How to evaluate scanner alerts and stale odds

A scanner alert is a claim: "this price is available, and it's better than it should be." Many alerts fail that test by the time you act. The price has already moved, the market is suspended, or the sportsbook won't take your stake. How often that happens, for your books and markets, is the most important thing to measure before paying for a scanner.

What "stale odds" means

A price is stale when a sportsbook hasn't yet updated it after the rest of the market moved. Stale prices are exactly what +EV and arbitrage scanners hunt for, and they're short-lived by nature: either the sportsbook updates the price, or it notices that people are taking it.

There's a second kind of staleness: the scanner's data can be behind the sportsbook. If the scanner shows a price the sportsbook has already changed, the alert was never real.

Researchers who used their own odds dashboard in a real-money betting experiment ran into the second problem: odds on the dashboard updated with a delay, so the games they could actually bet on were only a subset of the opportunities their strategy identified. They also missed many opportunities simply because nobody was watching the dashboard around the clock.^[1]

Why an alert might not be bettable

• The price moved between the scan and your click.
• The scanner was behind the sportsbook's real price.
• The market was suspended, common in live betting.
• Your stake was cut because of account limits or market maximums.
• The bet was held for review and rejected or repriced.
• It's an error price. House rules can let a sportsbook void, or resettle, bets at odds "materially different from those available in the general betting market" (BetMGM's Massachusetts wording), which describes a stale price exactly.^[2] In Massachusetts, obvious errors must be defined in the operator's internal controls.^[3]

A simple alert log

During a trial, record every alert you try to act on:

After a week or two, two numbers fall out: fill rate (the share of alerts you could place in full at the shown price) and slippage (how much worse the price was when you got on, on average). Both go straight into the break-even calculator.

Signs an alert is likely to fail

• The price is far better than every other sportsbook: either a big opportunity or an error.
• The market is niche or low-limit.
• The alert is for a live market and you're not already looking at that game.
• The same sportsbook has rejected or repriced several of your recent bets.

What this means for choosing software

Vendors describe speed in general terms. What you can verify yourself is how many alerts you could act on, at your books, in your markets. A cheaper tool with a higher fill rate for your situation can be worth more than a faster tool covering books you don't use. See What to check before paying for an arbitrage scanner.