Profit % doesn't look correct

[tsharon]

I have an intraday strategy that buys and sells stocks that i am trying to backtest.....i run the backtest and have the parameters set to sell (close the long position) at either a 1.00% gain or after a certain number of bars (time has elapsed).....i ran the back tester and then looked at the Strategy Analyzer and looked at my list of Trades.....I did the backtest on the stock QCOM.....The first trade it listed was this:

EntryLong 5/17/06 9:30 am Price=46.39 Contracts = 100
ExitLong 5/17/06 9:35 am Price =46.85 Contracts = 100

under the Profit column it says i made $46.39
Under the Profit (%) column it says .05%

My question is about the .05% profit %????

If you do the math....buy at 46.39 and sell at 46.85.....the return is 1.00% not .05%.......can somebody explain to me why the Strategy Analyzer has the wrong value????

[eSignal_JasonK]

Hello tsharon,

First, please tell me what eSignal version and build you have (Help-->About).



If you are not using this version, please upgrade and try your back test again.

If you are using the most current version, please let me know what your current settings are (click on the Settings icon at the top of the Strategy Analyzer).

[tsharon]

I upgraded to the newest version (version 8.0 build 782)....and it still shows a .05% return....my settings are the same as the one you showed me except fot 2 fields...

My "initial Capital" is 100000
my "degree of risk aversion of investor" is conservative....

is there a way for me to send you the display of the trades from the Stregegy Analyzer so you can see it first hand?

[tsharon]

so what is the next step in figuring this out???

[eSignal_JasonK]

Hello tsharon,

I need to test your formula code on my end. Please post your formula using the attachment feature when replying on-line.

[tsharon]

Fomula

Ok,....here is my formula.....i set the percent below BB to 2% and the Goal Percent gain to 1.00% and used qcom over last 120 days intraday data......

[eSignal_JasonK]

Hello tsharon,

The profit % for each trade is the dollar amount of profit as a percentage of the account equity rather than the % profit based on the entry and exit price of the trade.

[tsharon]

is that the number i should be looking at to evaluate how successful the formula is??? Or is it better to look at the % profit/loss based on the entry/exit prices of the trades? if it is better.....is there a value somewhere that shows the % profit/loss based on entry/exit prices of the trade.

[tsharon]

can you please look at my last post and tell me what you think?

[demarcog]

Pardon my jumping in but as someone who has spent more time then I care to mention considering this issue, let me suggest there have been some fairly recent updates in the knowledge base on the Backtester which you may want to check out.

The shortest answer is look at the Equity Curve, if it looks like a perfect uptrend with minimal pullbacks (drawdowns), then it may be promising. Of course the larger the amount of data backtested, and the more trades the better. The guide suggest's 30 trades as the mininum required for statistical significance.

Analyzing the success of any trading is somewhat more complicated then one would think. Years ago, I would look at "Net Profit", the higher the number the better, right! Maybe.

When I would start trading the system the reality of the "drawdowns" become painfully obvious and make it extremely difficult to "stick"with the system. "past performance is no guarantee" and it's very easy to get shaken out after a series of losses, just when the system begins a series of successful trades, which I would possibly miss.

A system can only be successful (percent profitable) 50 percent of the time, but the average win is double the average loss and be a "good" system, so the higher the percent profitable the better, but again, several small losses, followed by bigger gains is fine.

I like to set the "account size" to however amount of capital is required to purchse the instrument, non-margined, to get an idea of the return on the investment. net profit/capital required. Obviously return on investment is the "industry standard" for measurement of any investment.

Be careful of over optimization, or curve fitting, the smaller the number of variables, the better. If you test 1000 moving average combinations and find 2 winners, you can be sure that it's doubtful that it will perform as well in the future.

Don't forget about slippage and commission! Overestimate slippage, because guaranteed you will get bad fills even on the most liquid instruments, and believeing that a "limit order" is going to get filled at the bid if buying or offer if selling means you are really going to miss alot of the best trades. Usually getting a good fill price, as someone who has traded alot, is NOT a good sign.

Also there are alot of issues with the backtester, review the doc thoroughly, becasue depending on your logic and selection of MARKET-THISBAR-NEXTBAR etc, the results can be inaccurate.

Then there is conversion to realtime, need to remove all the strategy.golong, goshort, etc, and test it during live trading hours. I'm seeing as I write this major differences in my results, due to "forward looking" conditions, ie., make money when tested, but not in actuality.

I like your code, I actually have a few BB related strategies with a few twists that I received taking a currency trading class which made money trading realtime in the class but I haven't gotten round to programming in EFS, if you would like to share the load, maybe we could work together.

Good luck, and feel free to contact me at [email protected], or through the board, although I don't sign in on a regular basis.

Glen