Do you need to trade more when you have a positive expectancy? or.. Is it better to identify the best time to trade that positive expectancy? Refinement and better profits need to be evaluated. Is it sheer numbers of trades or quality of trades? ES
Expectancy is measured from a defined set of metrics . Take every trade where these metrics are satisfied , Its that simple ..
But I suppose one could cherry pick trades within that set to improve upon the results? In other words positive expectancy is a good place to start? ES https://www.stator-afm.com/tutorial/statistics/positive-expectancy/ Positive expectancy is defined as how much money, on average, we can expect to make for every dollar we risk .
Usually, quality of your PnL (e.g. Sharpe/Sortino or such) is a trade-off vs your transaction costs (i.e. PnL per trade value). So the more you trade, the smoother your PnL is but the more you pay in TCs. If you have a way to tell that you have a slightly better or slightly worse trade, you should use that in sizing your trades but add some form of a threshold to reduce your trading frequency and thus lower transaction costs.
Dealing with averages can actually reveal a negative expectancy but one should not give up as there can be time filters applied without changing the methodology to change the positive expectancy to a negative expectancy. Simply trading the morning session verses 24hrs.... ES So we as traders need to determine, from a representative sample size of trades, a system’s percentage accuracy, average win and average loss. If the system doesn’t produce a positive expectancy then it should be left alone. To trade it would only lead to ruin!
Agreed! Someone said "RIP sle" in a post and I thought he actually died!! I inquired via PM but was notified that only his handle was "dead".
In general the lower the expectancy the higher the drawdowns/losing streaks. You can try and filter the signals from a low expectancy system and try and turn it into a higher expectancy system. Then you can bet larger and also pay less in transaction costs.
There are 2 sides to expectancy and in my world one is rated higher than the other . It all boils down to probability curves but i will take the system with a higher winrate over one with a lower winrate given equal expectancy . When building systems being aware of this is a huge advantage , most are clueless to it . 2 systems with same expectancy the one with highest winrate will have the smoothest curve and therefore a lower isk of ruin . Tails are worst case scenarios , a metric most are oblivious to . Clearly both sides of expectancy are important but one side deserves a higher priority . OK the can of worms is open
ES, In a positive expectancy trading system, as your average winning pct rises, your avg win to loss ratio will decrease and vice verse. Find something that suits you in the 30-70% avg winning pct area (40-60% is easier on the nerves). Past 70% and you're going to be taking upside-down trades (risking more than you stand to gain...don't do that)