Published August 1, before any game started. A full baseball slate analysed, fifteen games. Nothing edited afterwards. Below are two of them: the one game we would have bet, and one of the fourteen we would not, because the second kind is most of what you get. Football and basketball nights look exactly the same, with the same engines and the same working shown.
Seattle is asking you to risk $169 to win $100, and they have the third worst offence in baseball. They have scored four runs a game all season and have lost eight of their last ten. Minnesota scores nearly five runs a game and has won seven of ten. The market has priced the pitching matchup, which genuinely favours Seattle, and charged almost nothing for the fact that the favourite cannot score.
We played the game 22,000 times on a computer before it happened. Minnesota won 45 out of every 100. The price on offer only needs them to win 42 out of 100 to be worth taking. That gap is the whole bet: small, real, and in our favour.
A gap that size is usually a sign that we are missing something rather than a sign that we found money. So we attacked our own bet five different ways: we made Minnesota's rookie pitcher far worse than any source says he is, we made Seattle's ace as good as anyone claims, and we corrected an error we found in our own park adjustment that day.
The bet survived every one of them. It never got worse than a three point edge. It held up because the advantage does not come from the pitching, which is the part we are least sure about. It comes from a lineup that scores four runs a game being sold to you as a solid favourite.
Seattle's ballpark suppresses scoring more than any other in baseball and they are built for it, winning 30 of 54 at home. Their starter at home against a rookie is exactly where a good pitcher takes over a low scoring game. And a bet with this same shape lost us money the day before.
Minnesota, quarter of a normal bet size. Do not take it below +125. Below that the advantage is smaller than our own margin of error, and the bet stops being worth making.
A second sportsbook was showing the same Minnesota bet at a better price at the same moment, and a run total half a point lower. Shopping between two apps was worth more that day than everything our model found. We tell members that constantly, even though it is not what a betting product is supposed to say.
New York's starter has allowed 3.26 runs per nine innings this season. Chicago's has allowed nearly 7. That is one of the widest gaps between two starting pitchers all day, in a ballpark that usually helps hitters. Our simulator, handed those numbers, expected more than ten runs in the game.
The betting market set the line at six and a half runs. The lowest on the board.
When our number is four runs away from the market's number, that is not an edge we found. That is a confession that we are missing something. So we went looking, and here is what we found: the day before, in the same ballpark, the total was nine with a much better home pitcher. A drop that size with a worse pitcher can only be one thing. Wind blowing in, hard.
Our park adjustment is a fixed number. It cannot see today's weather. So we said exactly that on the page and did not touch the game.
No bet, and the honest reason. The market is holding information we do not have. Betting into that is betting against the weather.
If you are betting it anyway, and plenty of people will, now you know the single thing that decides it. Watch the flags. A game priced this low is telling you the ball is not carrying, which matters a great deal more than either starting pitcher.
That is two games out of fifteen. You get the whole board every night, in every league we cover, with the same working shown on each one. On a normal night one or two are worth real money and the rest are worth understanding.
Owl & Odds is an analytics and entertainment product, not betting advice. We never accept wagers. Must be 19+ or of legal age in your jurisdiction. Past results do not guarantee future results. If gambling stops being entertainment, help is available: The National Problem Gambling Helpline is free and confidential, anytime: 1 800 522 4700. Owl & Odds operates as an analytics and entertainment product.