Some placements are contested. Rather than quietly changing a number, this site surfaces the pull that reader feedback puts on a game. The pull dictates how far the community wants a particular game to move toward a different rating, and how much agreement is behind it.
Agree or disagree with what you see here? Corrections are welcome on every game. The steadier the pushback on a placement, the more likely it is to be revisited.
Every debated game below reads the same way. Here is an annotated Rocket League example.
The "Suggest a correction" form on any game proposes where its Micro / Meso / Macro levels should sit.
Every correction on a game feeds one combined pull. A "pull" is not really a voting system, but rather a "gravitational" effect. A single suggestion is considered to have a faint pull, and it only strengthens as more feedback points the same way. Volume and agreement both count.
The bar keeps its current placement and starts gravitating toward the proposed one. A locked original from Surnex never moves, but a community requested custom game placement is open to revision as the pull builds.
The aggregated feedback informs the call to make a change on the current placement. Your feedback is gathered and summarized. Received pushback on a placement is taken seriously and can lead to a change, though no fixed number of corrections is guaranteed to enforce a change.
Pull example
Both corrections here want the systems axis at low. The first argues boost is the only thing a match asks you to manage, and also raises the reads for coordinating with a teammate and guessing where the other team will commit; the second leaves the reads alone. The placement counts rotations and boost economy as a real systems layer, and keeps the reads lower because both teams can see the whole field the whole time.
Both corrections here want the systems axis at low. The first argues boost is the only thing a match asks you to manage, and also raises the reads for coordinating with a teammate and guessing where the other team will commit; the second leaves the reads alone. The placement counts rotations and boost economy as a real systems layer, and keeps the reads lower because both teams can see the whole field the whole time.
The optimization side of this was acted on: solving the board is now weighed as planning rather than as reading hidden information, after a correction pointed out there is no opponent and compared it to sudoku. What stays open is the claim that speed and efficiency play amount to an execution layer, which the placement holds is pattern recognition arriving faster rather than finer control of the mouse.
Draft and rotations argued as decisive at a high level of play, which would put the planning layer level with the aiming and the bush reads rather than under them. One of the two corrections also weighs those first two layers slightly lower.
The reads layer draws two corrections down to low. One argues it only shows up in PvP, which few players touch. The other judges the game by its early levels and would lower execution and planning a notch as well.
Two corrections asking for execution and planning layers alongside the reads, one of them arguing that night cycles get routed and that execution carries real weight at a high level. The placement holds door timing to be reaction rather than fine control, and power rationing to be triage inside the night, with no optimization layer standing outside it.
Two corrections push the reads axis up, arguing that a partly filled board is information you work out as you go rather than a plan you bring in. They split on how far, and the milder of the two concedes what the placement rests on: the board is fixed from the start and has one answer, so nothing new arrives to read. The same reasoning is what moved minesweeper's solving onto the planning axis.
A case for a much heavier execution layer, argued from speedrunning, where the run turns on portal placement and carrying momentum through them. The placement reads the game as it is normally played, where a solution guide ends it.
A case that modern fighting games lean more on system and matchup knowledge, raising the strategy layer relative to pure execution.
Managing heavy randomness argued as a real reads layer, since good information would strongly aid consistent runs.
Executing role abilities and managing the town's limited resources argued as light execution and planning layers on top of the social deduction.
A suggestion that route optimization is a lighter factor than a full strategy placement.
Optimal clue-giving argued as a light planning layer for the spymaster role.
A suggestion of a light reads layer in anticipating the opponent's play.
Chase play argued as more execution-driven, with strategy mattering less once both sides already hold the same information.
Matchup knowledge across a very large roster argued as a significant strategy layer, not a negligible one.
A case that fast rolldowns and precise unit positioning add a real execution layer, while most turns are spent optimizing your own board rather than reading opponents.
A suggestion that the reads layer sits below a top-tier placement.
An argument that since chess has no hidden information or randomness, an engine's advantage would hold regardless of opponent, so opponent-specific prep shouldn't count as a real reads layer.
A case that a good run turns on build planning and luck more than on input execution, arguing the micro layer should sit lower.
A proposal to lower the execution weight a notch, filed without a gameplay case to set against the sub-second parry windows the placement rests on.
At the highest difficulties much of the bullet layout is randomized past the point memorization helps, so survival depends on reading the screen and improvising in the moment. The correction places that improvisation on the reads axis.
Trading at high frequency turns on speed of execution, while over long horizons it turns on allocation and planning. The correction argues the category spans all three layers instead of resting on reads alone.
The reads layer here rests entirely on the PvP scene, which this correction holds too marginal to count against a game most people play solo. That layer was already borderline when the placement was made.
A correction weighing the execution layer higher than the placement does. No supporting case came with it, and the placement counts timed crouch-walking and noise management as work in service of staying unseen.
This correction puts most of the gains on learning track layouts and where to spend boost, a planning layer sitting under the high-speed driving.
A nudge upward on the reads axis, with no case attached. The dice and save-scumming were weighed when the placement was made and held low, since a fixed campaign converges once learned and there is no opponent to read.
This correction drops the planning axis to its lowest level and comes with nothing to argue it. Most of the weight in the current placement sits there, on the experience routes and money methods a long account turns on.
A push to raise the systems axis, filed without a case for it. The placement holds that once a matchup and gameplan are learned there is little left to route or manage, since a set resets each round rather than building toward a persistent economy or win condition.
