Review methodology: how TheBettingDude scores apps, sites and tipsters

Every ranked product on this site gets one score out of 10. The rubric has six criteria with fixed weights, but only five are scored today: accuracy joins once pick tracking is running. Nobody can pay to change a score, a rank or a badge. Here is exactly how it works.

The criteria and their weights

Criterion	Weight	What I look at
Accuracy	30%	Not scored yet. Once tracking runs: how the picks did against the odds available when they were posted, over a logged sample. Never the hit-rate on the product’s own landing page.
Transparency	20%	Is there a full, dated record with the losses left in? Is it clear who is behind the product and how picks are made?
Coverage	15%	Sports, leagues and markets covered, how often picks arrive, and how early before kick-off.
UX	15%	How fast you get from opening it to a usable pick. Ads, pop-ups, crashes and dark patterns cost points.
Price	10%	What you get for free, what the paid plan costs, and how easy it is to cancel.
Support	10%	Whether anyone answers, how quickly, and whether the product is still being updated.

Accuracy carries the most weight in the rubric because it’s the reason you’d use a prediction product at all. Transparency is second because a record you can’t check is worth nothing.

Accuracy isn’t scored yet

I haven’t logged any product’s picks yet, so no listing on this site has an accuracy mark. Until tracking runs and a product has a big enough logged sample, the accuracy criterion is left out and its 30% is shared among the other five in proportion. So today every score is built from five criteria, and their weights work out like this:

Criterion	Weight in the rubric	Weight today, with accuracy left out
Transparency	20%	28.6%
Coverage	15%	21.4%
UX	15%	21.4%
Price	10%	14.3%
Support	10%	14.3%

Some listings will never have picks to grade: bookmakers, odds tools, stats sites, podcasts. For them the five-criteria version is permanent. Where a score bar appears, accuracy shows as not rated.

I don’t borrow an accuracy number from the product’s own marketing. If I haven’t logged it, it isn’t scored.

How scores become a mark out of 10

In a full review, each criterion is marked from 0 to 10 and the final score is the weighted average, rounded to one decimal. No full reviews are published yet. Until they are, each ranked listing carries one provisional overall score out of 10, set from desk research against the same five criteria and weights. Marks per criterion are published with the full review, not before.

Score	What it means
9.0 – 10	Outstanding. Hard to find a real fault.
8.0 – 8.9	Very good. Recommended, with minor gripes.
7.0 – 7.9	Good, with flaws you should know about first.
6.0 – 6.9	Usable. Better options exist for most people.
Below 6.0	Not recommended. The verdict on the product’s page says why.

Lists are ordered by score, highest first. Ties are broken alphabetically, not by who asked nicely.

Badges are earned

• Editor’s pick: the top scorer of its type (apps, sites, tools and so on), with no serious issue open against it.
• Best free: the highest-scoring free product of its type. When that’s the same product as the Editor’s pick, it keeps the one badge and no Best free is given.
• Best for beginners: the product that explains itself best, judged on UX and transparency.

Badges can’t be bought. They follow the scores: when a score changes, the badge moves with it. Badges are assigned by hand from the scores, and not every list has all three yet: no “Best for beginners” badge has been awarded so far. Bookmakers get no badges, because bookmakers are described here, never promoted.

Featured placements

Some lists have a block labelled “Featured” above them. Featured blocks are placements picked by the site. They sit outside the numbered ranking and carry no score, and being featured says nothing about how a product compares with the ranked ones. Rankings are never for sale: no product can pay to enter them or to move within them.

Any reader can challenge a score at contact@thebettingdude.com. There is more about the site on the About TheBettingDude page.

What “provisional” and “draft” mean

• Provisional is the state of every listing until a person has confirmed the product’s name, link, price and key facts and the full review is written. Today that means every listing, because no full review has been published yet. The product still shows, because a directory with holes is less useful, but treat the score and the details as unconfirmed. The product’s page says when the full review is still in progress.
• Draft pages were written with AI help and haven’t been signed off by a person. They stay out of search engines until they are.
• I never claim hands-on testing that didn’t happen. If a fact about a real product isn’t known, the field is left empty.

Re-review schedule

This is the schedule I intend to keep. The site is new, so no page has been through a full cycle yet.

Page type	Planned review	Out-of-cycle update when
Category lists	Every 3 months	A notable product appears or shuts down
Reviews (once they exist)	Every 6 months	Price change, app delisted, major redesign
Guides	Every 6 months	A law changes or a new kind of bet appears
Glossary	Yearly	—

Editorial pages show a “last reviewed” date and the author. That date only changes when the content was actually checked.

Money and rankings

No affiliate links, no commission, no paid rankings. When sponsored placement opens, it will be labelled “Sponsored” and sit outside the numbered list. The business inquiries page has the rules.

Corrections

Found a mistake? Email contact@thebettingdude.com with the page and the problem. The aim is to fix factual errors quickly, update the page’s date, and note significant corrections on the page. Opinions stay opinions: a product’s owner can correct a fact, but not a verdict.