If you strip all the algebra away, the key points are: 1. Sharpe...
sysls@systematicls
62 views
Nov 27, 2025
~1 min read
Advertisement
1
If you strip all the algebra away, the key points are:
1. Sharpe decomposes:
a. Sharpe ≈ IC × √N
b. IC = skill
c. √N = number of unrelated bets
2. IC is correlation:
a. Between your predictions and what actually happens, after adjusting for how volatile each stock is.
b. It’s a clean, scale-free measure of forecasting skill.
3. IC is intensive:
a. It doesn’t care how many stocks you trade.
b. It’s “how good is each bet”.
4. √N is extensive:
a. More independent names -> higher Sharpe, but with diminishing returns (square root, not linear).
5. Research program:
a. Clean the returns with a factor model to get idiosyncratic returns.
b. Normalize predictions.
c. Regress risk-adjusted returns on risk-adjusted predictions → slope = IC.
d. Build models to maximize IC over time.
6. Separation of concerns:
a. Signals: maximize IC.
b. Portfolio & execution: maximize TC, i.e., how much of that IC you actually turn into PnL.
1. Sharpe decomposes:
a. Sharpe ≈ IC × √N
b. IC = skill
c. √N = number of unrelated bets
2. IC is correlation:
a. Between your predictions and what actually happens, after adjusting for how volatile each stock is.
b. It’s a clean, scale-free measure of forecasting skill.
3. IC is intensive:
a. It doesn’t care how many stocks you trade.
b. It’s “how good is each bet”.
4. √N is extensive:
a. More independent names -> higher Sharpe, but with diminishing returns (square root, not linear).
5. Research program:
a. Clean the returns with a factor model to get idiosyncratic returns.
b. Normalize predictions.
c. Regress risk-adjusted returns on risk-adjusted predictions → slope = IC.
d. Build models to maximize IC over time.
6. Separation of concerns:
a. Signals: maximize IC.
b. Portfolio & execution: maximize TC, i.e., how much of that IC you actually turn into PnL.
Gappy (Giuseppe Paleologo)@__paleologo
This morning I started putting together a long-ish note on the Transfer Coefficient, so I started to write about the Information Coefficient, and that sent me on a detour and I haven't even started on the TC yet. But these three pages are probably the shortest zero-to-one intro to the Master framework for cross-sectional quantitative investing, so I thought I would share it here. Send corrections and suggestions.
2
You're welcome, kindergarden kids with Jane Street interviews lined up.
3
For the rest of you who graduated university like the rest of the normies, know it's already too late for you... unless you CATCH UP by FOLLOWING me.
4
@Mariellangsaez Por cada 5 me gusta necesitas retuitearme una vez. No tienes otra opción.
