Trade Management for 2 Trading Schools (Continuation Vs Mean-Reverting:)

[palm]

Hi All,
I am now trading continuation systems and would like to add a mean-reverting system to compliment my portfolio. I compare the returns distribution of the two systems below. (pls see attachment)


The red curve is for MR while the green one is for CON systems. Both have positive expectations. You can see that the big chunk of returns is cluttered at small gains level for MR. My question is on exit strategies. Given these two systems, how would you design exit strategies differently.
Exit strategies issues are
1.) Stop-loss
-(Fix or volatility-based)
-(Tight or loose)
2.) Profit-target
-To have or not to have
-Tight-loose
3.) Time-Exit
-To have or not to have
Anyone has experience dealing with MR type systems?

Any other thoughts are welcomed. Thank you

 

[Adamus]

That's an unusual question. Framing the question in the context of a returns distribution shows immediately that the continuation system is not doing what I would thought a continuation system would do. Maybe I've got it wrong but I assume a continuation system is a trend following system? In that case I would expect the upper end of the distribution to be fatter than the MR system. Maybe you mean just a break-out system. In either case, how do you exit now without any stop loss?

It would also be useful if you posted the systems' win:loss ratio or average win and average loss, and the % winners.

 

[palm]

That's an unusual question. Framing the question in the context of a returns distribution shows immediately that the continuation system is not doing what I would thought a continuation system would do. Maybe I've got it wrong but I assume a continuation system is a trend following system? In that case I would expect the upper end of the distribution to be fatter than the MR system. Maybe you mean just a break-out system. In either case, how do you exit now without any stop loss?

It would also be useful if you posted the systems' win:loss ratio or average win and average loss, and the % winners.

Hi,
Thanks for your response.
I'll try to clarify my questions.
1.) I group breakout and trend-following under the same type (Continuation of price). Is there a fundamental difference between the two?
2.) The CON system is a Breakout system with Time-Exit. In fact, both system have the same fixed holding periods.
3.) The right tail of the CON system is in fact fatter. (shows in the graph)

WIN/LOSS Ratio
CON system: 45%
MR system: 51%
Profit Factor
CON: 1.46
MR: 1.45

 

[Adamus]

Before you get your hopes up, I'd better just say that the answers to your questions are the kind of answers that I've also been looking for in several years of writing mechanical systems.

I suspect that one of the reasons that I haven't got answers to those questions is that it would require a huge amount of work sifting through the profit and loss history of the system under a magnifying glass. First of all you'd need a lot of trades. I think hundreds at least, better would be thousands. Then you'd need to put them in a database or a spreadsheet and examine what percentage of the trades resulted in what different type of result - and I'm not just talking about final profit or loss.

Obviously a returns distribution is going to tell you something, and you don't have to use a graph all the time, you can look at the mean, the median, the standard deviation, the kurtosis, or stats like the Sortino ratio or the Sharpe ratio.

There are also several statistics available (at least in the analysis packages I've used) above and beyond the final return and the returns distribution.

There is Maximum Adverse Excursion - the furthest a trade goes against you before exiting.

Maximum Favorable Excursion - the amount a trade moves in your favor before exiting.

End Trade Drawdown - how much you give back after reaching the maximum favorable excursion

Despite having those statistics available for NinjaTrader which I use, the way they are calculated is not so good. It's just the blanket average for all trades. I want to see it for different levels, i.e. what does it look like for the really profitable trades vs the not-so-profitable vs the losses.

This in my book is the kind of info you need to work out whether the exit methods you are thinking of applying are actually doing what you thought, without any unintended or surprising side-effects.

 

[amnonco]

Adamus is right,
there's something wrong with these distributions. usually momentum systems (i.e trend following) has left fat tails, and mean reverting systems has right fat tails.
however this has a lot of variation due to specifics in your systems.
e.g you can have an intra day momentum systems with mean reverting like distributions...
so it all depends on the specifics of your systems.


to answer some of your questions, in general,
the use of a stop loss and/or a take profit will always have pros and cons (naturally stop loss cuts your large losses but negates some of those profits as well, and take profit vice versa).

what I usually do is this:
for momentum trades:
if its a short term trade it is usually better to use take profit, use stop loss that is a bit longer than your MAE.
if its long term use some sort of volatility stop or trailing stop. usually no other stop is needed but you have to manage position size accordingly.
for mean reverting trades:
if it's short term - again - i use a take profit targets. I use a wide stop + averaging under certain conditions in these systems.
longer term mean reversion systems to me is somewhat of a contradiction, so I don't use them.

about time exits - they are pretty much the best exits if you can't decide which exit to use. so if there is no fundamentally "correct" exit to the system, one that has something to do with the underlying concepts of the system, use a time stop.

another note:
if we look at your distributions, then your 2 systems are pretty much the same. meaning if you use an X% stop the "avoid these" will be gone and accordingly the "don't count on these" will also be gone.

so if the expectancy is positive in the "near 0" zone, you'd have a positive expectancy low variance system.

good luck

 

[meanreversion]

Hi All,
I am now trading continuation systems and would like to add a mean-reverting system to compliment my portfolio.

Anyone has experience dealing with MR type systems?

Any other thoughts are welcomed.

I've been researching mean reverting systems to add to the continuation (trend) system which I'm currently trading.

The three metrics I tend to look at to evaluate systems are Sharpe, Sortino (more applicable with a breakout system) and return/drawdown.

If I add two systems together, are the risk adjusted returns superior to either of the two components .. if they are, then there's a case for trading the two alongside each other.

Can you provide us with some more analysis, i.e. return/drawdown ratio for your trend system, and same again for your mean reversion system?

 

[palm]

Thank you all for your comments.
The reason I compare returns distribution of the 2 is that I want to understand how 2 trading systems with exact opposite philosophies both works.

Here are the systems that produce the returns distribution you saw above.

For Momentum system, long if closing price is above 10 days high. Hold for 5 days. No stops. Use volume filter to ensure tradeable signals.

For MR, long if closing price is below RSI(14) value of 20. Hold for 5 days. No stops. Only volume filter is applied.


I try to use the most basic system setup with fewest parameters. The distributions you saw above are based on the same holding period.

I reasoned that those guys who trade different systems (CON and MR) must use different type of exit techniques for them to both make money.

Update on my MR system testing. I test stop-loss/profit-target based on short term volatility and it tested out that 1:4 ratio of stops/profits work out best.

The system doesn’t fly very well on out sample data.