POSITIONING DECISIONS USING WEARABLES: SOME INSIGHTS FROM THE INVASION-TERITORIAL TEAM SPORTS Dr Bill Gerrard Professor of Business and Sports Analytics, University of Leeds N8 PRP Training Session Leeds University Business School, 16 th May 2018
POSITIONING DECISIONS USING WEARABLES:SOME INSIGHTS FROM THE
INVASION-TERITORIAL TEAM SPORTS
Dr Bill Gerrard
Professor of Business and Sports Analytics, University of Leeds
N8 PRP Training SessionLeeds University Business School, 16th May 2018
THE ICE-CREAM SELLER PROBLEM
• Context: sunbathers equally spaced along beach; same brand of ice cream sold at same price by both ice-cream sellers
• Positioning problem: where will the two ice-cream sellers position themselves?
THE ICE-CREAM SELLER PROBLEM: PREDICTED OUTCOME
• Predicted outcome: both ice-cream sellers will position themselves in the centre of the beach
• Rationale: only differentiator is distance; incentive for each seller to set up next to each other to grab all the market on one side and share the market on the other side
THE ICE-CREAM SELLER PROBLEM: BEHAVIOUR DYNAMICS
25%25%25% 25%
25%75%
50% 50%
POLITICAL ELECTIONS OFTEN INVOLVE THE SAME POSITIONING LOGIC
THE POLITICAL LANDSCAPE
LEFT CENTRE RIGHT
SqueezedOut
THE POSITIONING DECISION
• Positioning decision: choosing how to optimally
allocate people/things/activities within a specified
space
• Three components in all positioning positions:
- Objective: what are you trying to achieve?
- Context: the relationship between positioning and
outcomes
- Constraints: space; resources
POSITIONING: DIFFERENT CONTEXTS BUT OFTEN A COMMON SOLUTION
• Positioning decisions occur in very different contexts:
- Economics – product positioning
- Politics – electoral positioning
- Geography – locations of cities, towns, villages
- Military conflicts – deployment of military forces
- Invasion-territorial team sports – game tactics
- Policing – what are your positioning decisions?
• But often share a common solution
Optimal Positioning = Median Location
CENTRAL LOCATION THEORY
• Central location theory = spatial
theory in urban geography
• Attempts to explain the reasons
behind the patterns of
distribution, size and numbers of
cities, towns and villages
• Key idea: think of space as
consisting of a series of hexagons
at different levels
ANALYSING SPACE - THINK HEXAGONS NOT CIRCLES OR SQUARES!!!
Why are hexagons the best unit for spatial analysis?
Circles: equi-distant from centre to edge but overlap/leave gaps
Squares: nest together without gaps/overlaps but edges not equi-
distant from centre
Hexagons: nesting property and better approximation to equi-
distance property of circles
THE IMPORTANCE OF SPACE IN TEAM SPORTS
• Invasion-territorial team sports: various codes of
football, rugby, hockey and basketball
• Essence of invasion-territorial sports:
THE CONTROL OF SPACE THROUGH THE
TACTICAL COORDINATION OF PLAYERS
• Spatial dimension is crucial to producing
performance analytics that supports coaching
decisions effectively
THE DIFFERENT LEVELS OF SPATIAL ANALYTICS
Level 3: Trajectory Data
Level 2: X,Y Event Data
Level 1: Zoned Frequency Data
GPSVideo Tracking
Video Analysis(Pitch Grid)
Video Analysis(Pitch Map)
LEVEL 1: EXIT PLAY IN RUGBY UNION
Aviva Premiership Final, May 2015
Bath 16 – 28 Saracens (HT 3 – 25)
• Frequency data indicates that a key factor in Bath’s defeat was their high number of turnovers conceded
• But it is only zoned data that provides a real tactical understanding
• Lesson: too many ball-in-hand plays in own half can be dangerous
Source: Opta data
LEVEL 2: AVERAGE PLAYER POSITION IN FOOTBALL
• Average position calculated based only on player actions
• Example: Ajax , vs AZ, Eredivisie R25, 2015/16, 0 – 60 mins (before substitutions)
LEVEL 3: WEARABLES IN ELITE TEAM SPORTS
LEVEL 3: WEARABLE OUTPUTS –LOCATION; DISTANCE; SPEED
LEVEL 3: MEASURING DEFENSIVE STRETCH IN BASKETBALL
• Basketball analytics is most advanced in analysing space
• CHAD (convex hull area of defence) provides a measure of
defensive stretch
• Allows quantification of impact of “floor-spacers” on attacking
team who stretch opposition defences (i.e. CHAD)
CHAD
Space to be attacked Space to be
created
LEVEL 3: POSITIONING ALGORITHM FOR DEFENDING IN FOOTBALL
• Blue Squares = team in
possession
• Blue Circle = player with
ball
• Black Diamonds =
defending team
• GKs excluded
• Positioning decision involves weighting demands to cover space, opposing players and the ball
• Algorithm currently can give too much weight to covering space in some situations
• Arrows indicate coaching interventions to adjust positioning decision
POLICING AND POSITIONING
•What are your positioning decisions?