Choosing a game to play with other people is mostly a question about servers, not about genres. A game can be excellent and still be a bad choice at 11pm on a Tuesday if only a few hundred people are inside it. The BloxBus directory records a live concurrent-player figure for 78 fully-synced games, and at the moment of this sync those 78 games held 2,829,152 people between them. That single column is the most practical tool the site publishes for answering "will there be anyone else in there with us?"
Everything below comes from that column and its companions - lifetime visits, favourites, upvotes and downvotes. No popularity ranking is inferred from anything else, and where the directory genuinely has no data to support a claim, this article says so rather than filling the gap.
| # | Game | Playing now | Lifetime visits | Approval |
|---|---|---|---|---|
| 1 | Steal An Egg | 674,001 | 9,581,698,613 | 91.7% |
| 2 | Murder Mystery 2 | 227,178 | 31,860,236,990 | 90.8% |
| 3 | Brookhaven 🏡RP | 226,273 | 88,937,606,302 | 85.9% |
| 4 | [🌗] Steal a Brainrot | 215,211 | 74,180,681,525 | 85.1% |
| 5 | 99 Nights in the Forest | 96,237 | 30,276,562,882 | 90.5% |
| 6 | Blox Fruits | 95,013 | 65,115,037,760 | 92.3% |
| 7 | [SKY ASSASSIN] Jujutsu Shenanigans | 89,794 | 7,781,698,779 | 86.3% |
| 8 | RIVALS | 74,641 | 18,981,943,321 | 93.7% |
| 9 | Anime Dice | 70,730 | 139,686,048 | 99.1% |
| 10 | [🎃] Adopt Me! | 65,311 | 728 | 27.3% |
The shape of that table is the first thing worth noticing. The top four games - Steal An Egg, Murder Mystery 2, Brookhaven 🏡RP and [🌗] Steal a Brainrot - account for 1,342,663 of the 2,829,152 concurrent players on the directory, which is 47.5% of everyone currently playing anything tracked here. Extend to the top ten and the share reaches 64.8%. The remaining 68 games split 994,763 between them.
For a group deciding where to go, that concentration is good news. It means the busiest handful of experiences are busy by a margin wide enough that time zone and hour of day barely matter. It also means the long tail is genuinely quiet: a game outside the top twenty will often have fewer people in it than a single busy server inside the top five.
Murder Mystery 2 and Brookhaven 🏡RP are separated by 905 concurrent players - Murder Mystery 2 on 227,178 and Brookhaven 🏡RP on 226,273. That is a gap of well under one percent, and it is inside the range that a single afternoon can overturn. Treat positions two and three as shared. What separates them for practical purposes is not population but texture: Brookhaven 🏡RP carries 88,937,606,302 lifetime visits against Murder Mystery 2's 31,860,236,990, so the two have very different histories despite being identically crowded right now.
This is also the clearest illustration of why a live figure alone is a poor tiebreaker. Concurrency tells you about this instant. Visits tell you about the whole life of the game. Favourites tell you how who wanted a shortcut back. A group picking a game for one evening should weight concurrency heaviest. A group picking a game to commit to for a month should weight the other two.
The directory also records upvotes and downvotes, which together work out to an overall approval rate of 89.6% across 78 games. Within the top ten that rate varies more than the population figures suggest. RIVALS sits at 93.7% on 74,641 concurrent players, while Brookhaven 🏡RP records 85.9% on a much larger 226,273. A gap that wide between two games with comparable traffic is a real signal about how the people already inside feel about the experience.
Approval is not quality, and the directory does not present it as such. It is a vote anyone can cast, it accumulates over years, and a game that has been on the platform a long time has had more chances to collect both kinds. Used as a tiebreaker between two games you already want to play, it is useful. Used as the primary filter, it mostly re-ranks games by age.
Concurrency solves the "will anyone be there" problem, but some groups want the opposite: a server with nobody else in it. The directory currently lists private-server coverage for 12 of its 78 synced games. That is a minority of the catalogue, but those 12 games are not marginal - they include Murder Mystery 2, Brookhaven 🏡RP and [🌗] Steal a Brainrot, three of the four busiest titles on the board. In other words, the games crowded enough to be worth playing are also, disproportionately, the games where you can book a room and shut the door.
The roleplay category is where this matters most, because roleplay is the genre where uninvited strangers change the experience the most. Two of the directory's busiest games sit in that category, and both appear in the private-server list. If your group's plan is "we want to hang out and talk", filtering the roleplay listing by private-server availability first, then by concurrency second, will produce a better shortlist than either filter alone.
Start from the concurrent column and discard anything below the level where servers feel empty to you - for most groups that lands somewhere in the tens of thousands. Then apply one of two modifiers. If you want a public, busy room, keep the highest-approval survivors. If you want privacy, intersect what is left with the private-server list. The directory's own code page for the second-busiest game in the directory and the equivalent page for the third-busiest title both sit inside that surviving set, which makes them convenient places to check whether a game has any currently-listed codes before you commit an evening to it.
Two earlier write-ups on this site cover adjacent ground with their own snapshots: a friends-and-live-data roundup from an earlier sync and a second pass at the same question using live figures. Both are worth reading side by side with this one, because the ordering between them has shifted - which is itself the honest answer to "which game should we play": the board moves, so re-check rather than memorise.
Concurrency says nothing about whether a game supports your group size, whether it splits parties across instances, or whether its matchmaking will keep your friends in the same server. The directory does not record any of that, so nothing here should be read as a claim about it. It also says nothing about the current moment in any strict sense: these figures are a snapshot taken by a scheduled sync, and by the time you read this the numbers will have moved. What stays stable is the structure - a heavily concentrated top, a long quiet tail, and a private-server option concentrated in exactly the games where crowding is worst.