My system results... what do you think?

[Rickshaw Man]

Quote from maos:

Morgan, thank you on your reply, it was just the kind I was hoping for.



I believe that , at least in this context, by results you mean profit or %return...

Profit figures are one of the things I pay least respect to because I believe that they are influenced by so many different uncontrollable elements that they are all but unpredictable in any way.
During development I focused on measurments like %win, payoff /winloss ratio, %drawdown as well as robustness of overall strategy in the terms of keeping those figures as constant as possible over different stocks and market cycles.


I am aware of that, but what I wanted to accomplish when I finished developing this strategy is to see how it could cope with everything I could ''throw'' at her in terms different price data, changing market conditions, cycles, volatility and so on

I designed this strategy with all that in mind, trying to accomplish that the strategy checks the history of each stock it trades(on a year by year basis) and adjusts to that particular stock and it's characteristics. Which, I believe, is why it performs well over any portfolio i test it...So, I'm not entirely sure that I understand the effects the absence of entirely independent tests could have on my system performance.

I know for certain that what the overall market does affects the performance of my strategy, because it forms an index of stocks in a portfolio it trades and blocks certain trades in respect to what that index does...considering that every portfolio has 130+ stocks, the index that is formed resembles the market index(in some cases closely)...here again pops out the question of independent tests...suggestions?



Yeah, that is the main problem. I initially planned on developing three systems that would later be combined into one master system: one for rising markets(long), one for declining(short), and one for sideways markets(long)...the first two are finished and I plan on developing the third as soon as i get started trading what I finished because I can't afford to wait for six more months...

So, to answer your question, I believe(hope) I know the reason. The market stopped rising and went into a kind of sideways mode after an uptrend which is the ugliest period for the strategy at this stage.
Long strategy is still signaling trades( that after entering just keep falling...it isn't the pullback the strategy belives it is, but down move during sideways action which tend to be larger ) because the downtrend hasn't really started, which is the reason the short part of the strategy isn't making enough money to account for the losing long trades...the part where the sideways part of the strategy would fit in nicely.


tried already... both break down completely...




Systematic pullback buyers out there are welcome to share their experiences...

Your priority should be to minimize the losses in this directionless market.

Is their anyway you can test your strategy from 1965-1980, that may help you.

 

[boxster332]

Quote from man:

sharpe ratio of 1.3 and profit factor of 1.5 are relatively low IMHO, since the 150 stock universe is probably subject to several ex-post factors. i would recommend you paper trade it until you have forty trades and compare the two distributions of results.

some poster mentioned slippage of 1.7% - i admire everybody who can build a trading system with average trading length below two months on that transaction cost. hats off.

peace

I also agree that the SR of 1.3 is too low. Most people look for huge return figures but miss out on the fact that lower returns but higher SR's can be leveraged up to the same drawdown levels to achieve higher returns. Major investment banks require a minimum SR of 2 for short term prop trading when testing. Besides, low volatiltiy makes life much easier day to day.

 

[maos]

Quote from boxster332:

I also agree that the SR of 1.3 is too low. Most people look for huge return figures but miss out on the fact that lower returns but higher SR's can be leveraged up to the same drawdown levels to achieve higher returns. Major investment banks require a minimum SR of 2 for short term prop trading when testing. Besides, low volatility makes life much easier day to day.

boxter,

Since I'm still couple of days away from trading this system, I'm toying with different position sizes, with which I managed to raise the SR to slightly over 1.5

I'm aware that the value over 2 would be a good thing, but considering the nature of my system I believe that raising it further would be difficult without changing the entire exit logic. (The system chooses stop and target levels completely autonomous from me based on volatility, which greatly affects the variability of returns )

 

[science_trader]

I haven't read all the posts on this thread, but what I can say looking at your equity curve is that your system has worked very well for a short period of time and that all the rest is random. I wouldn't bet a cent on that one.

 

[horribilicus]

Maybe you will want to download the free code for the two systems "Nosewhistle" and "Earwax", which are over on the Turtle Trading Software bboard. These two systems are programmed for Wealth-Lab, just like your system.

You can run all three of them on bunches of different portfolios with plenty of different position-sizing approaches, and see how well your new system stacks up against these two freebies in a variety of situations.

The source code and the various Wealth Lab output statistics are shown on this web page http://turtletradingsoftware.com/fo...s=0&postorder=asc&highlight=nosewhis*&start=0

 

[man]

Quote from science_trader:

... looking at your equity curve ...

actually, after having spend quite some time with quantifying outputQuality, i more and more think that no fancy ratio can beat an experienced look at an equity curve.

 

[science_trader]

Quote from man:

actually, after having spend quite some time with quantifying outputQuality, i more and more think that no fancy ratio can beat an experienced look at an equity curve.

It's all about statistics : you reduce a set of data to a single statistic (mean, standard deviation, ...) and loose information at each step. Looking at the whole set of data can tell you far more than looking at a single statistic.

In finance, it is even worse as the set of data are very often unstationnary.

 

[Rickshaw Man]

Quote from horribilicus:

Maybe you will want to download the free code for the two systems "Nosewhistle" and "Earwax", which are over on the Turtle Trading Software bboard. These two systems are programmed for Wealth-Lab, just like your system.

You can run all three of them on bunches of different portfolios with plenty of different position-sizing approaches, and see how well your new system stacks up against these two freebies in a variety of situations.

The source code and the various Wealth Lab output statistics are shown on this web page http://turtletradingsoftware.com/fo...s=0&postorder=asc&highlight=nosewhis*&start=0

Seems you ran him off, maybe his strategy was really one of the two you mentioned.

 

[man]

stock systems with a sharpe ratio around 1 on a data set which is not cleaned against survivorship bias are problematic. i believe that you finally much more directly "feel" sharpe ratio rather than profit factor once you are trading ...

 

[maos]

Quote from Rickshaw Man:

Seems you ran him off, maybe his strategy was really one of the two you mentioned.

Still here...

I'm not posting because there is no need, all I wanted were some opinions...which I got, also I answered all of the questions that were directed to me.

P.S. If you take at least a rudimentary look at the strategies horribilicus mentioned you will see there is no similarity.