Statistics Question

S

shorty_mcshort

Say I have a strategy that trades the Emini S&P 500 swing trading (very short term swing). The average profit is 3.4 points and the standard deviation is 5.9 points of all trades combined (all winners and losers). If I subtract these two numbers I obviously get a negative number. What does this tell me?

Does it mean that I have less than 66% that in the future this strategy is a loser?

Eric
 
Hi Eric,

Tells you very little. If it wasn't a negative number you would have found the holy grail or (more likely) you have over-fitted your system.

You might want to look at the higher order moments (skew and kurtosis) to determine the shape of your distribution. That will give you more clues about the returns it generates. If you don't know what i mean you should look it up. Every trader should know about the basics of distributions.

If you want to figure out if the system is a keeper. You might want to run some monte carlo sims (can be done in excel if you don't have the tools), or apply Ralph Vinces excellent Risk of Ruin.

Two caveats:

1. How big is your sample - less than about 30 trades forget the stats. Get some more samples.

2. How did develop your system and generate your backtesting results? Very hard to avoid over-fitting every for experienced trades.

Hey good luck.

bolter
 
I can think of at least two ways to do the Monte Carlo:
1. If you have n trades, take n random draws with replication from your returns/trade, take the mean, repeat 10,000 times and plot the distribution of means. Order the means from low-high, take #250 as a lower bound on your future return
2. Instead of drawing from the returns/trade, randomly draw from the historical series of inputs (e.g. closing prices) to your strategy. Run your backtest, take the average return/trade, repeat 10,000 times etc.
 
how do you do #1 and what are you looking for to determine that it's a winner?

i guess same question for #2.
Quote from NoWorries:

I can think of at least two ways to do the Monte Carlo:
1. If you have n trades, take n random draws with replication from your returns/trade, take the mean, repeat 10,000 times and plot the distribution of means. Order the means from low-high, take #250 as a lower bound on your future return
2. Instead of drawing from the returns/trade, randomly draw from the historical series of inputs (e.g. closing prices) to your strategy. Run your backtest, take the average return/trade, repeat 10,000 times etc.
More...
 
Quote from bolter:

Hi Eric,

Tells you very little. If it wasn't a negative number you would have found the holy grail or (more likely) you have over-fitted your system.

You might want to look at the higher order moments (skew and kurtosis) to determine the shape of your distribution. That will give you more clues about the returns it generates. If you don't know what i mean you should look it up. Every trader should know about the basics of distributions.

If you want to figure out if the system is a keeper. You might want to run some monte carlo sims (can be done in excel if you don't have the tools), or apply Ralph Vinces excellent Risk of Ruin.

Two caveats:

1. How big is your sample - less than about 30 trades forget the stats. Get some more samples.

2. How did develop your system and generate your backtesting results? Very hard to avoid over-fitting every for experienced trades.

Hey good luck.

bolter
More...


bolter,

thanks for the info. I plan on doing some more research in this area. Should I assume that by reading books on LTCM that I should not just assume my distributions will be a bell curve? In other words it may or may not have fat tails or some other type of distribution.

1. my sample size is over 150 trades
2. i did my backtesting in excel so there was no optimization done at all. Nothing fancy just observing the market and manually going in excel and seeing how it would test out.

Do you recommend any books on this subject that is easily understood from a beginners perspective?

Eric
 
Quote from m4a1:

how do you do #1 and what are you looking for to determine that it's a winner?

i guess same question for #2.
More...

#1: You can do this is easily with any programming language or statistical package. Excel could work as well.
#2: I think there must be off-the-shelf backtesting software that allows this kind of monte carlo simulation. You can also re-program your strategy using a general programming language.
 
have you done it in Excel? how do you go about building a distribution based on your trades? then run a monte-carlo based on that distribution?
 
Quote from NoWorries:

I can think of at least two ways to do the Monte Carlo:
1. If you have n trades, take n random draws with replication from your returns/trade, take the mean, repeat 10,000 times and plot the distribution of means. Order the means from low-high, take #250 as a lower bound on your future return
2. Instead of drawing from the returns/trade, randomly draw from the historical series of inputs (e.g. closing prices) to your strategy. Run your backtest, take the average return/trade, repeat 10,000 times etc.
More...

Would love to know if this is possible in excel and if so how?
 
You must log in or register to reply here.
Back
Top