Clear rules, satisfying loop, fair difficulty, meaningful decisions
Not here: Ads, paywalls, crashes, amount of content
METHODOLOGY · score-model-v1 · PUBLISHED NOV 8, 2026
How we test mobile puzzle games
Every score on this site comes from hands-on play, a written observation log and saved evidence. Here is exactly how a number is produced — and what it can’t tell you.
01
The formula
S = 0.30G + 0.20A + 0.20F + 0.15C + 0.15U
Weights sum to 100%. Displayed to one decimal, rounding half away from zero: 7.25 → 7.3, 7.24 → 7.2.
How forced ads interrupt, when they appear, whether state survives, whether rewarded ads are truly optional
Not here: Ad count alone is never converted into points
Can you progress without paying? Are IAPs transparent? Any sales-driven difficulty spikes?
Not here: Ad experience, repetition
Do new mechanics or modes change how you solve over time?
Not here: Raw level count
Hit targets, mis-taps, performance, crashes, non-ad resume
Not here: State lost after an ad (counted under Ads)
Word labels, for a quick read
Outstanding
Great
Good
OK
Weak
02
Score anchors
Anchors calibrate judgment; they are not a checklist. In-between scores are allowed when the editor can say which anchor it sits closer to. Example: Ads.
0
Ads make the game basically unplayable, or break the level mid-solve.
2
Frequent forced ads wreck the rhythm, or repeated resume problems.
5
Typical casual level: mostly between levels, noticeable but finishable.
8
Few forced ads at natural breaks; rewarded ads clearly optional.
10
Verified online: no forced ads. Airplane mode never counts.
03
Missing data & double counting
Missing is never 0 or 10
If a dimension lacks evidence it stays empty. No composite score is calculated, and the game can’t enter numeric rankings.
One event, one dimension
A mid-level ad that resets the board is counted under Ads only — not again under Stability or Gameplay.
Initial vs. full reviews
Initial reviews check genre and playability. Only full reviews (3+ sessions over 2–3 days) get a composite score.
04
Tags: yes, no, or not verified
Tags describe play conditions. They help filtering but never change the score. Need-based lists accept only “yes” — “not verified” never passes.
Offline play YES One-handed NO No forced ads NOT VERIFIED YET
05
Rankings & ties
- Filter first, then sort by displayed score. No gaps left behind.
- Equal displayed scores share a rank; the next rank is skipped (1, 2, 2, 4).
- A list is called “Top N” only when at least N games qualify.
- iOS and Android are scored separately and never averaged together.
- Reviews older than 90 days, or after a major unretested update, leave numeric lists.
06
Limits & independence
What we can’t verify
We don’t buy in-app purchases. Late-game monetization is listed as a boundary.
Ads vary by region, account and network. Tester region is always recorded.
We never describe a test as “US IP” unless it was.
What money can’t buy
Ads, affiliates and sponsors never change scores, filters or ranks.
Sponsored placements are labeled next to the content.
Every manual override is logged with a reason.