Friday results: six misses from seven
Seven BAI8 Desk calls settled on Friday 9 October and six lost, against 3.4 expected wins. Every result, what the night did to the record, and what it proves.
On this page
Seven of our Desk calls settled on Friday 9 October (the NHL games early on Saturday morning UTC) and six lost. Only PSV against Heerenveen won. Our own probabilities expected about 3.4 wins from those seven, so this was a bad night, and it goes on the record with the rest.
The record after Friday
| Settled | Won | Lost | Expected wins | Brier score |
|---|---|---|---|---|
| 33 | 16 | 17 | 17.8 | 0.227 |
A Brier score is the average squared gap between our probability and what happened (0 is perfect; always saying 50% scores 0.25). At 0.227 we are only a little better than a coin.
The seven results, misses first
| Fixture | Our pick | Our chance | Best price | Result |
|---|---|---|---|---|
| Columbus Blue Jackets v Pittsburgh Penguins | Penguins | 51.3% | 2.06 | Lost |
| Detroit Red Wings v Seattle Kraken | Red Wings | 53.8% | 1.74 | Lost |
| West Ham United v Queens Park Rangers | West Ham | 42.7% | 1.57 | Lost |
| RC Lens v Lyon | Lyon | 37.7% | n/a | Lost |
| Borussia Dortmund v Werder Bremen | Dortmund | 53.7% | 1.41 | Lost |
| Málaga v Espanyol | Espanyol | 38.9% | 2.86 | Lost |
| PSV Eindhoven v Heerenveen | PSV | 59.0% | 1.27 | Won |
The record shows no best price for Lens v Lyon, so we leave it blank.
What a night like this means
None of these picks was above 60%. Six misses from seven is unlucky on our numbers, but not rare enough to prove the model broken; we cover how long a bad run can honestly last in how long a losing run can honestly last. What it does do is pull our record below expectation, and that gap is what we watch.
The positive-edge calls
Five calls had a positive edge at the price we saw. One won (Dodgers at Braves, 1.99). Four lost: White Sox v Guardians twice, Espanyol at Málaga and Penguins at Columbus. Five is far too few to judge.
What this does not tell you
Thirty-three calls is a small sample. It cannot tell you whether the model is well calibrated, and it says nothing about calls not yet settled.
18+. Bet only what you can afford to lose.
Common questions
What is BAI8's full record after Friday?
Thirty-three calls settled. The side we published won 16 and lost 17. Adding up our own probabilities, we expected about 17.8 wins.
How did Friday go for BAI8?
Badly. Seven calls settled since our last update and only one won, PSV against Heerenveen. Our probabilities expected about 3.4 wins.
How have BAI8's positive-edge calls done?
Five of the 33 had a positive edge at the price we saw. One won (Dodgers at Braves) and four lost, including Penguins and Espanyol this week.