Home / How we score
Best list
The Editors' score is our own number out of ten to one decimal, attached to every series card on this site. It is built from four measures with fixed weights, each ranked against the whole catalogue. Visitors' votes are separate. This page explains what goes in, what stays out, and why a score can move when we recount.
| Measure | Weight | What it reads |
|---|---|---|
| Story | 30% | Premise, characters and twists, read from the free episodes and the description |
| Heat | 30% | How charged the free episodes are, measured from their subtitles |
| Free value | 25% | How much of the series is free against the tokens to finish it |
| Newness | 15% | Full on release day, then fading |
Story is the biggest single measure at 30 percent. We read the subtitles of the free episodes together with the official description, then ask what a viewer actually gets: are the characters distinct, is the premise clear, does anything twist or escalate, or is it one situation repeated.
This is a judgment, written down as a rating, and we try to be consistent rather than generous. A series with a sharp one-line setup and a reason to watch episode 2 scores well. One that is a name, a bed and little else does not.
We read the free material only. We never guess at what happens in paid episodes, and the cards say when a series left us wanting more information. We also note when the description promises more than the free episodes show, because a gap between the pitch and the pages is worth knowing about before you spend anything.
Heat counts for 30 percent and is measured from the subtitles of the free episodes, not from any picture. The number shows how much heat the free stretch builds, through setup, tension and innuendo. Free episodes of open series are suggestive, so a high heat figure reflects heavy build-up and nothing graphic.
Free value counts for 25 percent. It compares how much of a series you can watch for free with the tokens needed to finish it. A series with two free episodes and a short paid tail does well here. One with two free episodes and a very long paid tail does not.
Together these two measures push the score toward series that deliver something before you pay, which is the point of a buyer's guide. A series can look generous and still score low on value if the paid tail is long, which is exactly the trap we want the number to catch.
Newness takes the last 15 percent. Candy releases series constantly, and a recent title is more likely to still be getting attention, so we give a small edge to what was counted most recently.
Every measure is a percentile, meaning a rank among all the titles we counted. A series sits at, say, the top quarter for story, and that rank is recalculated whenever we count again. If a batch of stronger series arrives, an older one can slide down without a word of its own having changed.
That is the honest trade-off of a relative score. It tells you how a series compares today, not how good it is forever. The latest count is dated November 5, 2026. A title counted only last week gets a small edge, but the edge fades as newer ones arrive, so a high newness score is a snapshot, not a trophy.
Visitors' votes live on the Most Voted list and next to each thumbs-up button. They are never blended into the Editors' score, and the Editors' score never reaches into the votes.
Keeping them apart lets you compare two opinions: ours, measured, and the visitors', tapped. When they disagree, that is useful information rather than a bug.
And we never invent either one. Votes are real taps counted on our server, we add none ourselves, and nobody can buy a score or a place. On the cards you will see both side by side: the Editors' score as a number out of ten, and the thumbs-up as a plain tally from visitors. Neither is adjusted to make the other look right, and a large gap between them is usually the most interesting thing on the page. If we ever change the weights, we will say so here.
Because the score measures story, free heat, free value and newness against the whole catalogue, not your taste. A series can be a favorite and still have a long paid tail or an older release date. Check the thumbs-up count for the visitors' view.
Yes, from the description alone, since they have no public episodes. Their story and heat are estimates, so treat a Spicy score as a rougher guess than an open series score.
No. Nobody can buy a score or a place, and Candy has no input on our numbers. The partner code on Play links is the only income, and it is the same for every series.
Catalog and Editors' scores from the Candy count of November 5, 2026. Premium prices checked October 30, 2026. Visitors' votes open; counts are written here every week.
Candy Top Shorts reviews adult AI series and links to Candy AI, a service for adults. Continue only if you are of legal age where you live.