|
PureBytes Links
Trading Reference Links
|
Ross, I dont think your measurement is really important. What is the
difference between 2 and 5 if five had a DD of 90% and 2 of 6%?
Did you publish the system here:
http://www.wealth-lab.com/cgi-bin/WealthLab.DLL/editsystem?id=26928
At least it is published from a guy from Australia (Adelaide) and he is
asking the exact same question. If you are not the same here is the
beginning of his description:
...This script is to demonstrate the importance of risk management.
The trades are entered randomly (long only) and exited using either a
percentage profit stop (PS default equals 4) or a percentage loss stop
(LS default equals 2). ...
Volker
||-----Ursprüngliche Nachricht-----
||Von: Ross Bond [mailto:Ross.Bond@xxxxxxxxxx]
||Gesendet: Wednesday, February 11, 2004 10:58 AM
||An: Omega List
||Betreff: Position Sizing Effectiveness
||
||
||
|| A position sizing question if I may:
||
|| When testing single MarketSystems with position sizing (using Fixed
|| Ratio and Tharp's %Volatility) I calculate the effect of applying
|| position sizing by the simple means of calculating the ratio of
|| [equity with position sizing /equity without position sizing], so if
|| a MarketSystem with position sizing applied returns a closed equity
|| twice the size of the same MarketSystem without position sizing then
|| the effect of applying position sizing is 2.0.
||
|| Given that means of evaluating the performance of position sizing on
|| a single MarketSystem, what sort of values would seem to be
|| reasonable? As an example I have some MarketSystems that return
|| position sizing effect values of 5.0 over a 10 year test period.
||
|| I also have MarketSystems that do much worse over much longer
|| periods, and if I go crazy with some curve-fitting I can get much
|| better returns over much shorter periods...
||
|| I would appreciate hearing from list members what sort of position
|| sizing effect values they have observed in testing or in trading.
||
|| TIA!
||
||--
||Best regards,
|| Ross mail to: Ross.Bond@xxxxxxxxxx
||
||
|