Why free pick sites publish so many picks

Open any of the big free pick pages and count. Forty ratings across a dozen sports, every day, each one given the same confident tone. It looks like abundance. It is arithmetic. Those pages do not earn money when a pick wins; they earn money when a reader clicks through to a sportsbook and signs up. Under that model, the correct number of picks to publish is every pick you can generate, because every one is another affiliate link and none of them has to be right. The reader’s job — sorting forty contradictory ratings into the two or three bets actually worth making — is left to the reader.

What volume does to a record

Volume is also the cheapest way to look good. A service that posts one pick a day is judged on that pick: a mediocre record is visible immediately, and there is nowhere to hide it. A service that posts forty a day gets to choose what it shows you afterwards. Six of the forty went 6-0? That is Monday’s headline. Nobody is asked to have followed all forty, and nobody totals them up. The math behind this is dull and brutal: spread enough picks across enough games and some subset of them will always have had a great week. The more picks there are, the more highlight reels there are to sell — and the less the whole set, taken honestly, tells you about the next one.

There is a second, quieter effect. When a page publishes everything, it never has to stand for anything. If a model liked a heavy favourite at a bad price, the pick still ships, because an unpicked game is a page with no affiliate link on it. The discipline — the willingness to look at a whole board and say “nothing here is worth your money” — is precisely what the volume model cannot afford.

What choosing three costs us

We publish exactly three picks a day, every day, and grade all of them in public afterwards, whatever they do. The constraint is the honesty mechanism: with three picks a day there is no subset to cherry-pick, no under-performing corner of the ledger to leave out of the headline. Our full record — 23-19-1 at the moment — is one unbroken line, losses dated and graded, the days nothing cleared included. On an empty day we publish the empty card and say why. That willingness costs us clicks every day, and it is the whole product.

The trade-offs are real, and you should know them before paying for anything of ours. We say no to most of the board, so on many days the two or three plays we pass on would have won. A card of three can go 0-3 while a page posting forty ratings points at the fifteen that cashed. And our prices currently come from one public sportsbook rather than a multi-book feed, which means the best available price somewhere else is sometimes a tick better than the one we publish — we note this on the record rather than pretend otherwise. If you want breadth, or a book-shopping engine, we are not it; the method page shows exactly what Three does and does not do.