trading kelly fraction is optimal, the problem is that the swings can be huge and if they make you lose your nerve, then it is pointless. Better to trade a smaller fraction i.e max 'bliss' and be able to stick to the 'system'.
When asked a compound question Ed ignored the half about position sizing and focused on risk management. When he says there's an optimal amount which can be determined via a bliss function he's referring to optimal risk not optimal position size. Ed says without risk there won't be long term gain but too much risk causes short term volatility. Ed suggests and I agree it's preferrable to exclude the subset of profitable trades which cause fluctuations in return. He says use a Sharpe ratio and MAR or "some other bliss function". These are risk management tools not position sizing tools. IMO a viable bliss formula can supersede and negate the need for position sizing concern. Bliss solves risk of ruin by eliminating it at the source rather than including risk and passing it to postion sizing.
Paraphrasing Ed ...a little risk is necessary because without risk there isn't much gain. IMO when things look the bleakest that's the moment when risk and reward are favorably aligned.
Striving for a perfectly straight equity curve doesn't make sense because it penalizes oversized gains for being outliers which cause variance and it ranks a flatlined EC with zero trades as ideal due to flaws in how the default tools focus component weight when the components haven't been separated and preprocessed before passing to the next function. The solution in it's simplist form is to minimize risk while also positively weighting growth in capital. The ultimate goal of bliss is to almost enitrely eliminate the need for considering risk of ruin and b able to nearly always go nearly all in on nearly every signal. There's a lot of people reading ET who are better equiped to take these worthwhile topics to the next level. I'm sorry this isn't much better than my earlier terse reply. edit... @Frederick Foresight pranked me again. ...ohh the hilarity. What do you gain by posting questions and deleting your posts?
Your posts on the subject of position sizing are the best I know of anyone ever having posted to Elite Trader. Please keep an open mind to the possibility of preprocessing your trades with Ed's undisclosed bliss function before passing them to position sizing. And please post more often!
I wasn't pranking you. But you didn't directly answer my question so I decided to bow out, in case I misunderstood. What I thought you wrote earlier made sense to me, and I just wanted you to clarify. The example I used for illustrative purposes was "countertrend" trades (however the trader defines them). They may be profitable on balance, but there is greater risk of loss if the timing is not just right than for, say, "trend" trades. And so, it might be worthwhile to avoid such trades for "bliss" and emotional balance. That is the sort of thing I thought you were alluding to, although it could have been something other than specifically my example of countertrend trades, such as time frame or whatever. But I don't agree with this comment: I think that the idea of nearly always going nearly all in, whatever the category of trade, is the royal road to ruin.
Counter trend setups enable tight stops because they either work relatively quickly or there's no reason to b there. Large stops cause undesirable variance everywhere regardless of what the setup is ...or how huge the target ...or how impossible it is to focus development on small stops. That's the gist of this paradigm. Humor bro ...count the times I used nearly in that sentence about going all in. 5? 10? Going all in is a joke but without lofty goals we're unlikely to find ways to attain them. Seriously, small stop, huge target, awesome R:R ratio, huge upside variance, tiny and stable downside variance. It requires effort to visualize separating the variance but that's the solution. Seriously, the ability to go all in all the time is the goal. Diversification will cover any residual risk of ruin for ten 100 year catastrophic failures. It's a huge waste to target development efforts at anything less than perfection.