Definition of MAE, MFE, ETD

[PLeary]

On the Trade Performance Report, it shows 'Average MAE', Average MFE', and 'Average ETD'.

What do these abbreviations mean?

[FMtrader]

On the Trade Performance Report, it shows 'Average MAE', Average MFE', and 'Average ETD'.

What do these abbreviations mean?

Go to the NT8 help guide here: http://ninjatrader.com/support/helpGuides/nt8/en-us/

Then go to "Trade Performance" then to "Statistics Definitions" for detailed descriptions

Basically:
Maximum Adverse Excursion
Maximum Favorable Excursion
End Trade Drawdown

[NinjaTrader_PatrickG]

FMtrader is correct.

Here is a direct link to that portion of the help guide:

http://ninjatrader.com/support/helpGuides/nt8/en-us/statistics_definitions.htm

[PLeary]

Thanks!

these are very useful stats, glad I know what they mean now!

[GuppyDRV]

Just a quick thought........is there a way to figure out on which bar since entry the MAE, MFE happened?

GuppyDRV

[koganam]

Just a quick thought........is there a way to figure out on which bar since entry the MAE, MFE happened?

GuppyDRV

int ReferenceBars = BarsSinceEntry(<parameters>) //replace "<parameters>" with the correct parameters for which [I]BarsSinceEntry[/I] you want to query.

//Assume that we are long (or just exited a long). then:
int MAEbarsAgo = LowestBar(Low, ReferenceBars);
int MFEbarsAgo = Highestbar(High, ReferenceBars);

//It would be the opposite for shorts.

You now know how many barsAgo the events occurred, so you can calculate each relevant barNumber if you want, by subtracting said value(s) from CurrentBar.

ref: http://ninjatrader.com/support/helpG.../lowestbar.htm

http://ninjatrader.com/support/helpGuides/nt7/highestbar.htm

[Dolfan]

OK, so I've run a strategy analyzer and I feel comfortable with the definitions provided by the forum as directed here but, when viewing the MAE/MFE/ETD data I find some puzzling results. When reviewing in tick mode the MFE and ETD tick values are identical but when viewing in currency mode they are not. Why is that? Is there some calculation that I need to account for? Here is a screen shot for reference. https://www.screencast.com/t/TCGWBzlVk Thanks in advance.

Best regards,

Dolfan

[NinjaTrader_PatrickG]

OK, so I've run a strategy analyzer and I feel comfortable with the definitions provided by the forum as directed here but, when viewing the MAE/MFE/ETD data I find some puzzling results. When reviewing in tick mode the MFE and ETD tick values are identical but when viewing in currency mode they are not. Why is that? Is there some calculation that I need to account for? Here is a screen shot for reference. https://www.screencast.com/t/TCGWBzlVk Thanks in advance.

The formula for Average ETD is:

• Average MFE – Average Trade

I assume if you look at your report in ticks mode you would see that the "Average Trade" value is zero. According to the formula for Average ETD, I would expect the MFE and ETD to be the same value when the "Average Trade" is zero.

If you look at your report in currency mode I suspect that "Average Trade" has a value which is then subtracted from the MFE to give you a currency which is different than the MFE.

You can find the definitions and formulas here:

http://ninjatrader.com/support/helpGuides/nt8/en-us/statistics_definitions.htm

[Dolfan]

I read the definitions and I feel reasonably sure that I understand them. So riddle me this, why is my "Average Trade" zero in tick mode but not in currency mode? Note that the analysis showed 3229 total trades.

Where am I going with this? I read a comment somewhere else that one person was looking at a trade's performance and it was suggested to adjust his profit target to take full advantage of the MFE number. I hope to adjust my strategy in testing for multiple instruments by doing the same thing. It is confusing and misleading when so many numbers come back the same, when you know they are not.

Best regards,

Dolfan

[NinjaTrader_PatrickG]

I will need to do some extensive testing to check the math for the average trade calculations in the different modes. I will follow up with what I discover. Thanks for your patience.

[Dolfan]

Thanks Patrick. I checked the other settings like percentage, pips & points and it seems that only in currencies mode is there "meaningful" numbers, although that could change with whatever your findings are.

An entire spreadsheet awaits your return.

Dolfan

[NinjaTrader_PatrickG]

Thanks for bringing this to our attention, Dolfan.

After analysis it turns out that the formulas here needed adjustment to be accurate. We are tracking this internally as NTEIGHT-11330.

Please look for this number in an upcoming release of NinjaTrader 8. Here is a link to the release notes page:

http://ninjatrader.com/support/helpGuides/nt8/en-us/?release_notes.htm

[Dolfan]

I wait with breathless anticipation for the result. In reviewing the strategy analyzer results, I calculate that the supposed distance of the MAE (and possibly the MFA & ETD) are not nearly accurate. For example, where it shows 3-4 ticks MAE and a $$ figure that coincides, I review the tick chart to verify and it may move 10, 20, maybe 30 ticks. This data is completely inaccurate and therefore irrelevant and unusable. Are there other errors in the process of being fixed that I should know about? Is there a list somewhere to track?

Dolfan

[NinjaTrader_PatrickG]

The MAE and MFE are averages of all trades in the results. I would not expect it to reflect the literal max for any one trade.

Can you go into further detail about how you're calculating that the values are incorrect?

I am running this backtest - it has few trades and is not difficult to check the math manually. This is the backtest I ran to verify your original report. Could you run this backtest and tell me what math you're seeing as incorrect?

• Sample MA Crossover
• ES 03-17
• 5 minute
• 2/28/2017 - 3/02/2017
• CME US Index Futures RTH
• all other settings are default

Please be as detailed as possible.

[Dolfan]

OK, this is the second time trying to enter all of this. For the record, expired tokens are a ROYAL PAIN IN THE GAZZOOOOO! I hate typing a lengthy response only to have it COMPLETELY LOST because of an expired token. ARRRRRGGGGGGGGGGGG!!!!!!!!!!!

OK, breathing easy now....

I ran your analysis and I see your point but I get different results running my strategy analysis. Could it possibly be because I am using RenkoBars (9) on my chart? I can see where the numbers over a long period of time will be skewed by the amount of time and data points collected but for the numbers to be so low, even averaged out, the probability of a position moving against me is very very low, as suggested by the analysis. I think we are all smarter than that. Even with a success rate of 63%, the idea that 37% of all trades will, on average, move ONLY 2 ticks against me is unrealistic. https://www.screencast.com/t/ApVQApaBbU

I ran my strategy with identical settings over a long period of time, 01/01/2008-present, with the only difference being the size of the RenkoBars (?) per instrument. I exported the analysis and combined them all into one spreadsheet. You can see that the MAE is amazingly similar with the exception of SI. https://www.screencast.com/t/64SWnp34EdY
This suggests to me that the market is very reliable, and we know that isn't so. Since my entering into this thread was about how to use MAE for stops and MFE for profit targets, I submit once again that this data is not reliable as presented OR, as I am interpreting it.

On another note, when trying tun a Optimization, it appears as though this is only meant for strategies that use moving averages. Is this accurate or are there a list of indicators that you can test using optimization? Thanks!

Best regards,

Dolfan