We'd all love to be wow'd by your stellar performance. Why not offer up a sample of the kind of results you're talking about? rt
The point of this thread was to get guidance on what an acceptable drawdown was per return. Obviously people will have different views when it comes to this. Bottom line is you complained that mine isnt acceptable. Great. When asked why you responded with something along the lines of you have lots of great ideas that you arent willing to share and are above the rest of us in terms of intelligence (which you may be, I don't particularly care). Bottom line is you have nothing to add(or chose to add) to the discussion. I would love to add to my win/loss ratio and % winning trades but to do that I would need to find a way to further improve my strategy. At this point in time I am not educated enough to do that. . . . . . . . Whew. LOL. I just do not understand people's logic of coming into a forum with no intent of adding to the discussion.
Seriously. ET needs to implement some sort of user reputation system to increase the signal to noise ratio of quality posts. If there was a reputation system, I think users would think twice about posting a useless comment. This is the problem with the visible post count.
Back to RoughTrader... So basically I should do an optimized backtest of say, 2006-2007, than see how those parameters do from 2007-today? Or something of the like?
The basic idea of walk-forward optimization is this: Your historical data has N sessions (days) Choose a window of this data with K sessions, K < N. K is the in-sample data length. Choose a length of data J to be your out-of sample data. In total, K + J <= N The process is iterative. You start by optimizing the strategy over the first K sessions. Once the optimal parameters have been found, apply them to the historical data starting from session K+1 until K+J. Record the PNL for each session K+1 -> K+J. Slide the in-sample window forward by J sessions. Optimize, retrieve parameter values, then apply the values to obtain session PNL values for out-of-sample sessions K+J+1 -> K+2J. Your in-sample optimizations should be performed over data windows: 1 -> K 1+J -> K+J 1+2J -> K+2J 1+3J -> K+3J etc. etc. Your out-of-sample PNL values are extracted from the following windows: K+1 -> K+J K+1+J -> K+2J K+1+2J->K+3J k+1+3J->K+4J you continue this iterative process until you have run out of data (i.e., K >= N). Then the out-of-sample PNL values are sequentially stitched together to form a realistic, forward-looking sequence of trades that mimics how the system behaves with unseen price data. The selection of K and J values is a subject of debate. I personally choose to keep J relatively small (5 sessions, reoptimizing every Friday after close), and do a loose ratio optimization to find the best K value. If you have ANY parameters that the equity curve is sensitive to, IMO it is imperative that this analysis be done. Otherwise, you are just left with the result of optimization of historical data. Such a result is NOT forward-looking at all, and as such does not provide the information necessary to judge confidence in the system. rt
In my personal experience [based on the strategies I have implemented so far (may change in future) and talking about *intraday* only] the very best ratio: Avg Profit ----------------------- Max Drawdown is about 25-30%. Given a strategy, it is strongly dependent on the security. [High ratios can be achieved only with a very small number of securities, often ultra] eg,: Daily Avg : 239.86 Min Balance seen : -943.02 Ratio Avg/Drawdown % : 25.44% [It is very possible that other people have better strategies, which achieve higher ratios] PS Drawdrawns are part of strategies! ;-)
IMO there is an issue here, but the basic idea is right. You are using the minimum balance seen, which is a negative value. This redefines a drawdown to only be an equity dip below the initial seed capital. The correct definition of a drawdown is any equity dip below a most recent equity high. To calculate this you must profile the entire equity curve, identifying each new equity high and subsequent dip from the equity high. Then the largest dip from all these dips becomes, by definition, the max DD. I personally choose to use the ending equity to max DD ratio, which is essentially the same thing as using the average session profit to max DD, except scaled by the number of sessions. rt