Hi,
I think laudable that dukascopy is refining the contest rules to encourage more "realistic" trading strategies, however I think the new rule 6.3 (no winning trade can make more than 50% of previous days equity) hinders strategies that are successfull at predicting price moves.
A strategy is good not only by predicting the direction of a move, but also the extent of it, and deciding when to cut the losses short or when to close a successful trade that the strategy decide would not go higher. By capping the moves that can be captured, the strategies can not capitalize on higher, successful prediction, therefore lowering the incentive to develop algorithms trying to predict when a move has reached the maximum potential for example. Overall, rule 6.3 discourages developing the strategy in an area that is as important, if not more, than simply deciding when to enter a trade, i.e: when to close a trade.
Rule 6.3 also does not address what happens with strategies that closes trades partially. For example, it is a valid money management scheme to close half of a winning position in order to capture some gain, then let run the other half, or close it if the market pulls back and before the position left becomes a losing trade. One can imagine many other strategies involving more sophisticated rules to close positions, but rule 6.3 will certainly discourage developers to investigate those.
A different rule that would approach the same goal of a lesser "gambling" position sizing strategy, but leave the incentive to develop strong position close strategies, would be to cap the size of any given trade to, for example, 5% of current equity.
I hope dukascopy will consider my comments for future contest rules
|
