[R-SIG-Finance] Mathematical Expectation for a trading system

[Mark Knecht]

On Fri, Oct 16, 2009 at 8:00 AM, Ulrich Staudinger
<ustaudinger at gmail.com> wrote:
> Hi,
>
> On Fri, Oct 16, 2009 at 4:34 PM, Mark Knecht <markknecht at gmail.com> wrote:
>>
>> Not enough information. Is this a long-only system? If it is then you
>> bought 4 and sold 7.
>>
>> Is this a system that goes long and short? If so then I cannot tell
>> what completes a trade?
>
> the trade system in this case could have been any long/short trade system.
> I think the winners/loosers ratio is only applicable to a limited amount  of
> trade systems and does therefore provide only limited confidence about the
> expectation of trade systems.

Higher math - you
Lower math - me


>
> Other ratios like LiquidationPnl / Trades make more sense, or for example, a
> curve marked-to-market values of the position once a trade occurs.

Possibly. I don't disagree. Can't when my level of understanding is so low.

One comment I'd make is that I wouldn't include trades that are in
process as they aren't historical. Personally I look only at closed
trades when evaluating my systems historically.

>
> Another aspect is also the definition of a trade, i define, and i think
> that's also the mainstream definition a trade as a transaction. What you
> call a trade completion is something i know as a roundturn (buy and sell).
> The term round turns is  not applicable to pyramidizing systems, i think.
>

I don't disagree, but once we've matched buys to sells and thrown away
everything that's in process maybe the overall average expectation ($
Made/#trades) isn't any different but the largest winner, largest
losers are?

I do a lot of modeling on paper where I use what I call the 6 Tharp
numbers and generate trade sequences automatically to study things
like drawdowns, risk or ruin, etc. I think they are applicable to
pyramiding systems but as I said originally

Higher math - you
Lower math - me

;-)

>
> Kind regards,
> Ulrich
>
>
>
>
> --
> Ulrich B. Staudinger
>