Confidence pick vs value pick: two different bets
These two answer different questions and often point at different matches. Mixing them up is one of the most common ways to misread a model.
Two questions, two picks
There are two very different questions you can ask about a day's matches. "Who is most likely to win?" and "Where is the price wrong?" The first is about the outcome; the second is about the odds. They rarely have the same answer.
The confidence pick
The confidence pick is the match where the model is most sure of the winner — the factors line up hard behind one player. It is chosen for conviction, not price. That means a confidence pick can be a very short favorite with negative expected value: likely to win, but not a good price.
The value pick
The value pick is chosen for the opposite reason: the model rates a player higher than the market does, so the price is longer than fair and the expected value is positive. It is often a live underdog, and it will lose more often than the confidence pick. The edge, if it is real, lives in the price over many bets — not in any single result.
Why they're tracked separately
Because they are different bets with different hit rates, mixing them into one record would be misleading. A basket of longer-priced value picks will win a lower share of matches than a run of short favorites — that is expected, not failure. Baseline keeps them as separate series so neither flatters nor drags the other, and publishes both on the Transparency page.
Which should you follow?
That is a question of temperament, not correctness. The confidence pick trades upside for reliability; the value pick accepts more losses for a mathematical edge over time. Both are informational analyses, not betting advice — and neither promises a profit. You can see both, side by side, on the daily pick page.
Frequently asked questions
What is the difference between a confidence pick and a value pick?
The confidence pick is the match where the model is most sure of the winner (chosen for conviction). The value pick is where the model rates a player higher than the market, giving positive expected value (chosen for price). They often refer to different matches.
Why does the value pick lose more often?
It is usually a longer-priced underdog. A lower win rate is expected — the edge, if real, comes from the price paying more than fair across many bets, not from winning most individual matches.
Can the confidence pick have negative EV?
Yes. A heavy favorite can be very likely to win yet priced too short to offer value. Confidence is about the outcome; value is about the price.
Why track the two picks separately?
They have different natural hit rates. Combining them into one record would misrepresent both, so Baseline reports them as separate series.