Tips & Best Practices

Common Use Cases

Finding Rising Talent

Goal: Identify up-and-coming players in your region.

Steps:

1. Use State View with 3 months timeframe
2. Set Min Entrants: 24 to focus on established brackets
3. Enable Hide outliers to avoid top players dominating the view
4. Look for players in the top-left quadrant
   • High win rate (60-80%)
   • Moderate opponent strength (0.2-0.4)

What you're seeing: Players dominating their current bracket level who may be ready to compete at the next tier.

Follow-up: Check their recent tournament results and consider them for regional PR or invitations to higher-level events.

Analyzing a Weekly Series

Goal: Understand performance trends at your local weekly.

Steps:

1. Switch to Tournament View
2. Enter the series name (e.g., Guildhouse, The Function)
3. Use 30 days to see recent trends
4. Leave Allow multiple series disabled
5. Select the specific series from buttons if multiple match

What to look for:

• Who's improving? (moving up and right over time - requires multiple queries)
• Who's consistent? (high win rate, stable strength)
• Who's challenging themselves? (lower win rate, higher strength)

Pro tip: Run this query monthly and screenshot results to track improvement over time.

Comparing Character Performance

Goal: See how different characters perform in your region.

Steps:

1. Use State View with 3 months
2. Go to Advanced → Character Filters
3. Enter one character (e.g., Fox)
4. Click Apply Filters
5. Note the average cluster position
6. Screenshot the results
7. Reset and repeat for other characters

What you're comparing:

• Average win rate by character
• Average opponent strength faced
• Number of active players per character

Example Analysis:

• Marth mains: 15 players, average 0.35 strength, 58% win rate
• Fox mains: 22 players, average 0.42 strength, 61% win rate
• Conclusion: More Fox players, facing slightly tougher competition

Preparing for a Major

Goal: Research opponents before a major tournament.

Steps:

1. Switch to Tournament View
2. Paste the tournament URL from start.gg
3. Set timeframe to 3 months
4. Look for players in your region or registered for the event

What to study:

• Win rates of potential bracket opponents
• Who they've beaten (check their strength)
• Character matchups you'll face

Pro tip: Cross-reference with the tournament's attendee list to focus on likely matchups.

Evaluating Your Own Progress

Goal: Track your competitive improvement.

Steps:

1. Run a state-level query monthly
2. Find yourself in the results
3. Track these metrics over time:
   • Your win rate (going up?)
   • Your opponent strength (going up?)
   • Your position relative to known players

Healthy progression:

• Win rate stays stable or increases
• Opponent strength gradually increases
• You move from top-left → top-right over months

Warning signs:

• Win rate dropping while strength stays constant (hitting a wall)
• Neither metric improving over 3+ months (need to change training approach)

Advanced Strategies

Finding Your Next Rival

Goal: Identify players slightly better than you to practice against.

Steps:

1. Find your position on the chart
2. Look for players with:
   • 5-10% higher win rate
   • Similar opponent strength (±0.1)
3. These are ideal practice partners

Why this works: Playing people slightly better accelerates improvement without being discouraging.

Regional Meta Analysis

Goal: Understand what's working in your region.

Steps:

1. Character Distribution:
   • Run queries for top-tier characters
   • Count players per character
   • Note win rates
2. Bracket Tier Analysis:
   • Query 1: Min Entrants: 32, Max Entrants: 64 (locals/regionals)
   • Query 2: Min Entrants: 64 (majors)
   • Compare player overlap
3. Trend Detection:
   • Run same query monthly
   • Track changes in player count and positions
   • Spot emerging trends

Tournament Series Comparison

Goal: Compare multiple weekly series.

Steps:

1. Query each series separately (Tournament View)
2. Screenshot each result
3. Compare:
   • Average opponent strength
   • Number of unique players
   • Win rate distributions

Use case: Deciding which weekly to attend for optimal competition level.

Data Interpretation Tips

Don't Over-Index on Single Metrics

Bad: "Player A has 0.01 higher opponent strength, therefore they're better."

Good: "Player A and B have similar strength, but A has 10% higher win rate - A is likely performing better currently."

Why: Small differences in metrics can be statistical noise. Look for meaningful gaps (10%+ win rate difference, 0.1+ strength difference).

Context Matters

Example Scenario:

• Player A: 75% win rate, 0.3 strength
• Player B: 55% win rate, 0.6 strength

Question: Who's better?

Answer: Depends on context!

• Player A dominates locals but hasn't tested themselves
• Player B competes at high level but is still improving
• In 6 months, Player B may surpass Player A

Takeaway: Consider player trajectory, not just current metrics.

Watch for Sample Size Issues

Small sample sizes lead to unreliable metrics:

Red flags:

• Player shows up but you've never seen them
• Extreme metrics (95%+ win rate, very high strength)
• Player has very few events in the timeframe

What to do: Use Min Entrants or Min Large Event Share to filter out low-sample players.

Workflow Optimization

Start Broad, Then Narrow

1. Run a basic state-level query
2. Review the overall distribution
3. Identify interesting clusters or outliers
4. Add filters to zoom in on those areas

Example:

• Initial query: All GA players
• Notice a cluster of mid-level Fox players
• Add filter: Character: Fox, Min Entrants: 32
• Analyze just that group

Use the Reset Button

Don't be afraid to Reset frequently:

• Clears mental clutter
• Prevents filter interactions you forgot about
• Fresh start when exploring new questions

When to reset:

• Switching between unrelated analyses
• Getting zero results and not sure why
• Starting a new session

Save Your Screenshots

Since there's no export feature:

• Screenshot interesting visualizations
• Label them with date and filter settings
• Build a library for longitudinal analysis

Naming convention example:

2024-01-15_GA_State_3mo_Fox.png
2024-01-15_Guildhouse_Weekly_30d.png

Common Mistakes to Avoid

Mistake 1: Comparing Different Timeframes

Wrong: "Player A is better because they had higher win rate in 3-month view than Player B in 30-day view."

Right: Use the same timeframe for all players you're comparing.

Mistake 2: Ignoring Opponent Strength

Wrong: "Player A has 80% win rate, Player B has 60%, therefore A is better."

Right: Check opponent strength. B might be facing much tougher competition.

Mistake 3: Over-Filtering

Symptoms: Getting zero results or fewer than expected.

Solution:

• Remove filters one at a time
• Check your state code spelling
• Increase timeframe

Mistake 4: Treating Outliers as Noise

Wrong: Always hiding outliers.

Right: Outliers can be meaningful:

• Elite players who should be studied
• Statistical anomalies worth investigating
• Data quality issues to report

Rule of thumb: Hide outliers for readability, but check who they are before dismissing them.

Next Steps

• Troubleshoot issues: Troubleshooting
• Understand core concepts: Understanding Metrics