Hi,

One rule in the lab is up 24% on its bets. We think most of that is luck, and here's why.

The rule

B4 bets the under on early-season NBA totals. The belief is simple: early in the season, books set totals a little too high, because teams haven't settled into how they really play yet.

The 24%

34 picks graded, 22-12. Bet one unit on each and you'd be up about 8 units. That's a 24% return.

It's a real number, and it's mostly luck. 34 picks is a tiny sample, and a hot streak looks exactly like this.

The number we trust more

The close. On average B4 beats DraftKings' closing price by +0.89 points a pick. It beat the close on 20 of the 33 lines that moved. For comparison, our "bet every favorite" control sits at +0.08.

That edge is real, but it's small. It's well short of the book's cut, so its prices alone don't support a 24% return. When the luck runs out, the close is what's left.

Why we expect it to fade

If the belief is right, books learn as the season goes on. Totals come down, and the gap closes. So the real test isn't whether B4 keeps winning. It's whether that +0.89 shrinks as the weeks go by.

B4 is a research rule. It never enters a pond and it will never be sold. We are watching it to learn how fast books adjust.

Every pick stays on the board, wins and losses: pixelforgeplays.com/lab

Your turn

Hit reply and tell me one thing you think books get wrong early in the season. If it can be tested on our price history, we'll test it and tell you what we find.

Casey

P.S. Dan pinned his graded record on his profile. You can see it at pixelforgeplays.com/b/ocelotanalytics