PUBLIC RECORD · COMPLETED RESULTS ONLY//PUBLIC RECORD · COMPLETED RESULTS ONLY//
THE DAY6:30 AM CST YESTERDAY GRADED8:00 AM CST MODEL BUILDS11:35 AM CST FIRST PITCH~1:00 PM CST ALT LINES PRICE2:00 PM CST RE-SCREEN~3:30 PM CST LINEUPS → PICKS FREEZE5:00 PM CST TOMORROW EARLY LOOK6:40 PM CST TOMORROW SCOUT7:45–11:45 PM CST TOMORROW BOARD REFRESHES
THE VITALS
model diagnostics · day 100 of tracking · rendered THU 09/03 · 8:26 PM CST
BET-STRENGTH RECORD61.5%
BET-STRENGTH LEGS · MODEL ≥1K OFF THE LINE?
ALL GRADED LEGS2,216
CI FLOOR?%
BREAKEVEN57.7%
VERDICTNOT YET PROVEN
01
TONIGHT’S LINE MOVEMENT
every model lean’s entry vs the line, shown once the arm’s game is IN PLAY · GOLD ROWS = OUR CARD, listed first · no closes stamped yet · PROVISIONAL: the graded call happens at close, on the RECORD tab
PITCHER
PICK
OPEN
NOW
DRIFT
no drift to show yet · rows publish arm by arm as games go in play
02
WHAT WE’VE LEARNED
the three things worth knowing about this model, graded on real slates
01picks that BEAT THE CLOSING LINE win 62% · only 50% when we don’t. that gap is the whole thesis: beat the close and the wins follow
02when our card and the sharp market AGREE we win 67%(46 picks · small sample, do not extrapolate)
03NO signal has graduated yet · every honest floor is still below the 57.7% breakeven. that is why the evidence remains provisional
03
CONVICTION PERFORMANCE
518 graded legs · the further our number sits from the line, the more it wins · ordering holds every bucket · honest read: none clears the 57.7% DFS breakeven yet
BUCKET
WIN%
N
conviction 0.0-0.5
51%
1100
conviction 0.5-1
54%
685
conviction 1.0-1.5
55%
294
conviction 1.5-2
67%
81
conviction 2.0-99
52%
46
UNDER THE HOOD · full diagnostics (signal scorecard · nightly run history)
SIGNAL SCORECARD · every scenario we track; a signal graduates only when its floor clears 57.7%
SCENARIO
N
WIN%
95% CONFIDENCE FLOOR
market ALIGNED
228
54%
48%
market FADED
156
52%
44%
market NEUTRAL
95
52%
42%
beat the close
149
62%
54%
did NOT beat close
183
50%
43%
discrepancy CONFIRMED
262
54%
48%
conviction >=1.0
78
56%
45%
conviction
401
52%
47%
carded & ALIGNED
46
67%
53%
RUN HISTORY · one row per nightly build; REC W% and CLV BEAT are cumulative
SLATE
STARTERS
CARDED
TOP EV
REC W%
CLV BEAT
2026-09-04
34
0
—
—
—
2026-09-03
6
1
+5%
100%
—
2026-09-02
24
3
—
67%
—
2026-09-01
30
3
—
67%
—
2026-08-31
23
1
—
100%
—
2026-08-30
25
4
—
50%
—
2026-08-29
22
3
—
33%
—
2026-08-28
30
5
—
80%
—
2026-08-27
11
0
—
—
—
2026-08-26
27
2
—
50%
—
REALITY CHECKThe execution card is capped at two plays. Every displayed app line is repriced from the saved K distribution at that exact threshold, requires model + multi-book sportsbook agreement, and must clear lineup/workload safety. An unconfirmed opposing lineup receives a disclosed probability penalty. Empty is allowed; the model's edge is not yet statistically proven. Bet responsibly.
MODEL OUTPUTS · TRANSPARENTLY TRACKED · NOT BETTING ADVICE. "OUR K%" IS A CALIBRATED PROBABILITY (CAPS ≈ 70% BY DESIGN).
THE MODEL IS AT-MARKET; CLOSING-LINE VALUE IS THE TEST. COLOR KEY · VOLT = OUR MODEL · GOLD = FROZEN/CARDED · BLUE = THE MARKET · ORANGE = OVER · VIOLET = UNDER · GREEN/RED = LIVE & RESULTS ONLY. THIS TAB IS TOMORROW ONLY · LIVE ACTION IS ON THE BOARD.