---
title: "Fight Outcome Model + Optimal Bet Sizing | AimRank"
url: https://www.aimrank.io/products/mma-outcome-model/
description: "A calibration engine built in our lab to be judged against a live market - the scoreboard publishes once a season of logged picks exists, misses included. Sold as an engine for your own outcomes. AI prototype from €2,990."
language: en
updated: 2026-10-09
---

[Products · #23 · Operations & Forecasting](https://www.aimrank.io/products/)

TIER B BLUEPRINT

# Fight Outcome Model + Optimal Bet Sizing

A calibration engine built in our lab to be judged against a live market - the scoreboard publishes once a season of logged picks exists, misses included. Sold as an engine for your own outcomes.

scaffold from **€2,990**

## How it works

Anyone can post winners after the fight. This model writes its probability down before the odds close, bets only when it disagrees with the market, and lets the scoreboard judge.

01

### Features that matter

Reach and stance matchups, takedown defense, age curves, camp quality, short-notice flags, judge tendencies - engineered from fight-level data, not vibes.

02

### Probabilities, not picks

The output is a calibrated probability with an interval - and the calibration is built to be tracked publicly, so 62% has to mean 62% over the season.

03

### Edge or no bet

The model's number is compared to the market's. No meaningful gap - no bet. Most fights are no-bets; the discipline is the strategy.

04

### Size to survive

Fractional Kelly with drawdown constraints and correlation checks - modern bankroll theory. Growth is a side effect of not being ruined.

## Methodology reference

The published method this build is based on, and what it did in production.

**Betfair** publishes the full model → fair-odds → Kelly-stake loop its data-science team uses to size bets.

Probability model + Kelly staking · as reported by [Betfair Quants ↗](https://betfair-datascientists.github.io/wagering/stakingMethods/)

Referenced for methodology only. AimRank is not affiliated with, endorsed by, or a customer of these companies; each links to its own public source. Sources last checked September 2026.

## The Four Guarantees™ - turned on ourselves

The scaffold wires all four in. The PoC proves them on your data, as of that day. Only Implementation keeps them true over time: watched, drift-checked, with recourse.

### Measured value

Closing-line value and Brier score, published every event once the season ledger opens. CLV is the one metric that can't be faked by a lucky streak.

### Defensible

Every pick timestamped before odds close; features and model version logged; losing runs stay on the record. Nothing is deleted, ever.

### Self-improving

Recalibrated after every event; drift on the market's sharpness watched, with an alert the moment quality slips; when the edge decays, the honest output is "no bet" - not a stretch.

### Yours & everywhere

One container, your cloud, full source handover - the same engine pointed at your domain's outcomes, with MCP tools your agents query directly.

## The scoreboard, sized honestly

Illustrative figures - no season ledger has been published yet; when it is, it includes every miss. No bankroll advice is offered or implied.

**CLV** the only scoreboard that survives luck - tracked per pick

**100%** of picks logged pre-fight, timestamped, immutable

**\~61%** no-bet rate - most edges aren't there, and we say so

**0** profit promises - that's the whole point

## Three ways to own it

| Tier | What you get | Price | |
| ---------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------- | |
| **Scaffolding** ★ | The full repo - probability model, calibration tracking, fractional-Kelly sizing engine, eval gate, audit log, MCP server. Trained on public fight data. You own the source. | **€2,990** | |
| **PoC** ★★RECOMMENDED | The engine trained and applied on your domain's outcomes - calibration (Brier, reliability) and decision-sizing quality measured against your historical data. Full code and the quality report are yours. No performance guarantee at this stage, by design. | from**€9,000** | |
| **Implementation** ★★★ | Production: wired to your data feeds, decision sizing live under your risk limits, monitoring + drift alerts, recalibration every cycle, monthly scoreboard - the agreed number (conservative) attaches here. | from **€27,000**\+ from €900/month assurance | |

★ = engagement depth. **PoC is the recommended path**: quality proven on your data before production money. The PoC carries no performance guarantee by design; the agreed number (conservative) attaches at Implementation, informed by the PoC report.

## Born in the lab, built for the harshest market

**The same math runs the sellable catalog.** Calibrated probability + capital allocation under uncertainty is exactly what prices inventory bets in #08 Demand Forecasting, overbooking risk in #11 No-Show Prediction, price moves in #18 Dynamic Pricing, and imbalance exposure in #20 Energy Forecasting. A betting market is just the only client that settles the invoice in real time. When we tell a CFO "our P90 means 90%", this page is how we intend to earn that sentence, in public.

## What this is not

**This is not betting advice, and it never will be.** Selling picks is a licensed activity in most jurisdictions and a responsible-gambling minefield - and "winning system" marketing is exactly the language AimRank bans itself from using. What we sell is the engine, applied to your outcomes; what you will be able to follow, once the season ledger opens, is the research: the model's probabilities, the market's answer, and a scoreboard that includes every miss. Anyone selling you guaranteed profits has already told you everything about their calibration.

## Ready to see your own number?

Request the build: within 48h you get a personal reply with the value sized to your volume.

No commitment · reply within 48h · your data stays in the EU