take a look to the screenshots you can see that the differences from the total net profits is equal the difference from the comissions. so test 400 had a total net profit from -3452,2 (AAPL) and comission from 800. the test 10 show you a total net profit from -2672,4 with comission from 20. given that it was until now logical, that the comission is reflected in the total profit. if it is not, i can´t understand the difference which are similar to the difference in commisions.??

in my opinion it is strange that the profit per month from the asset TASR is at the test with commission 400 per trade 32,44% and at the test with only 10 $ it is 32,40 %. The difference between them is not very high but in my opinion it is strange that the strategy with a higher comission yield a better rate than the strategy with the lower costs?? DO you know what i meen??? Maybe i´m crazy or i´m wrong in my ideas. I don´t know.
nice greetings sigrid

[NinjaTrader_Jason]

Hello Sigrid,

For example, the Total PnL for the RIO in the left screenshot:

Profit: 184.25
Loss: -8299.25
Total PnL: -8115

184.25 - 8299.25 = -8115.

You see commissions are left out of the equasion.

Looking at the differences in the "Total Net Profit" it indicates not the exact same backtest is generated. Make sure both backtests use the same strategy, histoircal data and other backtest parameters if you like to generate same results.

[schlumpfine75]

i took the same strategy. only the comission was different. how can you explain that the difference from the two results (totalnetprofit 10 comission and totalnetprofit 400 comission-) is equal the difference from the comissions?? in my opnion is this an indicator for using the same startegy. i dont understand this phenomen?? if you say that the total net profit is the difference from profit and loss, is than there (in profit and loss) the comission reflected??? SOrry that i´m a dummy, but i need to understand it because i write my masterthesis about this. Where is in your opinion the comisssion reflcetd? nice greetings sigrid

[NinjaTrader_VinceB]

Hi,

Jason will get back to you with his answer tomorrow.

[NinjaTrader_Jason]

Correct, the commissions are reflected in the Loss and Profit.

[schlumpfine75]

Ji Jason!

Thanks for your answers. But there are still questions and i wanted to know if it´s possible to call you and speak to you on the telephone?? I can´t understand how it is possible when i take the same strategy and once costs from 10 $ per trade and once 400 $ per trade, that the result from profit per month (included comission 400 $ per trade) could be higher than the profit per month from the strategy with a comission from 10 $ per trade?? I calculated the profit per month for my own with the formula from the Helpsite, but i came to another result. Which correct formula will be taken by the cumulated profit and by the profit per month? What happend with the commisions? How will they reflected there?? The thing with the cumulated profit is negative while the profit per month is positive i can´t understand. If you take a look to the formula at the attachment you will see that it is impossible? Nice greetings Sigrid

[NinjaTrader_Bertrand]

Hi schlumpfine75, Jason will follow up with you on Monday.

[NinjaTrader_Jason]

Hello Sigrid,

Can you please send me screenshots of the Summary-tab of the two backtest results to support[at]ninjatrader[dot]com. Please reference this thread and put 'Attn Jason' in the subject header.

Make sure both backtests use the exact same parameters and historical data.

Commissions are reflected in the Gross Profit and Gross Loss. Cumulated Profit = Gross Profit - Gross Loss.

Profit per month = cumulated profit / amount of months.

If you would perform two identical backtests, one with a higher commission rate, it should result in a lower Cumulated Profit and a lower Profit per Month.