Greetings Ramon, Graham, and all -- You are probably already aware, but many will not be.
There is danger in multistep optimization when the second step is applied to the results of the first step, all of the out-of-sampleness has probably been used up. Be certain to reserve some additional data for a truly out-of-sample test. Thanks, Howard On Mon, Oct 19, 2009 at 9:23 AM, ramoncummins <ramoncumm...@hotmail.com>wrote: > > > I asked the same question recently, custom backtester code for r-squared of > the equity curve can be found here: > > http://finance.dir.groups.yahoo.com/group/amibroker/message/141718 > > cheers > > Ramon > > > --- In amibroker@yahoogroups.com <amibroker%40yahoogroups.com>, > "grahamj42" <graham.john...@...> wrote: > > > > Not sure if this is what you are after, but this one of my current > projects. > > > > 1. define variable PassNum - normally set to 1 > > 2. using CBT, define/create system specific equity curve > > 3. run backtest > > 4. set PassNum = 2 > > 5. using PassNum & Foreign, apply an indicator to the equity curve and > use this as en entry filter. As a simple approach I have been playing with > ROC. > > 6. run backtest > > > > It is a bit klunky, but it works. > > > > Graham > > > > > > > > --- In amibroker@yahoogroups.com <amibroker%40yahoogroups.com>, DIANE > TONETTI <ftonetti@> wrote: > > > > > > > > > > > > KRatio is measurment of the smoothness of the Equity curve which works > > > fine for systems that dont compound ... > > > > > > For systems that do you could calculate your own KRatio based on the > log > > > of equity > > > > > > On Wed, Oct 14, 2009 at 8:36 AM, droskill wrote: > > > > > > I agree it would be an interesting choice - only issue is that, as > > > far as I can see, the backtester does not allow that kind of targeting. > > > > > > --- In amibro...@yahoogrou ps.com > > > <mailto:amibroker@yahoogroups.com<amibroker%40yahoogroups.com>> > , > > > "woodshedder_ blogspot" <woodshedder_ blogspot@ ...> wrote: > > > > > > > > Droskill, > > > > What about R squared? > > > > --- In amibro...@yahoogrou ps.com > > > > <mailto:amibroker@yahoogroups.com<amibroker%40yahoogroups.com>> > , > > > > "droskill" <droskill@> wrote: > > > >> > > > >> Hey all - > > > >> One of the things I would love to optimize around is equity curve > > > >> smoothness - but I don't see any parameter that really matches with > > > >> that. Drawdown is an obvious one as a smooth equity curve generally > > > >> doesn't have huge drawdowns - but I'm wondering if people have other > > > > >> ideas. > > > >> Thanks in advance! > > > > > > > > > > <mailto:amibroker@yahoogroups.com <amibroker%40yahoogroups.com>> > > > > > > < > http://groups.yahoo.com/start;_ylc=X3oDMTJvNm8wcWYyBF9TAzk3MzU5NzE0BF9wAzMEZ3JwSWQDMTAxMDY5MgRncnBzcElkAzE3MDU2MzIxOTgEc2VjA25jbW9kBHNsawNncm91cHMyBHN0aW1lAzEyNTU1MjM4MDQ- > > > > > > > > > >
