Odel
rankedtoday mcp

rankedtoday mcp

Local
@ryankramerPythonMITUpdated 1w ago

League of Legends math for AI agents: MMR estimation, climb projections, KDA, hours. No API key.

Ranked Today MCP

League of Legends math and cognitive-readiness reference for AI agents. Estimate MMR from LP flow, project a climb, calculate KDA and lifetime hours, and get the research on sleep and reaction time without the folklore.

No API key. No Riot account. No network calls. Every tool is a pure function and every estimate shows its arithmetic.

Web versions: shouldiplayrankedtoday.com/tools

Install

{
  "mcpServers": {
    "rankedtoday": {
      "command": "uvx",
      "args": ["rankedtoday-mcp"]
    }
  }
}

Or pip install rankedtoday-mcp.

Tools

ToolWhat it does
estimate_mmrInfers whether MMR sits above or below your rank from LP gain/loss asymmetry
climb_calculatorGames required to reach a target LP total, or whether the climb runs backward
kdaKDA with a role-aware read and the winrate caveat
hours_playedLifetime hours from account level, with the math and the error bars
hardware_latency_budgetHow much of a reaction-test score is your monitor and mouse, not you
readiness_referenceWhat the sleep/cognition research supports — and explicitly what it does not
season_countdownDays to a ranked season boundary
rankedtoday_toolsFree cognitive tests and League tools on the web

Design notes

Nothing here pretends to read hidden data. Riot removed public MMR from the API years ago, and there is no lifetime hours counter. estimate_mmr infers from the LP exchange rate and says so; hours_played converts account level and reports plus-or-minus 20% along with what it cannot see. Any tool printing an exact MMR from a summoner name is guessing without telling you.

readiness_reference includes the disconfirming evidence. It lists what the research supports (+49 ms after one all-nighter, 5x more attention lapses, the 516-player rank correlation) and what it does not — including a controlled study where sleep-deprived players' cognition tanked while game outcomes barely moved. Readiness is a probabilistic personal signal, never a win predictor.

climb_calculator will tell you the climb runs backward. Below your breakeven winrate, more games means less LP. It returns that verdict rather than a games-required number, because the honest answer is that this climb does not finish.

Development

git clone https://github.com/RyanKramer/rankedtoday-mcp
cd rankedtoday-mcp
pip install -e .
rankedtoday-mcp

About

Built by Ranked Today — a 3-minute daily cognitive readiness check for League players. Free cognitive tests, no signup.

Not affiliated with or endorsed by Riot Games.

MIT licensed.