[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Tharp's Expectancy



PureBytes Links

Trading Reference Links

Paul wrote:
> I'm specifically looking at Tharp's definition of expectancy, where he 
> states:  (p.148)  "Expectancy is a way of comparing trading systems while 
> factoring out the effects of time, position sizing, and the fact that one 
> is trading various instruments that have different prices."

Sounds like a circular semantic game to me. See below.

Paul again:
> Using _his_ definition, I think I can imagine a negative expectancy game 
> that would make money, since you could conceivably have a position sizing 
> algorithm that would "know" when to enormously increase your bet size on 
> winning trades, even if those winning %'s were very small.

Assuming you could do that (big assumption), you are changing to a new
system. Since position sizing is an integral part of your new system,
the definition above which implies trading the same size on every trade
is irrelevant.

Phil wrote:
> If you know which trades to load up on, then you should reprogram the system
> to take those trades and skip the rest. And then you should be able to
> demonstrate a "positive" expectation w/o regard to size..

Exactly. You could make millions flipping a coin if you only bet on the
winners. :-)

Paul again:
> The "knowing" which bets to leverage is not absurd.  There are many folks 
> who load up on what they consider to be higher probability bets.

Well, maybe. Like Phil said, if they can *really* do that consistently,
they should just stand aside on the low probability trades and never
have a loser. Personally, I'm a pretty good contrary indicator. The
trades I really "love" often turn out to be losers and the trades I
really "hate" often turn out to be big winners. Maybe it's just me. :-)

Paul again:
> Sorry, but I still don't see it.  Let's do a thought experiment.  We look 
> at a hypothetical system that's had 100 trades historically.  50 of them 
> returned 5%, 50 of them returned -10%.  Obviously, a negative expectancy 
> based on V. Tharp's definition (we've backed out and normalized the 
> position sizing, by examining only the return).  But what if you had had an 
> algorithm that somehow "knew" to leverage your winning bets by 4/1.  Now we 
> have a very profitable game, but Tharp's "expectancy" is still negative.

5% of what and 10% of what? I'd say account size is the most relevant
number. If you bet 4 times as much on the winners, you are increasing
your account 20%, not 5%, and the expectancy is positive. Maybe not by
the definition above but who cares? We define words so we all know what
we are talking about. If a particular word defined in a particular way
doesn't apply, choose another one and don't get hung up on definitions.

-- 
  Dennis