THE HAnDbook of TECHnICAl AnAlysIsThis implies that the system, under dynamic sizing, has a mean‐reverting ten-
dency or behavior when it is in profit. Since the order or sequence of wins and
losses has no effect on the outcome, this also means that it would take seven
winning trades to return the capital to breakeven after losing only four trades,
assuming that all trades were initiated with the same R/r ratio. This implies that
dynamic sizing has a loss‐promoting tendency or behavior when the system is
experiencing loss. To reduce the effects of asymmetry (and thus reduce the impact
of both the mean‐reversion bias when in profit and loss‐promoting bias when in
loss) the trader may:
■ (^) Reduce the %risk per trade
■ (^) Reduce the number of trades being compounded
Monte Carlo simulations of fixed and dynamic sizing systems indicate the fol-
lowing with regard to asymmetry:
■ (^) When R/r →1:1, there is no asymmetry for small %r in fixed sizing systems
■ (^) When R/r →1:1, there is no asymmetry for large %r in fixed sizing systems,
but the balance curve is more volatile
■ (^) When R/r →1:1, there is no discernible asymmetry for very small %r in
dynamic sizing systems
■ (^) When R/r →1:1, there is significant asymmetry for large %r in dynamic sizing
systems, but the balance curve is more volatile
It is also established via random simulations of hundreds of thousands of
trades that any trading edge or advantage in the system kicks in and begins to
impact the system favorably in the following manner:
figure 28.27 Asymmetry in the Dynamic-Sizing Approach.