[用語集](https://unicum.gg/ja/glossary.md)/[評価](https://unicum.gg/ja/glossary/category/ratings.md)

# Mode ratingMR

The performance rating unicum.gg uses for the battle modes outside Random Battles, built from average experience per battle and win rate, each read against the median of the people who play that mode.

別名：MR。

[WN7](https://unicum.gg/ja/glossary/wn7.md), [WN8](https://unicum.gg/ja/glossary/wn8.md) and [WNX](https://unicum.gg/ja/glossary/wnx.md) cannot be used there. All three are built against an [expected damage](https://unicum.gg/ja/glossary/expected-values.md) figure measured in random battles, so a Stronghold Skirmish is scored as a random battle nobody could have played: the mode is shorter and tier-limited, which puts a good player's damage at a fraction of their random average while their win rate sits near 90%. Measured on one vehicle, the same player scored 1,794 WNX in Skirmish against 4,708 in Random, at 89.5% wins against 71.9%. The honest answer is not a corrected WN8 but a different question, and MR asks it: how does this player compare to the people who play this mode.

It uses two axes. Average experience per battle is the effectiveness signal, because the game's own experience formula already folds in damage, [frags](https://unicum.gg/ja/glossary/kills-per-battle.md), spotting and survival time, so separate axes would count the same battle twice. Win rate is the outcome. Each is divided by the mode's own median, they are weighted equally, and a volume term then discounts a small sample so a hundred battles cannot top a ranking ahead of someone with thousands.

One last division is what makes the number portable: the score is scaled by the mode's own median score, so 1000 is the median player in every mode. Without it the modes would not be comparable even after normalising the two axes, because the volume term reads a [battle count](https://unicum.gg/ja/glossary/battles.md) and the modes are played in wildly different volumes: measured on EU, the median score ran from 893 in [Clan](https://unicum.gg/ja/glossary/clan.md) Wars VI to 1300 in Skirmish, which would paint an average Skirmish player green and an average player elsewhere orange for the same standing.

What MR does not claim is a comparison with WN8. Those rate a player against the whole playerbase; MR rates them against a self-selected group, and a mode that attracts strong players has a higher bar than one that does not. It says where you stand among the people you actually meet there.

## 関連用語

[前衛戦短いストロングホールドバトル形式で、Tier VI、VIIIまたはXで7対7または10対10で戦います。工業資源のための戦闘です。](https://unicum.gg/ja/glossary/skirmish.md)[勝率プレイヤーが勝利した戦闘の割合であり、個人の出力ではなく結果を測定する唯一の統計値です。](https://unicum.gg/ja/glossary/win-rate.md)[WN8プレイヤーのダメージ、キル、スポット、基地防衛に対する期待値に基づいてスコアを与えるコミュニティパフォーマンス評価で、勝率の項目も追加されています。](https://unicum.gg/ja/glossary/wn8.md)

[World of Tanksの全用語](https://unicum.gg/ja/glossary.md)

## Navigation

- [unicum.gg](https://unicum.gg/ja.md)
- [Discord](https://discord.gg/Hqbfb8YPbU)
- [GitHub](https://github.com/unicum-gg/unicum.gg)

## Indexes

- [llms.txt](https://unicum.gg/llms.txt): the API, the MCP server and the other machine-readable surfaces.
- [llms-full.txt](https://unicum.gg/llms-full.txt): the same, with every endpoint's parameters inline.
- [sitemap.md](https://unicum.gg/sitemap.md): every section of the site, down to the individual pages.