Most trades that "worked" have some combination of risk premium/structural alpha, and genuine alpha. That is, you were getting paid to either take a risk or perform a service, but also you got paid more than you "should" have because others had not noticed how good the trade was
-
-
Way too much data-mining alpha, tempted to say even Eugene Fama worried about this one.
-
Agree. You need to look at the track record of actual value investors (and tease out the value premium from their other exposures) or look at post-publication performance of trading strategies to avoid it.
- Show replies
New conversation -
-
-
It's really hard to explain to someone who hasn't done this a lot how Every. Single. Time. in-sample $/alpha are bigger than OOS. To the point where you yell "Dammit, I did everything right! I was a good little boy, super careful with data, etc!" And it's *still* overfit.
-
Then you see someone pitch you their new quant fund with some backtest/papertrade results and you just laugh (internally). You feel like a jerk, because they really do believe their
#s. And *they're* trying to be good too! Doing things right, honest, etc. It's just *that* hard. - Show replies
New conversation -
-
-
Providing liquidity can be a solid revenue stream, if executed well. Equity stat arb should work as long as market inefficiency exists — various investment horizons, risk constraint from mandates, geographical differences, etc
-
Equity stat arb is good with alpha in there
End of conversation
New conversation -
-
-
Overfitting ftw
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.