the live counters wake up the second a real number lands. no fake-empty states, no zeros pretending to be data.
if a number isn't real yet, we don't show it.
signal priority
Every pick on the slate gets a verdict pill. Here's what the four tiers mean, the thresholds that decide them, and how to size each one.
edge
≥ 8% edge
confidence
≥ 65% confidence
action
The shortlist. Both bars cleared.
Anchor tier requires the model to beat the line by a wide margin AND have high confidence in the underlying estimate. Most days produce a handful. Some days produce zero — that's the system working, not failing.
sizingStandard unit size (~2-3% of bankroll). Don't multiply just because the tier name sounds aggressive.
edge
3-8% edge
confidence
no confidence floor
action
Real edge, normal sizing.
The bulk of the published slate. Edge is meaningful (3-8% beats the line), confidence isn't necessarily anchor-grade. These are the picks you bet because the math is on your side — not because they're locks.
sizingHalf unit to one unit. The expected value is positive; variance is the price.
edge
< 3% edge
confidence
any confidence
action
Tiny edge — informational.
Marginal picks are below the threshold where the math is worth chasing for most bankrolls. We publish them so the slate is comprehensive — knowing what we're avoiding is as useful as knowing what we're picking.
sizingSkip unless you're managing a large enough roll that 1% edges matter, in which case you don't need our advice.
edge
≤ 0% edge
confidence
n/a
action
The line is fair or against us.
Books beat us, ties, or the model thinks the line is fairly priced. Published for transparency — the slate isn't curated to make us look good. You should not be betting these.
sizingZero.
/parlay/best-spots filters the slate to anchor-tier picks only. If a day has zero anchors, the page says so — better an honest empty than a promoted marginal.
