anti-slop-analytics
Review sports figures, tables, notebooks, and reports for chartjunk, fake certainty, cropped axes, baseline erasure, metric laundering, and weak reproducibility. Use when asked to clean up or audit analytical presentation.
WalrusQuant · 已有 AI 审核记录
安装与使用
运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。
npx skills add WalrusQuant/sports-analytic-skills --skill anti-slop-analytics技能说明
name: anti-slop-analytics description: > Review sports figures, tables, notebooks, and reports for chartjunk, fake certainty, cropped axes, baseline erasure, metric laundering, and weak reproducibility. Use when asked to clean up or audit analytical presentation. license: MIT metadata: version: "0.12.0"
Anti-Slop Analytics
Outcome
Produce an item-by-item keep, fix, or kill review. Every retained figure or
table must make a defensible claim and disclose its population, period,
denominator, comparison baseline, uncertainty, and reproduction pointer.
Pretty is optional. Honest is required. Presentation cannot rescue a weak design, a losing model, or an unverifiable calculation.
When to use this skill
Use it when reviewing:
- exploratory or diagnostic plots;
- model-comparison and calibration figures;
- dashboards, notebooks, reports, papers, or public posts;
- requests to make results look cleaner, stronger, or more impressive;
- any visual whose design may conceal a weak baseline or unstable result.
If no visual or table exists, create the analytical artifact first. Use
sports-visualization for figure construction, validation-design for the
evaluation itself, and results-reporting for the full narrative. This skill
audits presentation; it does not repair the underlying study by relabeling it.
Inputs and evidence
Request or inspect:
- the figure, table, notebook, or report;
- the one claim each item is intended to support;
- the underlying user-owned data or metrics artifact when available;
- the decision context and audience;
- the plotting command, script, or notebook cell;
- relevant filters, exclusions, and fold definitions.
If evidence is missing, audit what is visible and mark unverifiable properties
as unknown. Never infer provenance, sample size, uncertainty, or held-out
status from visual polish.
Audit workflow
- Inventory every figure and table with a filename or stable identifier.
- Write the single decision-relevant claim each item is meant to support.
- Verify sport, competition, grain, period, population, sample size, exclusions, and denominator.
- Confirm that the plotted quantity matches the title, caption, and claim.
- Check that a relevant null, naive model, or incumbent baseline is visible.
- Inspect axes, transformations, scales, bins, smoothing, color, ordering, facets, and annotations.
- Separate training, in-sample, held-out, and forward-looking evidence.
- Inspect statistical honesty: uncertainty, multiplicity, selection, missing segments, and sensitivity to the chosen time window.
- Trace the item to a user-owned data file, metrics file, notebook, or command.
- Assign
keep,fix, orkill; prescribe the smallest honest replacement.
Failure catalog
| Failure | Why it fails | Preferred replacement |
|---|---|---|
Certainty without n | hides evidence volume | report denominator and interval or fold spread |
| Cropped quantitative axis | exaggerates small changes | zero or a justified domain range with disclosure |
| Cherry-picked time window | turns a hot streak into a general claim | full relevant period plus sensitivity window |
| Baseline erasure | prevents judging incremental value | plot candidate beside locked baseline |
| Metric laundering | promotes a flattering secondary metric | lead with the predeclared primary metric |
| Training as validation | overstates generalization | clearly labeled held-out or walk-forward results |
| Ranking without uncertainty | treats noisy order as stable | intervals, rank distributions, or stability summaries |
| Accuracy without base rate | hides class imbalance and probability quality | baseline plus log-loss or Brier score |
| Mean without distribution | hides tails, skew, and fold losses | quantiles, interval, distribution, or fold table |
| Unlabeled smoothing | conceals a transformation | state method, window, and parameters |
| Dual axes | manufactures visual correlation through scaling | aligned panels or indexed series |
| Giant correlation heatmap | substitutes volume for a hypothesis | selected relationships with purpose and uncertainty |
| Decorative 3-D, pie, or gradients | adds non-data ink and distorts comparison | flat marks, direct labels, or a small table |
| Screenshot-only evidence | cannot be reproduced or checked | exported artifact plus source and command |
| Probability framed as destiny | converts uncertainty into a lock | probability plus calibration and decision context |
| Sport context removed | makes grain and population ambiguous | sport, competition, grain, and period in label/caption |
For additional examples, read
references/slop_catalog.md when classifying a
suspect item and references/replacements.md
when choosing a redraw.
Verdict rubric
| Verdict | Use when | Required action |
|---|---|---|
keep | claim, scope, baseline, uncertainty, and provenance are clear | retain; tighten caption only if needed |
fix | calculation is usable but the presentation can mislead | specify exact axis, label, baseline, or encoding repair |
kill | wrong grain, wrong evidence, unverifiable, redundant, or decision-irrelevant | remove or replace with a new artifact |
If a decorative chart can be fixed only by changing its core claim, kill it and redraw from the underlying evidence.
Sports-specific checks
- A doubled team-game panel has two rows per contest. Do not report its overall win rate as home advantage; filter home rows or return to game grain.
- Team and player rankings often have unequal schedules and sample sizes. Show opportunity, exposure, or shrinkage rather than raw order alone.
- Season averages can hide rule, schedule, roster, or measurement changes. Label eras and test whether the conclusion survives them.
- Probabilities require calibration evidence. Discrimination alone does not show that a displayed 70% forecast