After uploading a game file, our solver analyzes every decision point to compute the game-theoretically optimal strategy. This guide explains what you'll see and how to interpret the results.

## What is a "Decision"?

A **decision** is any point in the game where a player must choose between multiple actions. In poker, this includes:

- Whether to fold, call, or raise preflop
- How much to bet on each street
- Whether to bluff or value bet on the river

Each decision point receives its own analysis from our solver, showing you the optimal strategy.

## Navigating to Your Analysis

1. Go to your **Plays** list at [mieza.ai/gto](https://mieza.ai/gto)
2. Imported plays appear with their game name and a status indicator:
   - **Imported** — Parsed and saved, waiting for analysis
   - **Analyzing** — Our solver is computing optimal strategies
   - **Analyzed** — Analysis complete, ready to review
3. Click on any analyzed play to see its full history and solver results

## What You'll See

### Game History Tree

Each play is displayed as a tree of game states. Starting from the initial deal, you can follow every action that occurred. The tree visualization shows:

- **Nodes** — Each state of the game
- **Edges** — Actions taken (fold, call, raise, deal, etc.)
- **Decision markers** — Highlighted nodes where the solver has computed an optimal policy

### Solver Decisions

At each decision point, you'll see:

- **Actions available** — The set of legal actions at that point
- **Optimal policy** — A probability distribution over actions. For example, "Raise 65%, Call 35%" means the game-theoretically optimal play mixes between raising and calling
- **Recommendation** — The single best action and the solver's confidence level
- **Thinking time** — How long the solver spent computing this decision

### Understanding Mixed Strategies

Game theory often recommends **mixed strategies** — playing different actions with specific probabilities. This isn't random; it's the mathematically optimal way to prevent opponents from exploiting you.

For example, if the solver says "Bet 70%, Check 30%" at a particular spot, it means:
- Betting is slightly preferred, but always betting would be exploitable
- The exact mix makes your opponent indifferent between their responses
- Over many similar situations, mixing this way maximizes your expected value

## Key Metrics

- **Decision count** — How many decision points were analyzed in the play
- **AI Score** — An aggregate measure of how closely the actual play matched optimal strategy (when available)
- **Expected Value** — The long-run value of the optimal strategy at each decision point

## Tips for Interpretation

1. **Focus on big decisions first** — Large pot situations have the most impact on your results
2. **Look for patterns** — If the solver frequently disagrees with your play in similar spots, that's an area for improvement
3. **Mixed strategies are normal** — Don't be alarmed if the solver recommends mixing; it's mathematically optimal
4. **Context matters** — The solver assumes game-theoretically optimal opponents. Against weaker opponents, exploitative deviations may perform better