Re: Optimal f code for Tradestation Part II

[Dennis Holverstott <dennis@xxxxxxxxxx>]

Thanks to Bob Fulks for the informative posts. 

I did a little testing of my own on various pyramiding strategies. I
tried several methods of aggressive leverage to try to turn $10K into
$1M in 3 years trading the e-mini. I picked a system that is, long-term,
quite profitable but can have some nasty drawdowns. One such drawdown
was soon after trading started. It uses an ATR based disaster stop so I
used that stop value to determine risk and position sizing. I used quite
pessimistic slippage and commission numbers but ignored margin
requirements.

First try: full optimal_f, risk just over 26% of the account on each
trade, adjust position size up or down as account size and stop size
change. It started trading 7 contracts which quickly went down to 1
contract in the drawdown. At the bottom, the drawdown was over 90%. With
margin requirements, you never would have been able to trade it this
way. Ignoring that however, it recovered and, at the end of the 3 years,
was at $1.1M, well off its peak of $1.4M. It was trading huge size at
the end which is also unrealistic. The Sharpe ratio for the run was a
poor 0.86.

Second try: same as above but back off the leverage to the minimum still
able to achieve the $1M goal. That turned out to be at a risk of just
over 15% or about 60% of optimal_f. It started with 4 contracts. This
one was also off a former equity high at the end but it wasn't as bad as
the full optimal_f. Max drawdown was 60% and the Sharp had improved to
2.1.

Third try: start much smaller but, once size increases, never let it go
smaller. The risk to achieve the $1M turned out to be just under 9% of
the account. It started trading 2 contracts. Max drawdown was 45% and it
ended the test at a new equity high. Sharpe had increased to 3.2, the
best of the group.

Conclusion - about what I had found on previous tests. The notion of
trading big and decreasing size during a drawdown is flawed. By almost
every measure, it's better to trade small enough to survive the
inevitable drawdowns and, when your account grows, conservatively
increase the size and leave it bigger. If you decrease size after a
loser and increase after a winner, you are trading big on the losers and
small on the winners, just the reverse of what you would like to do. The
test also showed that the notion that you can't go broke if you keep
trading smaller on losers is flawed unless you are starting with really
big size. You can't trade less than one contract so that's the lower
limit of how small you can go. As well, you quickly run into margin
limits.

-- 
  Dennis