Cummulative Profit

[tradetree]

I have two strategies, call them A & B, and they are applied to a chart so I can run the "Strategy Performance" from the chart. They both have the same instrument, same start & end time and are both enabled. A has a cummulative profit of 231%, B has a cummulative profit of 165%. Changing the format tab from % to currency, A has a cummulative profit of $63460, B has a cummulative profit of $63290. How does this make any sense? They have radically different % gains, but almost identical $ gains! This is mathematically impossible. I was trying to figure if this could be due to compounding, but it can't, a percentage gain should still show up in the dollar gain.

[NinjaTrader_JoshP]

Please see here for how cumulative profit is calculated: http://www.ninjatrader.com/support/h...efinitions.htm

The cumulative profit of each strategy is dependent on the trades that particular strategy makes. They are completely independent of each other and each strategy has its own cumulative profit which cannot be compared to another strategy's cumulative profit on a currency vs percentage basis like what you have done.

[tradetree]

Why can't you make the comparison? Here is the definition you gave:
cumulative profit
This statistic returns a value representing a summation of all the profits earned by all your trades. It can be interpreted as a performance measure for your strategy.

Currency
SUM(profit * quantity * point value) of all trades
Percent
PRODUCT(1 + profit / entry price) of all trades - 1
Points
SUM(profit * quantity) of all trades

where profit is defined as (exit price – entry price) for long trades and as (entry price – exit price) for short trades, quantity is defined as the number of contracts traded, and point value is defined as the monetary conversion of each point (e.g. 100 for currency pairs).


In my case, the quantity = 1 for both. The point value = 100 and is the same for both. That means the only unique item in this equation is the profit of all trades. Since the profit of A = profit of B according to the output, that means the SUM( profit * quantity * point value ) for A = B.



For the percent equation, all terms are the same for A and B, except for the entry price. The only variable is entry price ( A ) vs. entry price ( B ). Yet they are trading over the same period and on the same instrument, so the entry prices should correlate.



The only way to prove that your calculations are correct (or incorrect) is to export the trades for each strategy to a spreadsheet. Is there a way to do this?

[NinjaTrader_Bertrand]

tradetree, you can right click in the performance reporting grid and export them to Excel to compare.

Is the trade sequence and trades # exactly the same for both strategy? Are they both run on the exact same charts including session templates so you're making a viable comparison here?

Thanks,

[tradetree]

The trade sequence is the only variable. If that makes the difference it will be counter-intuitive but I'll get to the bottom of it either way. I exported to excel and will implement the equation and report the results. If it matches NT I'll try to understand how it makes sense. Statistics can be very confusing, but they are supposed to give you something humanly useful. If two strategies give a very different percent gain, but the same dollar gain, then that is not that useful unless it is telling another story that I don't yet understand. For example, it may be saying that risk is much lower in one than the other for the same dollar gain.

[tradetree]

I verified the results and I match NT in excel. Now I'm trying to interpret what these statistics mean intuitively. I believe less money is being used for the strategy with the higher % gain, even though the ending $ gain is the same for the two strategies. So it could be viewed as risk/reward is better for the higher % gain. On the other hand, being that I am highly leveraged, a smooth equity curve with more trades is better, so the $ gain is more important. Does NT have any reference material for statistical trade-offs?

[NinjaTrader_JoshP]

tradetree,

Unfortunately we do not have material to read up on how to interpret the statistics. Your description below is pretty much dead on accurate though.

[tradetree]

Josh,
I have been doing a lot of research on automated strategies and statistics and I'd like to submit an idea for NT. I've used "Profit Factor" and "Win/Loss ratio" and "Sharpe ratio", but these all have serious pitfalls. There are a couple of other ratios that are far better and just as easy to calculate as the Sharpe. The "Ulcer index" goes well beyond Sharpe in terms of giving a statistic that traders can actually use. Sharpe does not include the effect of draw-down on account impact. The "Ulcer index" does. Here are a couple of links to learn more:

http://www.tangotools.com/ui/ui.htm

http://www.financialpicks.com/systems_101.html

Right now I have to export my data to excel just to calculate this index. It would be great to have it as part of the Strategy Performance statistics and optimization/back-test statistics.

[NinjaTrader_Austin]

Tradetree, thank you for the suggestion.

[Camdo]

Thanks TradeTree.
I was wondering myself what Cumulative Profit Percent was good for.
Perhaps the Ulcer Index could be programed as a method and executed in OnTermination.

[tradetree]

Do you mean just create our own method and submit it, or wait for NT to do it?