Improving Sharpe Ratio

[Delta007]

Does anyone have any recommendations to increase the Sharpe Ratio, is there specific factors I should be focusing in on such as altering the entry or adjusting the exit parameters or using smaller targets tighter stop losses etc.? My sharpe ratios are pretty low, .5 to .85 area. I'm using a multi-instrument strategy, 25 different momentum stocks. Backtesting on 1minute over 2yrs. This is an intraday strategy that always exits before the end of the day.

[happypappy]

Reduce your down draw, essentially it needs to be more accurate during entries to increase Sharpe....Apparently a ratio of 3 is good, I dunno how people manage to hit that level!

[Delta007]

Reduce your draw down, easier said than done. Probably doesn't help that I'm using very volatile stocks.

[happypappy]

Reduce your draw down, easier said than done. Probably doesn't help that I'm using very volatile stocks.

If you getting a bit of draw down, but the strategy is detecting correct entries per say. ie you go long at 90, it drops to 85 then heads to 95. You should be able to avg out and make adjustments using MAE?

After 20 or so entries, take the MAE and divide it by 2, then instead of entering move the entry target down......So strategy would enter @ 90, MAE is 5, divide it by 2 then enter at 87.5.

The longer you run it the less your MAE gets the more accurate your entries should get......in theory :O

[Delta007]

Interesting theory. Say your entry signal is based on a simple macd cross and you figure out you should be entering $1 below it according to the MAE test how can you make it enter $1 lower than the cross. You'd have to somehow get it to store the cross value which could get complicated.

[happypappy]

Off top of my head so not sure the code actually right but you get the idea....

As an example, in the code, the location where you currently have EnterLong(), you could replace it with another function and replace EnterLong() with mynewentry(Close[0]).....So trigger occurred, now enter a limit entry when the price drops to a certain point.....I think that'll work....

private void mynewentry(double CurPrice)
{
EnterLongStopLimitOrder(CurPrice - MAE Figure, CurPrice - MAE Figure,"new name");
}

There is more work outside of this, you need to control how long you keep this order open and or capture the Close[0] to a input variable.....So it doesn't adjust the value each time....Bit tricky but it should help you on the right track....

http://www.ninjatrader.com/support/helpGuides/nt7/index.html?performance_data_tabs.htm

Checkout the IOrder stuff as well, this will help you gather entry information which can help you control your entries and cancel the entry if you no longer want it active...

[CableTrader007]

Either increase returns or lower volatility, i.e standard deviation of returns. As long as the returns are there smaller stops and limits will decrease the standard deviation of returns. Also, look at combining strategies or adding more stocks (running same strategy over more instruments) has positive effects on sharpe.

[Delta007]

Thx happyhappy ill look into that. And CableTrader if you use stocks what type of stop/targets do you typically use and do you use ATR? On the particular strategy I'm working on now, using stocks under $150, through optimization a $2 stop and $3.5 target has given best results. Which I realize is quite loose. I've tried to use all kinds of trails anywhere from 50cents in profit to $2 or just using a trail such as low of 15min bar etc. but they just seem to negatively effect it.

[CableTrader007]

Delta (like the 007 in the name), Sharpe is all about performance over variation and looking for consistent performance without large swings. Another thought could be optimizing based on R^2 of the equity curve. Either performance must go up or standard deviation of returns must go down to improve Sharpe. I trade futures and forex so no stocks experience. What I look for is smooth equity curves.