loader
Game Reviews

Upvotes and Downvotes Across 87 Directory Games

Nov 01, 2026 Game Reviews
Upvotes and Downvotes Across 87 Directory Games

Upvotes and downvotes are the only column in this directory where players register a judgement rather than a behaviour. Visits, favourites and concurrent players all measure what people did; the vote pair measures whether they approved. Across the 87 games here with a full statistics set, players have cast 261,332,006 votes — 234,183,880 up and 27,148,126 down — giving an aggregate approval rate of 89.61%. That headline number is remarkably stable, and the interesting part is how far individual games sit from it.

The ten highest approval rates

#GameApprovalUpvotesDownvotes
1Anime Dice99.04%753,1847,312
2Slayers 297.53%955,52424,203
3+1 Speed Keyboard Escape96.72%10,091,026342,204
4+1 Loot To Forge96.60%175,0486,155
5[⚡] Ride A Pet96.20%468,28518,497
6Death Ball96.06%2,989,394122,477
7Build An Ant Empire96.01%59,4062,472
8Attack on Titan Revolution95.97%1,445,57660,741
9Bee Swarm Simulator95.61%3,046,441140,004
10 [HATCH WARS] Pet Simulator 99! 94.74%3,078,408170,870

Every game in this group clears 94%. Anime Dice leads at 99.04%, on 753,184 upvotes against 7,312 downvotes. The pattern across the whole ten is that they are mid-sized: none of them is in the directory's top five by concurrent players. 9 of the 87 games in the sample clear 95% approval, so this is a common profile rather than an outlier — games that players like tend to be liked emphatically.

The five lowest approval rates

GameApprovalUpvotesDownvotes
Escape Running Head74.94%314,071105,044
Tower of Hell73.63%4,388,6101,571,970
Timebomb Duels [Alpha]65.22%240,481128,243
BARRY'S PRISON RUN! (OBBY)60.71%409,050264,728
[] Adopt Me!27.27%38

The bottom of the table needs a caveat before it means anything. [] Adopt Me! shows 27.27% approval, but on 3 upvotes and 8 downvotes — a total of 11 votes. That is far too small a base to rank anything, and it is listed here only because it is the mathematical floor. The smallest honest entry is BARRY'S PRISON RUN! (OBBY) at 60.71% on 409,050 upvotes, and that one carries a real vote count.

The most-favourited game in the directory is also one of the least approved

BARRY'S PRISON RUN! (OBBY) is the single most-favourited game here with 105,267,054 favourites, and it carries an approval rate of 60.71% — 409,050 upvotes against 264,728 downvotes. That is 28.90 points below the directory average and the widest gap between a favourites rank and an approval rank anywhere in the sample. The honest reading is that favourites and votes measure different moments: a favourite is usually cast once, early, often before a player has formed a strong view, while a downvote is cast by someone who played long enough to be annoyed. A game that is very good at attracting first-time players and less good at keeping them will produce exactly this shape.

Tower of Hell is the second case, at 73.63% on 4,388,610 upvotes and 1,571,970 downvotes — 1,571,970 downvotes is the largest single downvote count in the directory. It is also one of only two games listed on the obby category page. High volume pulls the rate toward the mean in both directions, and a game with millions of votes will show a more moderate figure than a game with thousands, simply because a large audience contains more disagreement.

Sample size is the thing to check first

Approval rate is a proportion, and proportions from small samples move a lot. Of the 87 games here, 6 sit below 80% approval and 9 sit at or above 95%; the middle game by rank is Piggy [SEASON 9 - FRIGHT NIGHT] at 90.33%. Before quoting any of those figures, look at the vote total underneath it. A useful working rule: below roughly 100,000 combined votes, treat the rate as directional; above that, it is stable enough to compare. Filtering to the games with at least 100,000 votes leaves 82 of them, and the ordering at the top of that filtered list is led by Anime Dice at 99.04%.

Three of the top ten by rate — Anime Dice, +1 Loot To Forge and [⚡] Ride A Pet — are listed together on the simulator category page, and each carries a full vote record you can check directly. You can see the raw vote columns for Loot To Forge and Anime Dice on their pages, and the low-approval outlier on the Barry's Prison Run page.

What the aggregate hides

A directory-wide 89.61% approval sounds like a strong endorsement, and in a narrow sense it is. But the votes are not distributed evenly: the games with the largest vote totals contribute almost all of the 261,332,006, so the aggregate is really a statement about a handful of very large titles rather than about the directory as a whole. A game with 10,091,026 upvotes moves the average; a game with 59,406 does not. Read the aggregate as a weighted average, and read per-game rates as the actual signal.

The comparison tables from earlier in the month make the movement visible: the 68-game approval board was computed on a smaller sample, and the most recent ratio table shows the same calculation one sync earlier. Because the sample has grown from 68 to 87 games, the aggregate has drifted even where individual games have not — a reminder that any directory-wide average moves when the directory does.

Downvote counts are absolute, not relative

Sorting by the raw downvote column produces a different list from sorting by rate, and the difference is instructive. The largest single downvote total here belongs to [] Steal a Brainrot at 2,706,471, which is 10.0% of all downvotes cast in the directory — but it sits alongside 15,496,019 upvotes, so its rate is 85.13%, comfortably above the middle of the table. A large game collects a large number of every kind of reaction. Ranking by raw downvotes will always surface the biggest games, which is why the rate column exists.

Restricting to games with at least one million combined votes removes that distortion. 59 games clear that threshold, and their rates run from 96.72% down to 73.63% — a spread of 23.09 points among games that are all, by any measure, successful. Within that band the ordering is stable between syncs in a way that the small-sample end of the table is not, which makes the high-volume group the only place where comparing approval rates between two games is genuinely meaningful.

One more caveat closes the loop: votes are cast on the platform, not on this directory, and the figures here are a sync of whatever the platform currently reports. A game's rate can therefore move because players changed their minds, or because the platform reconciled its counts. Both show up identically in this table, so a single-sync move in a small game's rate is not evidence of anything.

Cover art for the highest-approved game in the directory