Using AI and Statistical Techniques to Correct Play-by-play Substitution Errors
Bibliographic record
Abstract
Play-by-play is an important data source for basketball analysis, particularly for leagues that cannot afford the infrastructure for collecting video tracking data; it enables advanced metrics like adjusted plus-minus and lineup analysis like With Or Without You (WOWY).However, this analysis is not possible unless all substitutions are recorded and are correct.In this paper we use six seasons of play-by-play from the Canadian university league to derive a framework for automated cleaning of play-by-play that is littered with substitution logging errors.These errors include missing substitutions, unequal number of players subbing in and out, substitution patterns of a player not alternating between in/out, and more.We define features to build a prediction model for identifying correct/incorrect recorded substitutions and outline a simple heuristic for player activity to use for inferring the players who were not accounted for in the substitutions.We define two performance measures for objectively quantifying the effectiveness of this framework.The play-by-play which results from the algorithm opens up a set of statistics that were not obtainable for the Canadian university league which improves their analytics capabilities; coaches can improve strategy leading to a more competitive product, and media can introduce modern statistics in their coverage to increase engagement from fans.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.008 | 0.050 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.005 | 0.004 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".