Well, you have to do look at some sort of risk-adjusted returns otherwise it all becomes meaningless. HFT produces really good returns on capital because of the high turnover rates (it’s pretty amazing to see what sort of working capital some HFT firms use, actually). If anything, what makes discretionary managers/traders more interesting is lower capacity constraints and alpha that can’t be replicated easily.
What's better, systematic or discretionary?
- Thread starter sle
- Start date
That's your view and doesn't have much to do with systematic vs discretionary. Separate argument.
Of course the best discretionary trader can pick the best performers, averaged over time, over a system picking a basket. Yes I assumed the best when I typed "at the peak level", not EVERY discretionary trader.
Returns on capital are ridiculous - it’s common to see 10-20% a month before technology costs (that’s where the catch is) and UHFT players produce up to 10x of that. It’s not the traditional drawdowns that an HFT trader should be worried about, its loss of alpha and the ensuing bleed (commissions, technology costs etc add up quickly).
PS. I don’t do true HFT, but since I have some intraday strategies as well as write my own execution algos I have some knowledge about the space.
Last edited:
Actually, simply by virtue of information complexity (a discretionary trader has a much higher number of inputs into his decision process than any systematic strategy), you would imagine that even the best discretionary traders tend to have lower win rates but higher expected value per trade. My experience as well as the paper d08 linked agrees with that intuition.
A "lower win rate" can occur for many reasons, only one of which is whether the approach is discretionary or systematic. Attempting to isolate "discretionary or systematic" as independent variables would be a nightmare, much less conducting the actual tests. Those who implement the approach I follow have win rates of about 68-76% and seemingly unusually high profit-to-loss ratios due chiefly to the fact that losses are automatically cut short and profits are allowed to run for as long as the market cooperates. To achieve this, however, one must be systematically discretionary; in other words, one must know exactly what to look for, enter at exactly the right point, but then manage each trade as a unique occurrence.
Just out of curiosity, why are you asking the question?
Don't quite agree there.
https://www.elitetrader.com/et/attachments/et2-png.175637/
You can have both decent win rate and payout. The "best" are even more better.
Because I am bored
something @lawrence-lugar wrote as a response to a system seller caught my attention and I figured we can discuss it.Through my years in finance, I have yet to see any non-systematic traders that have realized Sharpe ratios over say 1.5 in the long run so I will take this with an appropriate grain of salt. But I am not going to say it’s totally impossible either.
My sharpe is even better if I update it with this years. Probably because you are in funds management where sizes make it more difficult. Different worlds we are measuring I guess. For accounts under 10m there are plenty of traders with a few years sharpe above that on account browsers like FS/MyFXbook.