Exercises: Home Field Advantage Deep Dive
Exercise 1: Historical HFA Calculation
Calculate home field advantage from the following season data:
| Team | Home Record | Home Margin | Away Record | Away Margin |
|---|---|---|---|---|
| A | 7-1 | +8.5 | 4-4 | +1.2 |
| B | 5-3 | +3.2 | 6-2 | +4.5 |
| C | 6-2 | +6.0 | 3-5 | -2.0 |
a) Calculate each team's HFA using: HFA = (Home Margin - Away Margin) / 2 b) Which team has the highest HFA? c) Team B has a better away record. Does this affect their HFA interpretation? d) What's the league average HFA if these were the only teams?
Exercise 2: Component Analysis
Break down HFA into components for the following game:
Game: Seattle vs Jacksonville in Seattle, Week 10, 4:25 PM ET
a) Estimate crowd noise component (consider Seattle's reputation) b) Calculate travel component (Jacksonville to Seattle) c) Estimate familiarity component d) Estimate climate component (Pacific Northwest in November) e) Sum the components and compare to standard 2.5 HFA
Exercise 3: Travel Effect Calculation
Calculate travel adjustments for the following road trips:
| Trip | From | To | Timezones Crossed | Direction |
|---|---|---|---|---|
| A | Buffalo | San Francisco | 3 | East → West |
| B | Seattle | New York | 3 | West → East |
| C | Kansas City | Denver | 1 | East → West |
| D | Miami | New England | 0 | Same zone |
Use the formula: Travel_Adj = 0.2 × timezones + 0.3 (if west→east) + 0.15 (if distance > 2000 mi)
Exercise 4: Divisional Game Analysis
A team plays 6 divisional games and 10 non-divisional home games:
Divisional: 4-2, +5.0 margin Non-Divisional: 8-2, +9.5 margin
a) Calculate implied HFA for divisional games b) Calculate implied HFA for non-divisional games c) What's the divisional game penalty? d) Why might divisional HFA be lower?
Exercise 5: Dynamic HFA Model
Build a complete HFA estimate for this game:
Game: Kansas City vs New England at Arrowhead Context: - Week 14, Sunday Night Football - New England traveling from East coast - Temperature: 28°F - KC ranked as high-HFA venue (+0.8) - Not a divisional game
Calculate total HFA using all applicable adjustments.
Exercise 6: Playoff HFA Analysis
Compare regular season and playoff HFA:
A home team has: - Regular season home record: 6-2 (margin: +4.5) - Playoff home record: 3-0 (margin: +12.0)
a) Calculate regular season implied HFA (assume 2.3 baseline) b) Calculate playoff implied HFA c) What's the playoff boost? d) Is 3 games enough sample for playoff conclusions?
Exercise 7: 2020 COVID Analysis
Given the following 2020 data:
- Games with fans: Home win 55%, margin +2.8
- Games without fans: Home win 50%, margin +0.5
a) Calculate the difference in HFA b) What does this suggest about crowd effects? c) How should we adjust 2020 data in historical models? d) Did 2020 data change our understanding of HFA components?
Exercise 8: Stadium Design Impact
Compare enclosed vs open stadiums:
| Stadium Type | Home Win % | Avg Margin |
|---|---|---|
| Enclosed indoor | 58% | +3.2 |
| Open outdoor | 54% | +2.1 |
| Retractable | 56% | +2.6 |
a) Calculate approximate HFA for each type b) What causes enclosed stadiums to have higher HFA? c) Should retractable roof decisions affect your HFA estimate? d) Design a model that accounts for stadium type
Exercise 9: Thursday Night Special
Thursday Night Football data shows:
- Away team coming off bye: Home win 52%
- Away team playing back-to-back: Home win 62%
a) Calculate implied HFA for each scenario b) What's the short-rest penalty for the away team? c) How should you adjust HFA for TNF? d) Does home team rest situation matter equally?
Exercise 10: Rivalry Game Analysis
For a high-intensity rivalry (e.g., GB vs CHI):
Historical data at Lambeau: - Standard opponents: Home margin +8.5 - vs Chicago: Home margin +6.2
a) Is HFA lower against rivals despite intensity? b) What factors might explain this? c) How would you model rivalry effects on HFA? d) Does familiarity outweigh crowd energy?
Exercise 11: Time of Day Effects
Calculate HFA by game time:
| Kickoff (ET) | Home Win % | Sample Size |
|---|---|---|
| 1:00 PM | 55% | 150 games |
| 4:00 PM | 54% | 80 games |
| 8:00 PM (SNF) | 58% | 40 games |
| 8:15 PM (MNF) | 57% | 35 games |
a) Calculate implied HFA for each time slot b) Why might primetime games have higher HFA? c) Is the sample size sufficient for 8 PM games? d) How would you incorporate time of day into your model?
Exercise 12: West Coast Teams
Analyze west coast team home performance:
| Team | Home Win % | vs East Teams |
|---|---|---|
| SEA | 72% | 80% |
| SF | 65% | 70% |
| LAR | 55% | 62% |
| LAC | 48% | 52% |
a) Calculate the "East Coast penalty" for each b) Why do west coast teams benefit from eastern visitors? c) How does game time interact with this effect? d) Model the west coast home advantage
Exercise 13: Altitude Effect (Denver)
Denver's Mile High Stadium:
- Home record vs sea-level teams: 68%
- Home record vs mountain teams: 58%
- Overall home record: 64%
a) Calculate altitude-specific HFA boost b) How would you identify "acclimated" visiting teams? c) Is one home game enough for acclimation? d) What's the appropriate altitude adjustment for Denver?
Exercise 14: Model Validation
You built a dynamic HFA model. Test against actual results:
| Game | Your HFA | Static HFA | Actual Margin |
|---|---|---|---|
| 1 | 3.5 | 2.5 | 7 |
| 2 | 1.8 | 2.5 | -3 |
| 3 | 4.0 | 2.5 | 10 |
| 4 | 2.2 | 2.5 | 1 |
| 5 | 3.2 | 2.5 | 5 |
a) Calculate MAE for your model b) Calculate MAE for static model c) Did dynamic HFA improve predictions? d) What's the value of the improvement in spread terms?
Exercise 15: Market Comparison
Compare your HFA to market-implied:
| Game | Your Rating Diff | Market Spread | Your HFA | Market Implied HFA |
|---|---|---|---|---|
| A | -1.0 | -3.5 | 3.2 | 2.5 |
| B | +2.0 | -1.0 | 2.8 | 3.0 |
| C | 0.0 | -4.0 | 3.0 | 4.0 |
a) Calculate the discrepancy for each game b) Which games show potential value? c) Is the market systematically different from your model? d) How would you track HFA disagreement over time?
Exercise 16: Building Team HFA Database
Design a system to track team-specific HFA:
a) What data would you collect? b) How many seasons of data do you need? c) How would you handle team relocations? d) How would you detect changes in HFA over time?
Provide a database schema and key queries.
Exercise 17: Super Bowl Analysis
The Super Bowl is played at a neutral site:
- Designated "home" team: Slight uniform advantage
- Actual travel/rest: Often similar
Given historical Super Bowl data where designated home teams win 52%:
a) Is there any meaningful HFA? b) What might explain the slight edge? c) Should your model include Super Bowl HFA? d) How do you handle the "home" designation in predictions?
Exercise 18: Multi-Factor Scenario
Calculate complete HFA for:
Game: Bills @ Chiefs, AFC Championship Context: - Playoff (Conference Championship) - Kansas City high-HFA venue (+0.8) - Temperature: 18°F - Buffalo cold-weather team - Sunday, 6:30 PM ET - Bills traveled 1,100 miles, 1 timezone
Build the complete HFA estimate with all adjustments.
Exercise 19: HFA Trends
Analyze HFA trend over five seasons:
| Season | Home Win % | Avg Margin |
|---|---|---|
| 2019 | 53% | +2.1 |
| 2020 | 51% | +0.7 |
| 2021 | 53% | +2.0 |
| 2022 | 54% | +2.3 |
| 2023 | 54% | +2.2 |
a) What was the COVID impact (2020)? b) Has HFA recovered to pre-2020 levels? c) What's the "new normal" HFA? d) How would you project future HFA?
Exercise 20: Complete HFA Model Implementation
Build a comprehensive HFA function:
Requirements: 1. Accept home team, away team, and game context 2. Calculate base HFA (modern ~2.3) 3. Apply team-specific adjustment 4. Apply travel adjustment 5. Apply context adjustments (divisional, playoff, primetime) 6. Return final HFA estimate
Provide Python implementation with test cases.
Challenge Exercise: HFA Prediction System
Design a system that predicts HFA for all weekly games:
Requirements: 1. Database of team HFA values 2. Context extraction from schedule 3. Dynamic calculation for each game 4. Comparison to market implied HFA 5. Tracking and validation over time
Deliverables: - System architecture - Data model - Key algorithms - Weekly report format
Submission Guidelines
For each exercise: 1. Show calculations step-by-step 2. State assumptions clearly 3. Interpret results 4. Note limitations 5. For programming exercises, include commented code