Every analysis on Oddlytics returns a confidence percentage between 1 and 99 — never a flat 0 or 100, because certainty like that doesn't exist in prediction markets. The number reflects how strongly the analysis favors one side once it has weighed base rates, historical precedent for similar events, and what the current price already implies about market consensus.

The number alone isn't the point, though. What actually matters is the gap between that confidence and the market's own implied probability. A pick sitting at 95% confidence on a market already priced at 94¢ isn't an opportunity — the market already agrees. A pick at 70% confidence on a market priced at 40¢ is a much bigger gap, even though the raw confidence number is lower. Oddlytics calls this gap "edge," and it's what Handpicked Bets actually ranks by, not raw confidence.

Position sizing follows the same logic in reverse: the app never suggests more than 5% of stated capital on a single position, and caps it lower — 2%, sometimes less — when confidence sits under 60%. Low confidence isn't hidden or rounded up; it's stated plainly, with the sizing suggestion shrinking to match.

None of this is a promise. It's a structured way of saying "here's how sure this analysis is, and here's exactly why," so you can weigh it yourself rather than take a number on faith.