Evaluating Results - Position Sizing

[underdog430]

I'm testing a pivot based system on the SPY. The system does poorly when it is trading back and forth around the pivot and typically makes any profit on the last trade of the day. I was wondering if I could improve my results by changing the position sizing based on the number of trades.

For example, if I doubled the size of the first trade, then I would score big on the days when there is just one trade (about 9 in 60). Since individual trade losses tend to be small (but cumulatively significant) the associated increase in losses wouldn't be that great.

If that makes sense, then the idea of also doubling the second trade also makes sense. The higher the number or sequence of trades that you double, however, the more you get chopped up. Anyway, it looks to me like doubling the size of the first two trades makes sense. Please let me know if you see a flaw in my thinking, think these profits are too low for a viable system, or have any other thoughts. Thanks.


Backtesting past 60 days using a base of 100 shares of SPY. The number indicates which trades are doubled (0 = none, 1 = first, 2 = first and second, etc.).

# Profit (Change) Drawdown (Change)
0 1,430 269
1 1,709 (19.5%) 284 ( 5.5%)
2 2,302 (34.6%) 344 (21.1%)
3 2,390 ( 3.8%) 408 (18.6%)
4 2,484 ( 3.9%) 421 ( 3.1%)
5 2,718 ( 9.4%) 474 (12.5%)
6 2,684 (-1.2%) 487 ( 2.7%)
7 2,582 (-3.8%) 502 ( 3.0%)
8 2,800 ( 8.4%) 523 ( 4.1%)

Note that simply doubling the base figure gives 2,860 and 538.

[Doji3333]

Position Sizing and Money Management

Underdog,

I believe your ideas will improve your system. The other thing you could consider in Money Management (MM).

I have seen system do poorly trading a single contract, but improve greatly using money management.

Like...

Buy 2 on an Entry signal, then sell 1 at a PT (a few points from entry), then let the system trade the last contract.

Sometimes, this turns a loosing trade into a winner.

Combining these techniques may help your system greatly.

Brad