Chapter 9: Quiz

Instructions

This quiz assesses your understanding of Expected Threat (xT) and ball progression metrics covered in Chapter 9. Select the best answer for each question. Answers are provided at the end.

Section A: Fundamentals (Questions 1-10)

Question 1

Expected Threat (xT) measures:

A) The probability of scoring from a shot at a given location B) The probability that possession at a location leads to a goal in subsequent actions C) The number of passes required to reach the penalty area D) The difficulty of making a pass from one zone to another

Question 2

Why was xT developed as a complement to xG and xA?

A) xG and xA were inaccurate B) xG and xA only credit endpoint actions (shots and assists), ignoring build-up contributions C) xT is faster to calculate D) xT requires less data

Question 3

In a typical xT grid, which area has the highest xT value?

A) The center circle B) Own penalty area C) Central area inside opponent's penalty box D) The wings in the final third

Question 4

The value iteration algorithm for calculating xT works by:

A) Randomly assigning values until they stabilize B) Using gradient descent on a loss function C) Iteratively updating zone values based on shooting probability and transition probabilities D) Training a neural network on historical data

Question 5

A pass from a zone with xT = 0.02 to a zone with xT = 0.08 adds how much xT?

A) 0.02 B) 0.06 C) 0.08 D) 0.10

Question 6

What happens to xT added when a player loses possession?

A) It becomes zero (no credit or penalty) B) It becomes negative (the starting zone's xT is lost) C) It remains positive (the attempt still had value) D) It's calculated differently using a separate model

Question 7

A pass is considered "progressive" under the standard definition if it:

A) Moves the ball forward by at least 10 meters B) Moves the ball at least 25% closer to the opponent's goal C) Is completed successfully D) Enters the final third of the pitch

Question 8

Which statement about progressive carries is correct?

A) Progressive carries are always more valuable than progressive passes B) Progressive carries measure ball movement while a player dribbles forward C) Progressive carries only count inside the penalty area D) Progressive carries cannot be calculated from event data

Question 9

The main difference between xT and VAEP is:

A) xT uses a grid while VAEP uses machine learning on action features B) xT requires tracking data while VAEP uses event data C) VAEP only measures offensive value while xT includes defense D) VAEP is simpler to implement than xT

Question 10

Expected Possession Value (EPV) differs from xT primarily because:

A) EPV uses a coarser grid B) EPV incorporates player positions from tracking data C) EPV only works for set pieces D) EPV ignores ball location

Section B: Calculation and Implementation (Questions 11-18)

Question 11

When building an xT grid from scratch, the initial values should be set to:

A) Zero for all zones B) Shot probability × goal probability (conversion rate) C) Random values between 0 and 1 D) Equal values across all zones

Question 12

A 12×8 xT grid divides the pitch into how many zones?

A) 20 B) 48 C) 96 D) 192

Question 13

The transition matrix in xT calculation represents:

A) The probability of a team scoring from each zone B) The probability of moving from one zone to any other zone via passes/carries C) The distance between zones D) The time taken to move between zones

Question 14

Why might you choose a finer grid resolution (e.g., 24×16) over a coarser one (e.g., 12×8)?

A) Finer grids are faster to calculate B) Finer grids capture more spatial nuance but require more data C) Finer grids are more accurate with small datasets D) Coarser grids are always inferior

Question 15

When calculating xT per 90 minutes, you should:

A) Divide total xT by number of matches B) Divide total xT by (minutes played / 90) C) Multiply xT by 90 and divide by total actions D) Sum xT only from the first 90 minutes

Question 16

To calculate a player's xT added from a carry, you need:

A) Only the end location of the carry B) The start and end locations of the carry C) The carry distance in meters D) The time spent on the carry

Question 17

The convergence criterion in value iteration typically stops when:

A) A fixed number of iterations have been completed B) The maximum change in any zone's value falls below a threshold C) All values become positive D) The total sum of all values equals 1.0

Question 18

When comparing xT between different data providers, you should be aware that:

A) All providers use identical grid definitions B) Coordinate systems and event definitions may differ C) xT values are standardized across the industry D) Provider differences don't affect xT calculations

Section C: Player Evaluation (Questions 19-24)

Question 19

A central midfielder with xT per 90 of 0.35 compared to a winger with 0.40. The most accurate interpretation is:

A) The winger is definitively better at ball progression B) The values should be compared within position groups due to role differences C) The midfielder is underperforming D) xT per 90 comparisons across positions are always valid

Question 20

Which player profile would typically show the highest xT generation?

A) A defensive midfielder focused on ball retention B) An attacking midfielder who plays through balls and final third passes C) A goalkeeper D) A center-back on a long-ball team

Question 21

A player has high xT added but low assist totals. This likely indicates:

A) The player is unlucky with teammate finishing B) The player progresses the ball well but doesn't create final chances C) The xT model is incorrectly calibrated D) The player should shoot more

Question 22

Progressive pass distance is calculated by:

A) Summing the length of all successful passes B) Summing the forward component of progressive passes C) Counting progressive passes and multiplying by average pass length D) Measuring the longest single pass

Question 23

When scouting for a ball-progressing center-back, which metrics combination is most relevant?

A) Goals and assists B) Progressive passes per 90, progressive carries per 90, xT per 90 C) Pass completion percentage alone D) Aerial duels won

Question 24

A player's xT decomposition shows 70% from passes and 30% from carries. This suggests:

A) The player should dribble more B) The player generates value primarily through passing rather than carrying C) The model is biased toward passes D) Carries are not important

Section D: Team Analysis (Questions 25-30)

Question 25

A team generates most of their xT from the final third. This suggests:

A) They struggle in build-up play B) They play a direct style, bypassing midfield C) They are effective at getting the ball to dangerous positions D) Their defenders can't pass

Question 26

xT differential (xT generated - xT allowed) correlates with:

A) Possession percentage B) Match results C) Number of corners D) Yellow cards

Question 27

Comparing two teams: Team A generates 1.5 xT per match but creates 12 xG. Team B generates 2.0 xT but creates 10 xG. The best interpretation is:

A) Team A is more efficient at converting xT to xG B) Team B's xT model is broken C) Team A relies more on set pieces or individual brilliance in the box D) xT and xG are unrelated

Question 28

When analyzing transitions, xT is particularly useful for:

A) Set piece analysis B) Identifying how teams generate threat on counter-attacks C) Measuring goalkeeper performance D) Calculating expected saves

Question 29

A team shows low xT from the defensive third but high xT from the middle third. This pattern suggests:

A) The team can't play out from the back but creates well once in midfield B) The team's defenders are poor C) The xT grid is poorly calibrated D) The team never has the ball in their own third

Question 30

To compare xT generation across different leagues:

A) Raw xT values are directly comparable B) Adjustments may be needed for league quality and style differences C) Only EPV should be used for cross-league comparison D) xT cannot be calculated for different leagues

Answer Key

1. B - xT measures the probability that possession at a location leads to a goal
2. B - xG/xA only credit endpoint actions, ignoring build-up
3. C - Central area inside the penalty box has highest xT
4. C - Value iteration updates values based on shooting and transition probabilities
5. B - xT added = 0.08 - 0.02 = 0.06
6. B - Lost possession results in negative xT (losing the starting zone's value)
7. B - Progressive passes move ball 25%+ closer to goal
8. B - Progressive carries measure forward ball movement while dribbling
9. A - xT uses a grid, VAEP uses ML on action features
10. B - EPV incorporates player positions from tracking data
11. B - Initial values = shot probability × conversion rate
12. C - 12 × 8 = 96 zones
13. B - Transition matrix shows movement probabilities between zones
14. B - Finer grids capture more nuance but need more data
15. B - xT per 90 = total xT / (minutes / 90)
16. B - Carry xT needs both start and end locations
17. B - Convergence when maximum change falls below threshold
18. B - Coordinate systems and event definitions may differ between providers
19. B - Position-specific comparisons are more valid
20. B - Attacking midfielders typically generate highest xT
21. B - High xT but low assists indicates good progression but few final chances
22. B - Sum the forward component of progressive passes
23. B - Progressive passes/carries per 90 and xT per 90 are most relevant
24. B - Player generates value primarily through passing
25. C - Effective at getting ball to dangerous positions
26. B - xT differential correlates with match results
27. A - Team A is more efficient at converting xT to xG
28. B - xT helps identify threat generation on counter-attacks
29. A - Team can't play out from back but creates well from midfield
30. B - Adjustments may be needed for league differences

Scoring Guide