Value of Backtesting and Stops

[waxwing]

Quote from pepper_john:

two peak market: exit after the second peak.

I'm sorry, I don't understand, please explain in more detail? Which peak is the second peak? And to be clear, are we looking at a histogram of *trade returns* or histogram of price movements in the market?

positive autocorrelation: in a down trending market, exit whenever the market is lower, as its will be even lower the next period

Oh, OK, I thought we were talking about autocorrelation in the series of returns from the trading system, but you're talking about, again, the time series of price movements itself. Well, that is trivial

 

[waxwing]

Quote from science_trader:

Of course you would need paths in a perfect world. Usually pdf and autocorrelations of strategy returns will be fine.

I'm just saying that in my experience, setting stops to "cut out" the part of the distribution where a lot of losses are, often doesn't work at all. But I guess you have a lot more experience than me

Tell me, have you often encountered clearly bimodal distributions of returns in your trading strategies?

 

[science_trader]

Quote from pepper_john:

what about the max draw down?

No negative month.

 

[science_trader]

Quote from waxwing:

I'm sorry, I don't understand, please explain in more detail? Which peak is the second peak? And to be clear, are we looking at a histogram of *trade returns* or histogram of price movements in the market?


histogram of trade returns.


Oh, OK, I thought we were talking about autocorrelation in the series of returns from the trading system, but you're talking about, again, the time series of price movements itself. Well, that is trivial

The same for autocorrelations. On my side I mean autocorrelations of the trade returns.

 

[science_trader]

Quote from waxwing:


Tell me, have you often encountered clearly bimodal distributions of returns in your trading strategies?

In 2 strategies I have never put into practice (too low returns). To say it fast, this was the kind of strategy that can work perfectly or be completely wrong.

 

[Holmes]

Quote from pepper_john:

Vow. Yearly return of more than 200% with Sharpe ratio of 5. Is this for single-strategy or a portfolio of strategies?

Yearly? Is that all you can muster?

Monthly would be more appropriate to get me interested.

Livermore did 150% weekly!

:eek:

Holmes

 

[waxwing]

Quote from science_trader:

The same for autocorrelations. On my side I mean autocorrelations of the trade returns.

Yes, I understood this. My previous posts were a reply to pepper john, who said the opposite.

But then back to my original question - why do autocorrelations give value to stops? If I see a clear autocorrelation in my trade returns, I would add a rule about position sizing dependent on previous returns. If done right, this should both increase the profitability of the strategy and remove the autocorrelation.
How do stops come in to that?

 

[waxwing]

Quote from taowave:

For example if we have a histogram of returns with one peak at, let's say, +$1000 and one at -$1000, it doesn't follow that by setting a stop somewhere around, let's say, -$500, we will increase the expectancy of the system, as far as I can see - the price paths are hidden in histograms

which is why you need a histogram that discovers the path.....

I like to graph MAE and MFE relative to bars since entry.

Very good point, thank you for mentioning it. Sorry I didn't notice your post before.

I'm going to go and update my backtesting code now

 

[science_trader]

Quote from Holmes:

Yearly? Is that all you can muster?

Monthly would be more appropriate to get me interested.

Livermore did 150% weekly!

:eek:

Holmes

If you could lend me some money at a good rate...

 

[cqg_support]

Without surfing through 17 pages of posts, I'll say this, experience has taught me 2 things.

1. Pigs are Cute and Hogs get slaughtered!!!!!!!!!

2. Any trade system worth trading will employ multiple exit strategies including stops and % and $ risk trailing, so as your profits increase, you can limit your downside and protect against sharp market movement against you.

CQG has a backtesting suite in the Professional Version of our software, complete with the above mentioned stops pre programmed, all you need to do is enter the % or $ figure.

Hope it helps....

Good Luck