NT8061- Strategy Performance Reporting doesn't work 100% with multiple strategies

[sledge]

I ran 2 strategies that took the SAME EXACT TRADES on the SAME EXACT DAY.

1 of the strategy performance results is simply just wrong, and reminds me of the reporting back in the BETA days that was supposedly corrected.

Avg MAE/MFE/ETD/%Trade?, are wrong just like previously - because the one strategy spits out a bad chart of the bad computations.

Examine: chart with same trades.

Control Center -> Strategies->right click strategy->Strategy Performance->Real Time

See attached images.

This was reproduced on different machines in NT8061.



I knew something was up when 1 strategy was displaying this type of chart which was very close to all the issues in BETA when I get the < graph.

https://www.screencast.com/t/jWYUSG7E1

While the other chart looked more correct and random:

https://www.screencast.com/t/c7qALJ7CH7Hr

And both strategies have about 95% of the same trades, so it should match except for a few spots, not the entire thing.

Maybe this will fix the Efficiency charts for me also.




Also of note: The trade performance report is wrong also - since it is sourcing the same incorrect values and averaging them together: (correct value #1 + incorrect value #2)/2

[NinjaTrader_Jim]

Hello sledge,

Thanks for opening the thread.

The math on the Strategy Performance Analysis is consistent with what is outlined in the help guide. MAE can be calculated with the information available in the window, as with ETD. We would need to know the Entry price for MFE to verify this value.

Could you provide screenshots of the Strategy Performance window with Trades chosen in the Display Selector?

I referenced the equations for these excursion and drawdown values here: https://ninjatrader.com/support/help...efinitions.htm

I look forward to assisting further.

[sledge]

Hello sledge,

Thanks for opening the thread.

The math on the Strategy Performance Analysis is consistent with what is outlined in the help guide. MAE can be calculated with the information available in the window, as with ETD. We would need to know the Entry price for MFE to verify this value.

Could you provide screenshots of the Strategy Performance window with Trades chosen in the Display Selector?

I referenced the equations for these excursion and drawdown values here: https://ninjatrader.com/support/help...efinitions.htm

I look forward to assisting further.

Say what? Reread my post again.

I posted images.

They should be the same
One is incorrect.

[NinjaTrader_Jim]

Hello sledge,

We will need to see the trades to see how these values were calculated from exactly.

Just by looking at the Analysis, we see one trade with a loss of 100. This could equate to an MAE of 100. Also, if the ETD can be calculated as MFE - Average Trade, then 200 could be a valid value. I use 'could' because we do not have the trade history that would note how these values got calculated with entry and exit prices.

We will need to verify the trades to troubleshoot further.

I look forward to your reply.

[sledge]

Hello sledge,

We will need to see the trades to see how these values were calculated from exactly.

Just by looking at the Analysis, we see one trade with a loss of 100. This could equate to an MAE of 100. Also, if the ETD can be calculated as MFE - Average Trade, then 200 could be a valid value. I use 'could' because we do not have the trade history that would note how these values got calculated with entry and exit prices.

We will need to verify the trades to troubleshoot further.

I look forward to your reply.

Maybe I wasn't completely clear in my original post.

2 strategies produced the same exact trade on the chart(4 @ xxxx.xx), and the same exact exit (4 @ yyyy.yy)

Therefore they both would have the same MAE, etc when looked at in the strategy performance.

They do not.

If I run the same strategy by itself, it produces the correct result.

I'll write up a simple test sample when I get a chance.

These reports have been generally wrong from the start of the entire beta. It's been frustrating to say the least.

[NinjaTrader_Jim]

Hello sledge,

I agree that same data set + same logic should equal the same results. If something has gone wrong, we will need to find out where it is going wrong. As the Trade Performance is based on the trades, we will need to make sure NinjaTrader has consistent information to calculate consistent results. I.E. Are the trades improperly recorded for Account Performance?

I've done some tests on my end with the SampleOnOrderUpdate strategy on a 5 second data series with the simulated data feed, and I can't create diverging results.

NinjaTrader 8 Release 7 just went live, so we should also make sure that this issue is actual in the latest build.

I would greatly appreciate a sample script if necessary and any steps I can take to reproduce and get the issue reported.

[sledge]

I would greatly appreciate a sample script if necessary and any steps I can take to reproduce and get the issue reported.

NT8070 sample:


Run strategy1 and strategy3 on the same ES chart.

I ran on 5/1/2017 and it produced different results in MAE/MFE/etc section.

Note: EnterLong @ 10am EST and ExitLong at 3:00pm EST appeared to work on the first trade. When I switch to EnterShort/ExitShort - bingo.

You may adjust the times. And it might also be related to losing short trades only.

Also to be clear, I added these to the same chart. In my original test - I had 2 separate ES charts. I have not tested adding to strategy tab in control center only.

[NinjaTrader_Jim]

Thanks sledge,

I've been able to reproduce and the issue has been reported.

EDIT: Our QA team has reviewed the matter and the ticket has been designated as a bug with the new ticket ID NTEIGHT-11839.

You can reference this ticket ID in the release notes of the version of NinjaTrader 8 that includes the fix. Release Notes can be found here: https://ninjatrader.com/support/help...ease_notes.htm

If you have a question about the status of this issue before a fix gets released, feel free to bump this thread and ask about the status of the ticket number.

Thanks again for the report!