May 5, 2026 · Pre-launch
Picks backed
by data.
Algorithmic Run Line picks for MLB. Validated against 1.5 seasons of historical
odds. Every pick tracked publicly with opening odds, closing line value, and
outcomes. No cherry-picked wins.
Status — May 2026
Currently posting
Run Line picks only, the single market that has
validated against our deployment criteria (positive ROI and CLV across 4 of 4
independent backtest periods). Moneyline and Over/Under remain in development —
see
How it works.
Today's pick
Run Line
Deployed
Public picks begin at launch
No picks yet
Daily Run Line picks will start publishing here once production deployment
completes. Each pick will be posted before games begin with the opening DraftKings
price, then graded automatically.
Run Line
Track record
Deployed
Top Run Line pick per day at 5% edge floor. ~1 pick per day on most game days.
Backtest below covers April 2024 through August 2025. Live record starts at zero
on the day public picks begin.
Backtest performance
April 2024 — August 2025
Backtest
Win rate
59.0%
148W — 103L
Median CLV
+1.80%
vs DK close
| Period |
Dates |
N |
Win rate |
ROI |
CLV |
Live record
Since launch
Live
Public picks begin at launch
0 — 0
The live record starts at zero on the day public picks begin. Every pick will be
posted before games start with opening odds, then graded automatically.
No retroactive edits.
Markets in development
Not yet posting
In Development
Two markets that haven't yet met our deployment criteria. We're tracking them
privately and will add them publicly once they pass the same validation bar
Run Line cleared.
Moneyline
Tracking CLV — ROI not yet validated
The Moneyline model produces a strong closing line value signal across our
backtest data, suggesting it's identifying mispriced lines. ROI hasn't yet
tracked the CLV signal, so we're holding it off the live product until both
indicators meet our deployment bar.
Backtest n: 538
CLV: +2.95%
Over / Under
v1.2.2 development underway
The current totals model uses team-level batting aggregates which haven't
shown a CLV signal in our backtest data. v1.2.2 introduces lineup-confirmed
batter features (per-batter xwOBA indexed by handedness, weighted by lineup
spot) to address this directly.
Backtest n: 620
CLV: +0.00%
Pick history
Every pick, every result
Every Run Line pick will be logged here once public posting begins. Date, matchup,
side, opening odds, closing line value, and result. No retroactive edits. Filter
and sort to verify any claim about the track record.
No live picks yet
History begins at launch
Once daily Run Line picks start publishing, the full pick-by-pick history will
populate here automatically. For backtest performance, see the
Record tab.
Methodology
How this works
True Line is built on a strict pre-commitment discipline: validation rules and
deployment criteria are locked in writing before results are known. The full
criteria are in the deployment criteria doc.
Model
v1.2.1 — logistic regression for moneyline and run line win probability,
ridge regression for totals. Features include
Statcast pitcher rates (K%, BB%, HR/FB%, Barrel%, HardHit%, CSW%),
team-aggregated xwOBA, traditional WHIP, recent form, and park factors.
Re-trained daily on a rolling historical window.
Selection
For Run Line: top 1 pick per day where the model's edge over the opening
DraftKings price exceeds 5%. Some days have no qualifying pick — we don't
force content. Approximately 78% of game days produce a pick.
Closing line value
Every pick is recorded at the opening DraftKings price. After games start,
we measure CLV against the closing price. Positive CLV is the leading
indicator of genuine edge — it means the books moved toward our pick
after we made it. ROI is the lagging indicator and has more variance.
Deployment bar
A market only goes live publicly if backtest data shows aggregate positive ROI,
aggregate positive CLV, AND positive ROI and CLV in every individual test period.
Run Line cleared all five criteria. Moneyline and Over/Under haven't yet — they
stay in development until they do.
Audit trigger
A deployed market is automatically paused if any of these fire on a rolling
100-pick window of live picks: median CLV falls below −1%, ROI falls below
−10%, or both ROI and CLV go negative simultaneously. The pause is publicly
disclosed; the market doesn't resume without re-validation.
Display unit
All public picks shown in units (1u = 1 unit). This is the
standard sports-betting convention, allowing each customer to scale to their
own bankroll. The operator's personal stake size is internal and not part of
the public display.
History
True Line had earlier model versions that are
not reflected in the
live record.
v1.0 (April 2025): Briefly live, paused after we caught a
settlement bug in run line tracking that inflated recorded results.
v1.1 (April 2026): Production-deployed but never validated
against our current criteria. Results were not posted publicly.
v1.2.1 (May 2026): The model deployed here. Validated
against 1.5 seasons of backtest data including a full out-of-sample 2025
test. Modest validation per pre-committed rules — strong CLV signal, modest
ROI improvement over baseline. Run Line met all deployment criteria;
Moneyline and Over/Under did not.
Disclaimer:
Picks are generated by a statistical model for informational and entertainment
purposes only. Past performance does not guarantee future results. Sports betting
involves risk — never wager more than you can afford to lose. If gambling is
causing problems for you or someone you know, call 1-800-GAMBLER.