Skip to content
Beta · Phase 1 Anti-Tout Science Desk

Trout Tout Track real tout ROI vs our public model picks.

Per-source pick-accuracy leaderboard. Sample-size shrunk (Bayesian), 95% confidence intervals. Negative lift means the source's calls have underperformed the sport-position baseline — that's a trout. Fade it.

Building sample…

No source-accuracy data yet for this sport/window. The first cron tick that pairs mentions with player-game outcomes will populate this leaderboard. Check back in a few hours.

NBA

Sources making the boldest player calls (high-confidence + strong sentiment, last 90d). Hot-take index sums signed conviction; red = bullish-loaded, green = contrarian/bearish.

1 POD The Bill Simmons Podcast 222 +3477
2 YT JxmyHighroller 59 +3439
3 YT Portland Trail Blazers (Official) 54 +3523
4 YT Thinking Basketball 37 +2746
5 YT Mike Korzemba 34 -1052
6 YT ESPN NBA 30 +1176
7 YT All The Smoke (Barnes/Jackson) 27 +2062
8 YT Bleacher Report 15 +372
9 NEW CBS Sports NFL 15 +265
10 YT The Athletic NFL Show 13 +850
11 YT HoopsHype 12 +1053
12 YT CBS Sports NFL 8 +410
13 NEW ESPN NFL News 4 +225
14 YT The Pat McAfee Show (YouTube) 2 +143
15 YT Chris Mannix (Sports Illustrated) 2 +143
16 NEW Pro Football Focus 2 +86
17 NEW Acme Packing Company (GB) 1 +45
18 NEW Windy City Gridiron (CHI) 1 +42

What the columns mean

Lift
Shrunk point estimate of how much better (or worse) the source's calls performed vs the sport-position baseline. Positive = the source helped, negative = trout.
Hit rate
Win % on explicit picks, excluding pushes. 52.4% is breakeven vs -110 juice.
95% CI
Confidence interval around the point estimate. Wide bars = thin sample; thin bars crossing 0 (or 50%) = the source is statistically indistinguishable from noise.
n
Sample size — observations contributing to the estimate. n < 10 is flagged as a small sample (faded CI bar).
ROI
Return on investment if you'd faded (or followed) the source at -110 unit-flat. Reported in the per-source detail page.
CLV
Closing-line value — how much the line moved after the source's pick, in the source's direction (avg %). The strongest long-run skill signal.

Methodology. Lift is computed over a 14-day window after each episode published. Shrinkage uses an empirical-Bayes prior with pseudo-count 50 toward neutrality (lift = 0 or hit-rate = 50%). 95% CIs are Wald (lift) and Wilson (hit rate). Small samples (n < 10) appear but with faded CI bars.

Not investment / betting advice. Historical accuracy doesn't predict future performance, especially with thin samples. Use this as one signal among many.

We use cookies for essential site functionality. With your consent, we also use cookies for analytics and performance monitoring. See our Privacy Policy.