|
Robpf
 Member
Posts: 13
Joined: 10/11/2012
Location: Pittsburgh, PA
User Profile |
I have been experimenting with single strategy runs to try and determine which strategies work best in different market states. I have done this by allocating trades only in one market state and enabling only one strategy. This appears to work well and trades are only taken in that market state. I went back and added a volatility trade ranking in hopes of improving this performance. When I do this trades are taken in all market states to spite the fact that there are no trade allocations in the other market states. I clearly don't understand how Trade Ranking is working in the flow of determining trades.
|
|
Jim Dean
 Elite
   Posts: 1059
Joined: 10/11/2012
Location: L'ville, GA
User Profile |
Sounds like a bug to me.
My understanding of how it’s supposed to work is the same as yours.
Compliments on the thoroughness of your testing.
|
|
Mark Holstius
 Elite
    Posts: 744
Joined: 10/11/2012
Location: Sleepy Hollow, IL
User Profile |
Hmmm
I've used the same process, and haven't had any problems with including a ranking function.
Can you expand on what you mean by no trades allocated(?) in the other market states? Do you have the % equity set to zero for the other states?
Also, in those states that you don't want to see trades, I've found that I still need to have 1 strategy selected.
Can you post a snag with more details?
Mark
|
|
Robpf
 Member
Posts: 13
Joined: 10/11/2012
Location: Pittsburgh, PA
User Profile |
Under Allocation I left Percentage of Equity set to a default of 10 for all market states but under allocation ratios I put a default of 10 for max longs of one market state only. All of the other max longs and max shorts were set to default of 0. I thought this would limit trades to only that market state with max longs allocated
|
|
Robpf
 Member
Posts: 13
Joined: 10/11/2012
Location: Pittsburgh, PA
User Profile |
I just tried it again on another machine and did not have the same problem. Something must not be compiling right. On the second machine there was no change with the addition of the market ranking. I will look more into later tonight. Thanks
|
|
Mark Holstius
 Elite
    Posts: 744
Joined: 10/11/2012
Location: Sleepy Hollow, IL
User Profile |
If anyone is wondering what process Robpf and I are referring to as a way to select strategies by their performance in a particular Market State, I've attached a pdf file with some basic instructions on how to do it...
It's a good way to see which strategies work well in each Market State, but it is a bit tedious.
Mark
Attached file : Choosing Strategy By Market State.pdf (1030KB - 427 downloads)
|
|
Harry G
 Member
 Posts: 35
Joined: 10/11/2012
User Profile |
Thank you...that was very helpful!
Harry
|
|
Robpf
 Member
Posts: 13
Joined: 10/11/2012
Location: Pittsburgh, PA
User Profile |
I can't explain what happened yesterday. It is a very complex program and I suspect there are bugs that only appear after multiple runs. I think it is a good idea to just randomly restart the program. I am making some progress in lowering drawdowns. The single strategy approach allows you to see effects of individual changes in more granularity. I agree it is tedious. Since I am looking a long only trading for IRAs it is particularly challenging to limit drawdowns in down periods. My latest insight is to limit the number of trades and use ranking to select only the best. The total returns will be less but the return to drawdown improves. I am currently using only 1 long trade on the single strategy testing and getting better results and I suspect that even with multiple strategies a small number of trades will be optimal. I am not big on optimization but this is one area I will definitely optimize when I get there.
|
|
Mark Holstius
 Elite
    Posts: 744
Joined: 10/11/2012
Location: Sleepy Hollow, IL
User Profile |
FYI: Be sure to update OT2018 to PR2I that came out today.
I found that it can make a big (usually positive) difference if you uncheck "Close All Trades when leaving this market".
Mark
|