How to compare strategies that trade infrequently / frequently (and select one of them)

R

rvsw

I need inputs from members on how to compare and select trading strategies for implementation from a pool of strategies.
I have 2 trading strategies which are in play almost throughout the year and in the back testing, provide an annualised return of about 12%.
I have another 3 trading strategy is matured in play infrequently ( less than a total of 60 trading days in a year) and therefore provide an annualised return which is about 3%. However, if we just consider the return over the trading period, it is as high as 30%
since I have limited capital,

1) I have to decide whether I should allocate capital entirely for the 2 trading strategies that are in play almost throughout the year or always OR
2) allocate part capital to the will to trading strategy is and then have capital always available at hand for3 trading strategies that are infrequently implemented .

the broader question is that how do we compare trading strategies that have different frequencies of trading and therefore, very different annualised returns even if the rate of return for trade remains the same across the trading strategies that are being compared.

I realise that there may not be a cut and dried answer to this so I'm looking for mainly best practices owhat criteria do people cconsider while selecting trading strategies for implementation

 
Your question is certainly valid, however your data are not. You need many observations more (hundreds and thousands) than you have, and especially so over multiple market environments. (A record-setting bull market skews towards trend-following, eh. o_O)

Your current focus on return ignores risk. Entirely?? :confused: Dunno.
But there are plenty of reward-to-risk measures out there for every flavor.

For you and your limited capital, however: you need to focus on how far away from profit any trade you make, can go. "Maximum Adverse Excursion" and "Maximum Favorable Excursion" MAE and MFE need to be your short-term friends.

When you get more numbers (more executed trades, over months, seasons, markets and years), the others will make more sense to use.
 
You might laugh, and since you used "days" I am assuming signals comes off daily charts, but if stock been around 50 years, I will test over all of them plus 5000 different symbols going as far back to 50 years. Yep great deal of work, and fun to see the "what ifs" during Bull markets, which I deem less values. What I am interested in is major reversal points and how much drawdowns from the extremes. Am sure many traders might not test their holdings since they bought low, but I think in terms of equity curve of open positions.

Small sample size is like building your home on edge of a dirt cliff, enough rain....and you don't know when. I been in your spot long ago, I have made the mountain of mistakes more than once, unless you can answer all the "what ifs", you have that house in California on dirt cliff.
 
Annualize their returns and risk metrics. Sharpe doesnt annualize nicely (mathematically) from daily for example, but you're not going for perfect you're just trying to get a number that sorta-makes-sense. You're a practitioner and not a scientist.

This puts each strategy on a level playing field for comparison. If the annualized return of a strategy is 3% but you're only making 60 trades a year your sharpe will also likely be significantly lower, as more profitable round trip trades necessarily increase the sharpe. In this case you may want to look at the profit factor of each. In these moments it's more opinion. If you do make 30% in 60 trades a year and your non-sharpe risk metrics are ok - why pick the other strategy? The logic doesnt make sense. Either your math is wrong, or I a missing something here.
 
Last edited:
rvsw said:
I need inputs from members on how to compare and select trading strategies for implementation from a pool of strategies.
I have 2 trading strategies which are in play almost throughout the year and in the back testing, provide an annualised return of about 12%.
I have another 3 trading strategy is matured in play infrequently ( less than a total of 60 trading days in a year) and therefore provide an annualised return which is about 3%. However, if we just consider the return over the trading period, it is as high as 30%
since I have limited capital,

1) I have to decide whether I should allocate capital entirely for the 2 trading strategies that are in play almost throughout the year or always OR
2) allocate part capital to the will to trading strategy is and then have capital always available at hand for3 trading strategies that are infrequently implemented .

the broader question is that how do we compare trading strategies that have different frequencies of trading and therefore, very different annualised returns even if the rate of return for trade remains the same across the trading strategies that are being compared.

I realise that there may not be a cut and dried answer to this so I'm looking for mainly best practices owhat criteria do people cconsider while selecting trading strategies for implementation
More...
@tommcginnis gave you excellent advice. Make sure you read his post carefully.

Question: Have you computed the expectancy and win rate of each strategy?

My approach (for strategy with positive expectancy) is counter intuitive:

1. Strategy that has low probability but very high payoff I trade frequently, but each trade is only fractional Kelly.

2. Strategy that has high probability but small payoff I trade infrequently but each trade tends to be at least one Kelly.
 
rvsw said:
I need inputs from members on how to compare and select trading strategies for implementation from a pool of strategies.
I have 2 trading strategies which are in play almost throughout the year and in the back testing, provide an annualised return of about 12%.
I have another 3 trading strategy is matured in play infrequently ( less than a total of 60 trading days in a year) and therefore provide an annualised return which is about 3%. However, if we just consider the return over the trading period, it is as high as 30%
since I have limited capital,

1) I have to decide whether I should allocate capital entirely for the 2 trading strategies that are in play almost throughout the year or always OR
2) allocate part capital to the will to trading strategy is and then have capital always available at hand for3 trading strategies that are infrequently implemented .

the broader question is that how do we compare trading strategies that have different frequencies of trading and therefore, very different annualised returns even if the rate of return for trade remains the same across the trading strategies that are being compared.

I realise that there may not be a cut and dried answer to this so I'm looking for mainly best practices owhat criteria do people cconsider while selecting trading strategies for implementation
More...

Before you go into analyzing, one would need more information.

1. What market did you test on (stocks, forex,...).
2. If stocks did you consider survivorship bias?
3. If Futures, did you consider rollover?
4. Did you consider worst case scenarios, based on your money management settings?
5. Did you consider volume?

Many more questions but those are the basics.
 
I also think that you need to rethink your profitability calculations. At the end of the day it is not at all important how many days you are trading, but how much profit you can make on your invested capital.
 
You must log in or register to reply here.
Back
Top