Market_Betting — Multi-Track Dashboard

Cross-track overview

Every track’s currently open, still-actionable flags (game/event not yet started, line still live), with each track’s own top 10 shown separately below, ranked by that track’s own edge value — a model probability edge for pick’em/weather/politics/props, or net profit per dollar risked for arbitrage. Tracks are kept separate here deliberately, since a probability edge and an arbitrage profit-per-dollar aren’t really the same unit — see each track’s own tab for the real methodology behind its numbers.

— Open, actionable flags (all tracks)

— Tracks reporting live data

— Best edge right now (any track)

— Long-dated politics rows shown below

🎯 Focus — highest edge, across every track

The largest-edge rows across all five tracks right now, largest first. The only rows left out are ones that literally cannot be placed anymore (game already started, contract already resolving, not fillable, or a legal restriction) — marked ⛔ Blocked elsewhere on this page. Everything else here is ranked purely by edge size: if you trust the model, start at the top.

Track	Opportunity	Side	Venue	Edge	Timing

Track 1 Pick’em — top 10 by edge

Opportunity	Side	Venue	Edge	Timing

Track 2 Arbitrage — top 10 by net profit/$1

Opportunity	Type	Venue	Edge	Timing

Track 3 Weather — top 10 by edge

Opportunity	Side	Venue	Edge	Timing

Track 4 Politics — top 10 by edge

Rows marked Long-dated resolve in more than 60 days — a real edge here is not the same kind of bet as a same-day flag in another track.

Opportunity	Side	Venue	Edge	Timing

Track 5 Props — top 10 by edge

Opportunity	Side	Venue	Edge	Timing

Track 1

Fixed-line pick’em (PrizePicks / Underdog)

— Open flags

— Closed & graded

— Average CLV edge

— Positive-edge rate

Running average CLV-equivalent edge

Running average of clv_edge_at_close across every closed flag, in the order each line closed — not a running sum, which grows without bound as more flags close and stops meaning anything. This is the project’s long-run trendline — expected to include real losing stretches.

Currently flagged opportunities

Open lines from the live hourly pipeline, sorted by model edge. Check exactly two PrizePicks flags to size them as a 2-pick Power Play entry below.

	Player	Game	Stat	Side	Platform	Line	Model edge	Game time

Sizing calculator

Runs the exact same math as scripts/sizing/sizing_engine.py — quarter-Kelly, the PrizePicks account-risk dampener, the same-game caution factor, and the 5%-of-bankroll cap — entirely in your browser. Your bankroll figure is never saved anywhere; it resets the moment you leave or reload this page. This stays exactly as manual as the script: one entry (two legs) at a time, and it has no memory of any stake you've already placed, so update the bankroll figure yourself between entries.

Selected legs

None selected. Check two PrizePicks flags in the table above.

Current bankroll ($)

Recently closed flags

The last 25 lines to resolve, most recent first.

Player	Stat	Side	Platform	Closing line	CLV edge at close	Closed

Could not load data/clv_log.csv. This page reads that file from its own build, so a fresh Cloudflare Pages deploy is needed after any pipeline run for this page to update.

Track 2

Cross-venue arbitrage (Kalshi + Polymarket)

— Flags this run

— Single-venue (YES+NO)

— Cross-venue pairs

— Fillable at flagged size

Currently flagged arbitrage opportunities

Output of the latest automated run of scripts/run_arbitrage_pipeline.py, sorted by net profit per dollar risked, after each venue's real taker fee is subtracted. This table replaces itself completely on every run — it is a point-in-time snapshot, not a running log the way the pick’em CLV chart above is. Sizing a flagged pair is still a manual step (sizing_engine.py arbitrage size) — this page never sizes or places anything.

What this row actually is: the same real-world event, priced on two different exchanges (Kalshi and Polymarket). "Market A" and "Market B" are the opposite sides of that same event — one YES-leaning, one NO-leaning — priced so that buying both together, right now, costs less than $1 combined. If it resolves, exactly one leg pays out $1 and the other pays $0, so the gap between what you paid and that guaranteed $1 is your profit, regardless of which side wins.

Click the "Market A" and "Market B" links below — they open that exact contract on Kalshi, or a live search for its title on Polymarket (Polymarket doesn't have stable per-market URLs to link to directly).
Confirm the title, side (YES/NO), and price on the venue still match this row — prices move between when this page was built and when you look.
Run sizing_engine.py arbitrage size to get a real position size, then place both legs yourself, at both venues — this only works as arbitrage if both legs actually fill. A half-filled position is a directional bet, not arbitrage.

Status	Type	Venue A	Market A	Ask A	Venue B	Market B	Ask B	Net profit / $1	Fillable	Legal footprint

Could not load data/arbitrage_flags_latest.csv. This page reads that file from its own build, so a fresh Cloudflare Pages deploy is needed after any arbitrage pipeline run for this section to update. See Session 3.5's notes if this file has never existed in a deploy yet.

Track 3

Weather / climate markets (Kalshi)

— Open flags

— Closed & graded

— Average CLV edge

— Positive-edge rate

Currently flagged contracts

Open lines from the daily weather calibration pipeline, sorted by model edge. Kalshi is the only venue this track ingests, so there is no cross-venue consensus benchmark here — the closing-price benchmark (a contract disappearing from a fresh pull) is what grades these flags instead, and it works well for this track specifically because weather contracts resolve in days, not weeks or months.

How to find this contract on Kalshi: click the ticker in the last column — it opens that exact market on Kalshi's site. Buy the side shown in "Side" (YES or NO) at or below the price shown in "Mkt price" — that's the edge this row is claiming. This table is a point-in-time snapshot from the last automated run, and this track only has flags when the model disagrees with Kalshi's own price by enough to clear the edge threshold — an empty table means no disagreement was found on the latest run, not that the pipeline is broken.

Status	City	Target date	Forecast	Strike	Side	Market price	Odds	Model edge	Lead time	Find it on Kalshi

Recently closed flags

The last 25 contracts to resolve or drop off the board, most recent first.

City	Target date	Side	Closing price	CLV edge at close	Closed

Could not load data/weather_clv_log.csv. This page reads that file from its own build, so a fresh Cloudflare Pages deploy is needed after any weather pipeline run for this section to update. See this session's handoff notes if this file has never existed in a deploy yet.

Track 4

Down-ballot politics (Kalshi + Polymarket)

— Open flags

— Closed & graded

— Average CLV edge

— Avg. time to resolution

Currently flagged races

Open positions from the daily politics pipeline, sorted by model edge. These are long-dated positions — a race can sit open for weeks or months before it resolves, unlike pick’em or arbitrage flags, which close same-day. The Time to resolution column is the whole point of this track’s frontend card: it tells you, at a glance, how long your capital would stay tied up if you took this position. Rows waiting longer than 60 days are highlighted.

Status	Candidate	Party	State	Race	Venue	Market price	Odds	Model edge	Time to resolution

Recently closed flags

The last 25 races to resolve or drop off the board, most recent first.

Candidate	Party	State	Venue	Closing price	CLV edge at close	Closed

Could not load data/politics_clv_log.csv. This page reads that file from its own build, so a fresh Cloudflare Pages deploy is needed after any politics pipeline run for this section to update. See this session's handoff notes if this file has never existed in a deploy yet.

Track 5

Sportsbook player props (DraftKings + FanDuel + BetMGM)

— Open flags

— Closed & graded

— Average CLV edge

— Positive-edge rate

Currently flagged props

Open lines from the 3-hourly DraftKings/FanDuel/BetMGM pipeline, sorted by model edge. BetMGM's real data is sourced via Rotowire, not a direct BetMGM pull — see SESSION_LOG.md's Session 6.9 entry. The Risk column is this track's own account-limiting-risk indicator: every row carries an Acct. limit risk badge, because sportsbooks limiting or banning consistently-winning bettors is this project's own best-corroborated risk for this venue type (PROPS_PLATFORM_RISK_MULTIPLIER, applied equally across all three platforms — no source in this project distinguishes any of them). A second Field vig badge appears only on rows whose own price still includes that platform's one-sided field vig (implied_prob_includes_field_vig) — a real per-row signal, unlike the platform-level badge, which is the same on every row today.

Status	Player	Team	Stat	Side	Platform	Line	Odds	Model edge	Game time	Risk

Recently closed flags

The last 25 props to resolve or drop off the board, most recent first.

Player	Stat	Side	Platform	Closing price	CLV edge at close	Closed

Could not load data/props_clv_log.csv. This page reads that file from its own build, so a fresh Cloudflare Pages deploy is needed after any props pipeline run for this section to update. See this session's handoff notes if this file has never existed in a deploy yet.