 |
A2ONE
Registered: May 2012
Posts: 6 |
05-22-12 11:05 PM
How should I evaluate my Sharpe ratio - using annual or monthly returns? I use NinjaTrader (which calculates based on monthly data), and I have incredibly low Sharpe ratios. I don't know if I should refine a strategy b/c this ratio is not very strong, and I know it is vital.
For instance, my main strategy has a (non-optimized) backtested Sharpe ratio of 0.55 in NinjaTrader, but these other statistics:
267% cumulative profit in 2.5 years, 30% max drawdown, 0.60% per trade.
Or for another strategy, over 12 months: 301% cumulative return, 8% max drawdown, 1.8% per trade, but only a 1.20 Sharpe ratio.
Should I only be concerned about the annual Sharpe ratios, which would be roughly 1.9 and 4.2, respectively?
|
| |
|
Edit/Delete • Quote • Complain |
jelite
Registered: Nov 2009
Posts: 35 |
05-23-12 02:24 PM
Annualized Sharpe >= 2 is very respectable in my opinion. I would focus more on return/drawdown ratio, which in your case (IF it's valid...) is high because that tells you how much you can make given a certain risk threshold.
|
| |
|
Edit/Delete • Quote • Complain |
Stoxtrader
Registered: Jul 2006
Posts: 228 |
05-23-12 04:30 PM
Quote from A2ONE:
How should I evaluate my Sharpe ratio - using annual or monthly returns? I use NinjaTrader (which calculates based on monthly data), and I have incredibly low Sharpe ratios. I don't know if I should refine a strategy b/c this ratio is not very strong, and I know it is vital.
For instance, my main strategy has a (non-optimized) backtested Sharpe ratio of 0.55 in NinjaTrader, but these other statistics:
267% cumulative profit in 2.5 years, 30% max drawdown, 0.60% per trade.
Or for another strategy, over 12 months: 301% cumulative return, 8% max drawdown, 1.8% per trade, but only a 1.20 Sharpe ratio.
Should I only be concerned about the annual Sharpe ratios, which would be roughly 1.9 and 4.2, respectively?
A rule of thumb is to have a sample size of at least 30, in order for the backtest to be somewhat predictive of future results. So monthly returns for 2.5 years (30 months) suffices, whereas monthly returns for 1 year (12 months) does not. If both strategies were for 2.5 years or more, they could be compared more easily.
|
| |
|
Edit/Delete • Quote • Complain |
gmst
Registered: Jul 2011
Posts: 3696 |
05-27-12 06:36 AM
Quote from A2ONE:
How should I evaluate my Sharpe ratio - using annual or monthly returns? I use NinjaTrader (which calculates based on monthly data), and I have incredibly low Sharpe ratios. I don't know if I should refine a strategy b/c this ratio is not very strong, and I know it is vital.
For instance, my main strategy has a (non-optimized) backtested Sharpe ratio of 0.55 in NinjaTrader, but these other statistics:
267% cumulative profit in 2.5 years, 30% max drawdown, 0.60% per trade.
Or for another strategy, over 12 months: 301% cumulative return, 8% max drawdown, 1.8% per trade, but only a 1.20 Sharpe ratio.
Should I only be concerned about the annual Sharpe ratios, which would be roughly 1.9 and 4.2, respectively?
Ninja and tradestation have screwed up here. Sharpe ratios are always quoted in annualized terms. so 1.9 and 4.2 are what you quote - which are excellent numbers btw. Good luck.
|
| |
|
Edit/Delete • Quote • Complain |
A2ONE
Registered: May 2012
Posts: 6 |
05-28-12 02:47 PM
Ok, that is what I was hoping to hear. Thanks for the replies.
Note, these are only backtested results, so I'm not going head over heals until the results are real.
One-
|
| |
|
Edit/Delete • Quote • Complain |
| Receive
an email whenever a new post is added to this thread by subscribing
to it. |
|
|
|
|
