Unreal "strategy", was: How to interpret backtest report

Dear All,

One of my friend forwarded me his strategy back test, it looks promising, please check attachment and provide your suggestions. Code is not provided, he is asking money for the same.BAcktest%20report Charts_Page_1 Charts_Page_2 Charts_Page_3

I wouldn’t even consider buying a strategy without real transaction history.

I could go into Amibroker and put 3 variables in, hit optimize and come up with the worlds best strategy.

Usually if my report looked like that, I’d be checking my code for some forward looking bias. Very simple to do and code check won’t catch all forward looking code.

6 Likes

Isn't it worth asking why he doesn't run the strategy in the market instead and make far more money than selling it you or a bunch of people?

3 Likes

Those results are screaming "coding mistake", not "holy grail".

5 Likes

For original poster, here is a free of charge "holy grail" code that will generate backtest results much better than what you posted. It will generate 100% winning trades and zero drawdown. It has just one disadvantage - it is TOTALLY USELESS

// USELESS "I am winning all the time" code
Buy = Sell = 1;
// these two lines make it winning all the time without any drawdown :-)  
BuyPrice = Low; 
SellPrice = High;

Bottom line: do NOT buy and do NOT trust ANY black boxes ("systems" without code)

Backtest is useful technique to verify your ideas for trading. But you need to know that formula used is correct one.

8 Likes

Thanks Sir,
Still I want your comment on results itself.

Did you mean that if code is OK then it is a good strategy?

You have been given comments - it's a "Holy Grail" code and it has been highlighted to you that they don't exist! Do you really believe the CAR on that report is realistic and achievable in the real world?

No, in simple words, that strategy is fake. You did NOT see the code yet it becomes quite obvious that the strategy is fake with >400% CAR with 2% exposure. Average % profit is 0.19%. Really? This means that 0.19% slippage kills it. This screams u-n-r-e-a-l, fantasy world.

It's time to learn one simple fact in life - there are NO free lunches. No-one is going to sell you good system. If the system was really working this guy would invest his or borrowed $1M and quadruple it in a year. But he decides to sell it. Really? How ridiculous is that.

If you want good system you have to find/develop YOUR OWN.

15 Likes

Resp. Sir,
Great findings in this discussion. I personally attended trading session sitting beside, it is real time no future reading but without automation, order putting is Manual and delayed.

You caught it perfect, 0.19 % is slippage which is killing the strategy. Hence he might be selling the same. gr8 sharing. Very thankful to all of you.

@dranokar,
If you sat beside him while he traded the system and you observed the results, then buy it. Seriously. Hire a couple of the ace programmers on this board to perfect the automation then sell it to a hedge fund manager for $50,000,000.
If that strategy actually works it is worth billions in the real world. But therein lies the problem...

1 Like