> sometimes a proper way to plan a new feature is to look at other >who already > have travelled the same path: this way create a point of >reference from > which we may discover we can do better .
Agreed....if it is a discussion, amongst users, about the range of portfolio approaches that we are taking etc and hopefully providing Tomasz with some useful feedback. >it allows to test both approach hicks was referring to: the "one >cash pool" >and the "assign x% of equity to each system") and made >absolutely no >reference to MPT, Vince or whatsoever. If you want to implement it yourself then I believe it is possible in AFL i.e. I would say that I could implement your version and Hicksies version, in AFL, without a mention of Vince, or MPT, except for the fact someone might call me on it and I really have a lot on my plate so I am not keen to go off and do a speculator. I think Paul Ho also said he can implement it in AFL? That is the only point I am disagreeing with you on -- that it can't be done in AFL. --- In [email protected], "ang_60" <ima_c...@...> wrote: > > --- In [email protected], "brian_z111" <brian_z111@> wrote: > > > > > > However, the point is .... whether or not it can be done depends on the > > model you are using, and the complexity of that model.... how many systems > > you want to trade and on how many symbols and whether you are in RT or > > daily bars etc. > > > > > Brian, > > sometimes a proper way to plan a new feature is to look at other who already > have travelled the same path: this way create a point of reference from which > we may discover we can do better or . simply transport their scheme into AB > (ideas have no copyright). > > I've already done this: although I like Amibroker the most, and I would like > it to became my one-stop software for everything I do in the trading world, I > know at least 4 retail products that allow built-in multisystem testing > (price varying from 410 usd$ to 3000 $), and I own three of them. > > So, I can honestly say that your point above is just your respectable point > of view but not at all an absolute necessity. > > In my humble opinion, the best possible multisystem testing engine is already > marketed, it is the one leaving the most flexibility to the user (for example > it allows to test both approach hicks was referring to: the "one cash pool" > and the "assign x% of equity to each system") and made absolutely no > reference to MPT, Vince or whatsoever. > > Greetings, > > Angelo >
