A Substack and a few signal packs, written by someone who's been wrong enough times to be honest about it. Fifteen years of insider-buy data, real out-of-sample drawdowns, the trades that worked and the trades that didn't. If you want a guru who promises 300% returns, this isn't it. If you want sharper decisions and the math behind them, welcome.
Every strategy shown with its worst 12-month period. Sharpe, Sortino, win rate, average trade, max drawdown. If it looks too good on paper, I'll tell you why.
If someone's 300% returns are consistent and real, they're managing a billion for Point72, not writing Substack for $10/mo. Steady 8–15% edges with discipline. That's the game.
Form 4 is public. Most retail parsers get false positives from 10b5-1 plans, trust distributions, tax-lot gifts. The pack filters those out. What's left is signal.
Nothing here is advice. You're an adult with your own account. What I ship is data, backtests, and how I think. How you size, when you enter, whether you even take the trade — that's you.
Fifteen years of Form 4 filings, filtered for significance (>$500k, open-market buys, non-10b5-1, non-gift). Backtested long-only, quarterly rebalance, 60-day hold. Delivered as SQL + Python notebook + monthly CSV of flagged filings.
Post-earnings-announcement drift is one of the oldest, most-documented anomalies in quant finance. The pack quantifies it cleanly: by size bucket, by surprise magnitude, by sector. Useful for swing traders who want to know when the crowd underreacts.
Classic 12-1 momentum, rebuilt with modern risk filters. Not the 2001 Jegadeesh paper — the 2025 version that survives high-correlation, low-vol regimes. For traders who can hold 40 names and rebalance monthly.
I'm Efe. CS senior, trilingual, built ML systems in production before I started managing my own book. Biskin Quant is the Substack and packs I wish existed when I started. Not because I'm an oracle — because I've been wrong enough times to write it honestly. The backtests are real. The drawdowns are real. The things I don't know, I'll tell you I don't know.
The goal isn't to replace your own judgment. It's to sharpen it. Read the free issue, decide if the signal pack is worth $149. If it's not, don't buy it. If you find a mistake in the math, email me. I'll fix it publicly and credit you.
Issue 01 is the insider-buy deep dive — how the data gets filtered, the full 15-year backtest with OOS, the three filings that matter this month. Free. In your inbox. Then roughly weekly after that, unless I have nothing honest to say.