ScoresAndStats review: free predictions with the price beside each call

What a prediction looks like

Each game gets a card with the spread, the total and the moneyline, and a one-line call. The Brewers at the Dodgers on 14 October read: "We make them 47.5% to win in Los Angeles; the price says 39.0%", and the call was the Brewers on the moneyline at +145. Many cards say "no bet", with the reason: the Stars at +120 had a gap of 3.9 points against the price, "short of the 5 a bet takes".

The site says how it builds its numbers, sport by sport. In the NFL it turns both teams' results into points against the spread. In baseball it sets a chance of winning against the price at a named sportsbook "with its margin taken out". In hockey each team's home and road record becomes a chance of winning. College football needs a gap of a field goal or more before it bets.

Most of the time the line is about right, so "no bet" is often the true answer for a game. A model with a fixed threshold is also easy to hold to account, because it can't decide after the fact which games it really liked.

What it can't tell you is whether the number is any good. A chance of winning built from home and road records is a simple model, and simple isn't wrong, but its quality shows only in results. That brings me to the record.

The record you'd want

The site says every prediction "stays on its page after the game, exactly as published" and is "never rewritten once the result is in, so you can judge us on the record". The FAQ says the prices on each card "carry the date and time they were read". It defines a unit as "one standard stake, used for every bet". All of that is how a record should be kept.

But on the pages I read, I didn't find the record added up: no season won-lost, no units, no closing line comparison. To judge it you'd have to open every past card one at a time. The method is laid out; the result isn't. Until it is, the record is the site's to publish and nobody else has checked it, which means it isn't ATB Verified.

If you follow it, I'd keep that tally myself. Note the price on the card, the time it was read, the price you got and the closing price. The closing line value guide explains why that last number beats a won-lost record over a short run.

The Premium page

Here the site contradicts itself. The home page says "there is no subscription and no sign-up wall". The FAQ page, titled "ScoresAndStats Premium Support Page", describes a subscription with consensus reports at $9.99 a month or $99 a year, add-ons for AI picks and handicappers' picks, and all three for $45 a month or $350 a year. It says there are no refunds. The terms describe subscriptions that renew automatically. The FAQ's sign-up link sends you to the free MLB predictions page.

It may be that the membership has ended and the FAQ page hasn't caught up. The site doesn't say, and I won't guess. If you're offered a ScoresAndStats subscription, ask what it includes, whether it renews and what the refund terms are, in writing. The home page also carries advertisements for other paid picks services at $19.95 and $29.95 for 30 days. Those are ads, and they aren't covered by this review.

Using it

As a free second opinion with its working shown, it costs you nothing to read. If you're curious how a number like this is built, what a sports betting strategy is is the place to start. The site's own FAQ says no one can guarantee winners, and it's right: no prediction takes the risk out of a bet. Set your loss limit first, and bet each pick at the same small share of your bankroll.

When the gap is small, it says no bet. I'd want the totals on the page too.