OSD / Studio / Arb Quality

Arb Quality — arb survivability post-move

How Arb Quality works

Arb-flow bets work differently from EV-flow bets. An arb leg is the side of an arbitrage that pays the soft book's inflated price; the candidate's "edge" is the gap between the soft book's price and the consensus (sharp) price. By design, arb candidates sit far from the market — that's the whole thesis.

What this page measures

"Arb quality" is the survivability of an arb's edge after the line moves. An arb at +40¢ that immediately collapses to 0 once sharp money hits is low-quality (you got picked off). An arb at +40¢ that holds steady for 5+ minutes while you place is high-quality. The page replays settled arb positions to measure how often the placed leg actually realized its predicted edge vs. how often it was bait that vanished.

Why it's useful

Identifies arb books that consistently produce vanishing edges (sharps adjust faster than we can place).
Distinguishes "stable arb" patterns (book lags consistently) from "fast-fade" patterns (book corrects within seconds).
Informs arb.minStabilityAgeSec tuning — how long the arb edge needs to persist before we trust it enough to place.

Distinct from alignMaxOddsDiff

alignMaxOddsDiff is the candidate-vs-market distance gate (intended as a stale-line-trap detector for EV flow). Arb flow is the inverse pattern — the candidate IS supposed to be far from market. The 14-day analysis from 2026-04-28 confirmed arb-flow gap is the alpha; the alignment gate's arb-scoped thresholds are set to effective-disable values (400¢) for that reason.

Research/replay tool. Not a runtime gate. See docs/gates.md#align-max-odds-diff for the related candidate-vs-market gate canonical contract.

Data

Remote API Base URL

Remote User ID

Remote API Key

Load remote positions to begin.

Controls

Run Settings

Run Mode

Objective Metric This is the score the optimizer maximizes when ranking candidates.

Minimum Bets for Recommendation Requires at least this many bets for a recommendation to be eligible.

Sweep Setup

Single sweep tests one knob across a range.

Parameter

Explicit Values (comma list, optional)

From

Step

Minimum Sharps Required 0 disables sharp-count gating.

Max Combo Evaluations Run-engine safety cap for Grid and Sharp sweeps (not a segment filter).

Guardrails

Min Distinct Days Requires recommendations to span at least this many settled days. Set 0 to disable.

Max Worst-Day Degradation ($) Rejects recommendations when worst-day profit drops by more than this amount. Set 0 to disable.

Min Green Day Rate (0-1) Requires this minimum share of profitable days for recommendations. Set 0 to disable.

Segment Filters

Sub-selection filters below are applied after Run Mode + Objective Metric.

Flow

Mode

League

Book

Market Family

Data Quality & Output

Allow overrides

Optional Config Overrides JSON (applied last)

Missing Data Policy

Rows to Show

Load remote positions to start.

Summary

Rows Loaded

Settled Matched

Baseline (ROI)

Best Delta

Baseline snapshot: ROI -, Profit -, Win Rate -, Bets -

This is historical replay on bets you already took. It cannot estimate bets that were never placed.

Parameter results are ranked here first. In Auto mode, charts below use the top-ranked parameter row.

Arb Coverage

Arb Rows

ROI Present

Stability Present

ROI Gate Evaluated

Arb ROI Min / Med / P95

Coverage Flags

Run studio to inspect arb ROI/stability data coverage and whether arb gates can move outcomes.

Param	Baseline	Best	Bets	Baseline Score	Best Score	Delta

Run a single-parameter sweep or Auto mode to populate these charts.

Shaded band = top-performing value zone (top 20% by selected objective metric).

Gold dot = best score in this sweep.

Line + dots = metric result for each tested value.

X-axis = tested parameter value.

ROI Curve

Run a single-parameter sweep (or Auto mode) to render ROI curve.

Profit Curve

Run a single-parameter sweep (or Auto mode) to render profit curve.

Win Rate Curve

Run a single-parameter sweep (or Auto mode) to render win-rate curve.

Bet Count Curve

Run a single-parameter sweep (or Auto mode) to render bet-count curve.

Distribution

Rows are tested parameter values. Each row shows aggregate results for that value.

Parameter Value	Bets	Pass Rate	Distinct Days	Green Day %	Worst Day	Win Rate	ROI	Profit	Stake	Avg EV