|
PureBytes Links
Trading Reference Links
|
Chuck wrote:
> << We had a very nice run-up last year which took
> us from a one lot to a nine lot. Then the meltdown occurred and
> the system exceeded its historical max DD by a factor of 2. >>
> My point exactly. What would your drawdown have been if you were only
> trading a one-lot (or maybe two) prior to the meltdown?
The drawdown is undoubtedly worse. But what's the end effect,
assuming you don't let the drawdown kill you? I.e. are you worse off
right after the drawdown, and after the drawdown period ends, if you
pyramid the contract size?
I did an experiment. I created a spreadsheet that generated trades
for a hypothetical system that were evenly distributed between -$1250
and $3750 ($5000 offset by $1250, a 75% chance of a winning trade,
average trade is $1250). I did this for 30 trades. Then for 10
trades I offset the trades from -$3750 to $1250, a 25% chance of a
winning trade with an average trade of -$1250, giving us our
"meltdown" period. Then I resumed the original 75% winning pattern
for 20 trades.
I started the account with a $100k balance, and computed the running
account balance for an account that traded 5 contract throughout, and
another that started out with 5 and added an additional contract for
each $20k in account equity. This is fairly aggressive leverage,
since it means we're risking 5*$1250 = $6250 or 6.25% of the account
on the initial trades. So you'd expect some serious drawdowns.
In almost every case, the pyramided account came out FAR better.
Even if you bailed at the very bottom of the drawdown with the
leveraged version, you were usually STILL ahead of where you'd have
been if you traded without pyramiding. Some cases bottomed out
slightly below the non-pyramided account, but not by much. I'll post
a sample GIF in another message.
If you trade through the meltdown, the pyramided account digs itself
out of its drawdown within 1 or 2 trades of when the non-pyramided
account does, and rapidly leaves it far behind. In fact since it
STARTED the drawdown at a much higher level, it's usually well ahead
of the non-pyramided account even at the bottom of the drawdown.
Even in the cases when its more-severe drawdown makes it dip below
the conservative account during the meltdown, it surpasses the non-
pyramided account's absolute value within a few trades after the end
of the meltdown period.
So it seems to me that IF you are willing to stand a worse drawdown,
you are far better off in the long run to increase your size. And
even if you bail at the worst possible time, you're still usually
better off.
This assumes, of course, that you don't hit such a bad drawdown that
you bankrupt yourself. But that's pretty hard unless you're
insaneley over-leveraged. I extended my "meltdown" period to TWENTY
trades, and even with the 6.25% leverage I seldom saw the equity even
go below the non-pyramided account, let alone go negative. But oh
man did I end up with some amazing account balances -- for the
pyramided account, that is.
Furthermore, the "pyramid forever" scenario isn't really realistic.
I think most traders WITHDRAW FUNDS from their accounts periodically.
In that case the leveraged example will be far ahead, giving you far
more funds to draw on. Chances are you will have withdrawn a large
amount from your account before the meltdown hits -- quite possibly
more than the non-leveraged account contains WITHOUT withdrawals! --
and you'll still have a large account to build from. The
withdrawals, while providing you a steady stream of income, will also
reduce the degree of pyramiding in the account, and thus the severity
of the drawdown.
Now obviously these results will differ for every system, and for
every set of trade results. But I think it should be somewhat
representative of what you could expect. Given these results, I
don't see why anyone would want to trade strictly at a fixed bet size
-- unless you don't want to accept a constant level of risk, you're
just satisfied with the returns you get from that bet size, and you
don't want to increase them. If you want to increase your results,
and you're willing to accept a steady level of risk AND larger
drawdowns, then this test says fixed bet sizes are far better in
almost every case.
Gary
|