Re: Position trading 100+ futures markets times 8 systems

[Mark Johnson <janitor@xxxxxxxxxxxx>]

Writing about using a mechanical system to trade
100 instruments, then 200 instruments, then 300
instruments, Bob Fulks says:

  BF> So the benefit decreases fairly rapidly but the
  BF> transaction costs increase linearly with the
  BF> number of equity curves so there should be an
  BF> optimum number of commodities to trade using
  BF> this method.

Bob, you've made an assumption: that transaction
costs increase linearly with the number of tradeable
instruments in the portfolio.  This assumption isn't
true in all cases.  For example it isn't true in my
case.

For me, and for futures traders in general,
transactions costs are linearly proportional to the
total number of CONTRACTS traded.

In my case, I use a positionsizing approach known
as "fixed fractional betsizing".  It seeks to keep
the total exposure (the total number of contracts,
across all positions in all instruments of the
portfolio) proportional to the account size.  Fixed
fractional positionsizing operates along these lines:

#Contracts per trade =
   Constant * Bankroll / #instruments_in_portfolio

Tripling the number of instruments in the portfolio
cuts the (number of contracts per trade) by a factor
of 3, and increases the total number of trades by
the same factor of 3.  Therefore the total number of
contracts traded ("total exposure") remains constant
when the portfolio size is adjusted.  Thus the total
transaction costs remain constant when the portfolio
size is adjusted.  For me and for other multi-
instrument portfolio traders who also use fixed
fractional position sizing.

Mark Johnson