Do hot handicappers stay hot?
We split every established record in half by pick order and compared the two halves. Recounted at each site update. Last update November 3, 2026.
The finding
Accounts that opened above plus 10% ROI (average +19.2%) had a median second half ROI of +2.5%, and 54% of them were up in the second half. Accounts that opened below minus 10% (average -15.0%) had a median second half of -8.4%, with 31% up.
Hot starts cooled sharply: from an average first half ROI of +19.2% to a median second half of +2.5%. Across all 115 accounts the two halves are somewhat related: the correlation is 0.19 (95% interval 0.05 to 0.36), where 0 means no relationship and 1 means the second half repeats the first. The average account went from +2.6% ROI in its first half to +0.0% in its second, and 50% of accounts were up in their second half.
A strong start is weak evidence of an edge. A record moves toward break even as it grows, so judge a handicapper on a long record graded at the posted odds.
Second half results by first half ROI
| First half ROI | Accounts | Average first half ROI | Up in second half | Median second half ROI | Average second half ROI |
|---|---|---|---|---|---|
| Cold: below minus 10% | 13 | -15.0% | 31% | -8.4% | -6.5% |
| Slightly down: minus 10% to 0% | 34 | -4.4% | 38% | -4.4% | -4.6% |
| Slightly up: 0% to plus 10% | 44 | +4.2% | 64% | +4.8% | +4.2% |
| Hot: above plus 10% | 24 | +19.2% | 54% | +2.5% | +2.3% |
The smallest group has 13 accounts, so a single account moves its percentages by several points. Read the groups as direction, not precise rates.
What this shows and what it does not
- It describes the past. These are measured results for accounts already on the board. They are not a forecast for any account and not a betting system.
- It does not say good handicappers do not exist. A small positive relationship is consistent with a few accounts holding a real edge and many hot starts being luck.
- Past ROI never guarantees future profit. Even the accounts that stayed up in their second half may not stay up in the next one.
Method, sample and bias checks
- Source: the full graded pick history of each account in the most recent verified snapshot, the same history Watchdog Pro members see. Every pick was posted in public before its game started and graded against the final score at the posted odds. Held, duplicate, repost and parlay leg picks are excluded.
- Selection: accounts with at least 100 graded picks in that verified history whose history reproduces the headline record exactly: 115 accounts. The rankings show 135 accounts at 100 or more picks; the difference is accounts that crossed 100 with picks graded after the last verified snapshot.
- Split: each history in pick order, oldest half against newest half (odd counts put the extra pick in the newer half). ROI is units won divided by units risked.
- No look ahead in the grouping: accounts are grouped by their first half only, and the outcome is measured on picks made after that half ended.
- Survivorship: to be included an account had to keep posting until it reached 100 picks. Accounts that quit after a bad or a lucky start are missing, so the cold group may look better than all cold starts really do. The study speaks for accounts that keep posting, which are the accounts a bettor can actually follow.
- Uncertainty: the 95% interval comes from 2,000 bootstrap resamples of the published pairs with a fixed seed, so anyone can reproduce it.
- Population: accounts that post picks in public on the platforms Watchdog monitors. They are not a random sample of all bettors.
- Nothing here is betting advice.
Look at real records
The four largest independently graded records on the board. Every profile shows the record, units, ROI, sample size, sport by sport results and recent form, free. Accounts are shown by anonymous Watchdog ID only.
#259F31
537 graded picks
-9.49 units, -1.9% ROI. Full breakdown by sport, market, odds and month.
#E2F536
481 graded picks
-24.15 units, -5.0% ROI. Full breakdown by sport, market, odds and month.
#9D8D0D
466 graded picks
+4.21 units, +0.9% ROI. Full breakdown by sport, market, odds and month.
#8E2EDA
321 graded picks
+9.13 units, +3.4% ROI. Full breakdown by sport, market, odds and month.
What Watchdog Pro adds
- Eligible pending picks from tracked accounts before their games start, with each account’s graded record beside the pick.
- The complete graded pick log for every account, not only the 50 most recent picks.
- Follow accounts and get an email when one of them posts a pick before the game.
Pro never reveals who an account is. Free visitors keep the rankings, the research and every profile breakdown. See how Watchdog grades a pick.
See the long record, not the hot week
The free rankings sort by sample size, units and ROI.
Updated November 3, 2026. 21+. Gambling problem? Call 1 800 GAMBLER.