Betting scanner testing methodology: how WagerBench tests scanners

WagerBench tests a betting scanner by logging its alerts on live slates and checking each one against the sportsbook's bet slip, from one disclosed state, on terms any buyer could get. This page is our betting scanner testing methodology: what we measure, how we record it, and what our tests can't tell you.

A scanner review is only as good as the test behind it. Vendors publish profit screenshots, sportsbook counts and speed claims, and almost none of that can be checked from the outside. Setting out the protocol lets you judge the evidence for yourself and hold us to it. Every review links back here, and when a review departs from this protocol, it says so on the page.

What a scanner test is trying to answer

A buyer wants to know five things, and they're what we measure:

1. What it really costs. The regular price, not the first-month promo, plus what the trial requires and how cancellation works.
2. Which sportsbooks it actually covers for you. Not the headline count, but the books you can legally use in the state where we test.
3. Whether its alerts can be acted on. An alert you can't bet at the listed price is noise.
4. How fast it is. How old a price is when it reaches you, and how long it takes to get from an alert to a bet slip.
5. Whether its numbers hold up. For +EV tools, whether the flagged prices beat the closing line more often than chance would suggest.

Features, layout and support come after those five. A beautiful dashboard that surfaces stale prices still loses money.

Why executability comes first

The gap between an alert and a bet that can actually be placed is the problem the whole category runs into. Researchers who ran a value-betting system on real football odds reported that prices had sometimes already changed by the time their own dashboard displayed them.^[1] Commercial odds feeds refresh on a cycle too: The Odds API, a commercial odds aggregator, documents refresh intervals of 60 seconds for pre-match main markets and 40 seconds in-play, with faster rates for exchanges.^[2] Any tool built on polled data is showing you prices that are at least a little old.

So our core test is simple to describe and slow to run. For each alert we log, we open the sportsbook and record whether the bet slip shows the same price, a worse price, or no market at all.

Betting scanner testing methodology: the setup

One state, disclosed. Sportsbook availability is set state by state. Massachusetts, for example, has seven licensed mobile sports wagering operators.^[3] A scanner can't be more useful to you than the books you can open. Each review names the state the testing was done from, and coverage findings apply to that state only.

Same games, same window. When we compare products, we run them side by side on the same slate, at the same time, with the same sportsbook accounts selected. A test that runs one product on Sunday NFL and another on a Tuesday night in the NHL isn't a comparison.

Paid plan, our own account. We test on a plan we pay for, or a trial on the same terms any buyer gets. If a vendor gives us free access, the review says so at the top.

Timestamps on everything. Each alert gets the time it appeared, the market, the listed price, the sportsbook, and the time we reached the bet slip. Screenshots back up the log.

What we record per alert

We don't need to place a bet to learn whether a price was available. When we do place real bets during testing, the review discloses stake sizes and results, and never treats a short run of profit or loss as evidence that a tool works.

Sample size and what it means

A few dozen alerts can't separate a good tool from a lucky week. Each review reports how many alerts were logged, over how many days, and across which sports. Results based on small samples are labeled as such. Closing-line results in particular need hundreds of bets before they say much, and we'll say plainly when a sample is too small to support a conclusion.

Evidence labels

Every claim in a review carries one of three labels:

• Tested by WagerBench. We observed it ourselves, under this protocol.
• Vendor-reported. The vendor says it, and we cite where. We haven't confirmed it.
• Not yet verified. We couldn't confirm it either way.

Prices always come from our dated software database, with the check date shown next to them.

What our tests can't tell you

A test from one state, over a few weeks, on a handful of sports, describes that window. Coverage differs by state, products change every season, and sportsbooks restrict winning accounts, which changes what any tool is worth to a given bettor (see why sportsbooks limit winning bettors). We don't promise profit, and no review should be read as a prediction that a tool will make money for you.

Independence

WagerBench may earn affiliate commissions from some products. Commissions never decide ratings, placement or what gets tested, and the affiliate disclosure lists which programs are active. Vendors can't review a page before it's published, and they can't pay for changes to it.

Retesting

We retest when a product changes its pricing, its plan structure or its core features, and at least once a year otherwise. Each review shows its test dates and the date of the last retest.