How RiftMind Calculates LoL Counters, Matchups, and Synergies

Data Source and Scope

RiftMind uses League of Legends match records collected through Riot data sources and processes them into patch, region, rank, champion, role, matchup, synergy, and build aggregates.

Public champion matchup and synergy pages focus on Ranked Solo/Duo games. Very short games are excluded by requiring at least 600 seconds of game duration.

Champion roles use Riot position fields, preferring team position and falling back to individual position. Only Top, Jungle, Mid, ADC, and Support are included in role-based stats.

Patch Window

Only pairwise matchup and synergy stats use a rolling patch window. The current window covers 3 patches ending at the displayed patch.

Champion role stats, builds, runes, and item summaries are patch-specific aggregates, not rolling patch-window aggregates unless a page explicitly says otherwise.

Games in the pairwise window are currently counted equally. The window improves sample size, but it can smooth over very recent champion, item, rune, or system changes.

Rank Buckets

Rank scope is assigned at the match level from the most common ranked solo tier among participants. If there is a tie, the higher tier wins the tie-break.

Low Elo currently aggregates Silver, Gold, Platinum, and Emerald matches.

High Elo currently aggregates Diamond, Master, Grandmaster, and Challenger matches.

Low Elo starts at Silver. Unranked matches remain outside the Low Elo and High Elo aggregates and are not selectable.

Expected Win Rate

The main counter and synergy signal is observed win rate versus expected win rate. This is intended to reduce the distortion caused by champions that are globally strong or weak in the same patch, region, rank scope, and role.

For enemy matchups, RiftMind combines the champion-role baseline with the opposing champion-role baseline inside the pairwise patch window using a log-odds model. In plain terms, expected win rate estimates how often the champion should win if both champions keep their usual role-adjusted strength.

For allied synergies, RiftMind combines both allied champion-role baselines with the same log-odds approach.

The displayed delta is observed win rate minus log-odds expected win rate, shown in percentage points. Negative matchup deltas are stronger counters; positive synergy deltas are stronger ally pairings.

Sample Thresholds and Sorting

RiftMind shows sample size next to pairwise records. Public pairwise stats currently require at least 500 games.

Counter summaries sort by expected-win-rate delta ascending, so the top entries are matchups where the champion performs worse than expected.

Synergy summaries sort by expected-win-rate delta descending, so the top entries are ally pairings where the pair performs better than expected.

Confidence intervals and formal uncertainty bands are not currently displayed, so close deltas should be interpreted cautiously even when they pass the minimum sample threshold.

How to Read All-Role Counters

Same-role counters are the closest proxy for direct lane or role matchups, such as Akali Mid versus Twisted Fate Mid.

All-role counters include enemy champions in any role. These can reveal draft-level pressure, teamfight tools, jungle influence, scaling patterns, or anti-carry effects, but they should not always be read as direct lane counters.

Use all-role counters as draft context, then cross-check same-role matchups, rank scope, patch window, and sample size before making strong conclusions.