compare_prob
Judge an external probability (e.g. a Polymarket/Kalshi price) against our sharp fair line — ONE call. Resolves the fixture, de-vigs the sharp book to a fair probability (power de-vig for 3-way 1x2), and reports the edge fair_prob − external_prob in percentage points, the ROI, and a verdict (good...
This record as markdown: /tools/dev-infersports-infersports/compare-prob.md
What compare_prob does on Infersports
AI agents call compare_prob to retrieve information from Infersports without modifying anything. It is typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
| Parameter | Type | Required | Description |
|---|---|---|---|
date | object | — | |
query | string | Yes | |
sport | object | — | |
period | string | — | |
outcome | string | — | |
market_type | string | — | |
external_prob | number | Yes | |
external_label | object | — |
Parameters from the server's own tool schema.
Why compare_prob is rated Low
This tool retrieves and analyzes existing odds data to compare probabilities and calculate edges. It produces informational output (fair probability, ROI, verdict) for decision support but does not execute trades, place bets, move money, modify data, or trigger external financial actions. The emphasis on "DETECTION ONLY" and that "the call is yours" confirms the tool is purely analytical/advisory.
From the tool's definition Tool description states "DETECTION ONLY: InferSports never ingests prediction-market data, sizes a stake, or picks — it gives you the sharp reference and the gap".
Attacks that exploit this kind of access
The rule that runs compare_prob safely
PolicyLayer is an MCP gateway: it sits between your AI agents and Infersports, and checks every tool call against a rule you set before the call runs. Nothing changes on the server itself. For compare_prob, this is the rule to start with:
compare_prob is read-only, so it stays allowed. Everything else on the server is denied unless you say otherwise.
The button opens the PolicyLayer dashboard: create your workspace, connect Infersports, apply this rule, and every compare_prob call is checked against it from then on.
Questions about compare_prob
Judge an external probability (e.g. a Polymarket/Kalshi price) against our sharp fair line — ONE call. Resolves the fixture, de-vigs the sharp book to a fair probability (power de-vig for 3-way 1x2), and reports the edge fair_prob − external_prob in percentage points, the ROI, and a verdict (good / marginal / no_edge). DETECTION ONLY: InferSports never ingests prediction-market data, sizes a stake, or picks — it gives you the sharp reference and the gap; the call is yours. Args: query: natural-language fixture, e.g. "France vs Argentina" or a single team. external_prob: the external implied probability for outcome, in (0,1). Pre-net it for the venue's fee/spread (e.g. a Polymarket YES ask of 0.55 → 0.55). market_type: "1x2" (default; the prediction-market-comparable moneyline), "asian_handicap" (only ±0.5 maps cleanly to a binary), or "totals". period: "full_time" (default) or "half_time". outcome: which leg the probability is for — home/draw/away (1x2), home/away (AH), over/under. external_label: optional source label echoed back, e.g. "polymarket" | "kalshi". sport: optional filter — "football" or "basketball". date: optional UTC date "YYYY-MM-DD" to disambiguate same-name fixtures. Read caveats before acting: a 1x2 fair is regulation 90-min (a prediction market that includes extra time / "to advance" is a different market); quarter/integer AH carries push mass. On an ambiguous query status is "ambiguous" — do not guess. status is "no_line" when no sharp fair is available to judge against. It is categorised as a Read tool in the Infersports MCP Server, which means it retrieves data without modifying state.
compare_prob accepts 8 parameters: date, query, sport, period, outcome, market_type, external_prob, external_label. Required: query, external_prob. The full parameter table on this page comes from the server's own tool schema.
Register the Infersports MCP server in PolicyLayer and add a rule for compare_prob: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Infersports. Nothing to install.
compare_prob is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the compare_prob rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for compare_prob. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
compare_prob is provided by the Infersports MCP server (https://api.infersports.dev/mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
More on Infersports, and thousands of servers like it.
This server
Across the catalogue