PureBytes Links
Trading Reference Links
|
I think your missing the point. The gambling analogies were used to
illustrate the invalidity of certain statements claimed about statistics and
probability. I do not think they were used in comparison to the market or
to the success or failure of any particular trading system.
Jeff
----- Original Message -----
From: "Guy Tann" <grt@xxxxxxxxxxxx>
To: <metastock@xxxxxxxxxxxxx>
Sent: Monday, April 17, 2000 7:44 PM
Subject: RE: Money Management Stops
> Kevin
>
> I'm not sure whether I agree with any of the gambling analogies when used
> for comparison to trading the markets. My primary reason for this is that
> in gambling (roulette and craps are examples) your analogies are 100%
> correct, IMHO. When compared to blackjack, there are some variances based
> upon the play of players before and after you that impact the deck and the
> play.
>
> When compared to trading the market, I don't think these analogies are
> applicable at all. Again, that's just my opinion.
>
> I'll agree that it's possible to have 1,000 events go against you in a row
> and then have 1,000 go for you in a row. However, if in testing this
> system, you ever decided to trade it, you would be broke in no time at
all.
>
> If on the other hand, you develop a system that is capable of 80%
profitable
> trades, proper testing would give you some idea as to the distribution of
> these losing events.
>
> In our case, even when we develop a new weighting schema, we back test the
> system for 6 months, 5 years and 18 years. While we're primarily
interested
> in the last 3 years, we also need to see how the system would have done in
> the real world over a longer period of time. Rather than manufacturer a
> random list of numbers going back many years, we just use the actual
> market's performance over that longer period. In our case, again, large
> losses are exceedingly rare, thank goodness. Repetitive losses are also
> very rare. On average, for the past 15 years, we average no more than 6
or
> 7 losses a year while still maintaining our minimum profitability ratio of
> over 70%.
>
> If in back testing, we ever encounter repetitive losses of the type you
> describe, even while still maintaining good profitability, we would never,
> ever trade it. In all of our back testing, we take repetitive losses into
> consideration as well as the system's profitability. A system that wipes
> you out without giving you the opportunity to make a "come back" is hardly
a
> system.
>
> Regards,
>
> Guy
> Fax (630) 604-1589
>
> -----Original Message-----
> From: owner-metastock@xxxxxxxxxxxxx
[mailto:owner-metastock@xxxxxxxxxxxxx]On
> Behalf Of Kevin243@xxxxxxx
> Sent: Monday, April 17, 2000 4:53 PM
> To: metastock@xxxxxxxxxxxxx
> Subject: Re: Money Management Stops
>
> The answer is no.
>
> It is completely possible to have a 1000 events go against you in a row
and
> then have another 1000 go for you in a row.
>
> The point is that you could be wiped out before the reversion back to the
> mean.
>
> What you are describing is like doubling your bet each time you lose
> figuring
> eventually, you will win. Not true. You will eventually find a streak
that
> breaks your bank first. Play long enough, and you will eventually lose
> everything. The house is counting out that. The house has a positive
> expectancy, you the gambler does not.
>
> Kevin Campbell
>
> In a message dated 4/17/00 7:43:39 AM Central Daylight Time,
> Michel.Amelinckx@xxxxxxxxxx writes:
>
> > Thanks for helping me out here. I guess my explanation was not very
> > understandable. And the example of the roulette table was even worse
> > actually it was WRONG because of course you still have the 0, so this
was
> > WRONG.
> > What I just tried to say, like described here below is if you have a
> system
> > with 70% prob. (of course these statistics are as good as you test
them)
> and
> > your system shows 4 or more LOSES IN A ROW. And because of this it is
> > deviating from your mean and thus the prob. of the next trade being
> correct
> > is higher. (There is of course a change that your system stops working
at
> > that point)
> > Is this incorrect ?
> >
> > Greetings
> >
> > Mickey
> > B
> >
>
>
|