Guide · Self-scouting with AI

Turn play-by-play into a scouting report with AI

How to use AI to turn your opponent play-by-play data into a scouting report to prepare your team for an upcoming game. Copy the prompt below and get your scouting report in minutes.

Updated 2026-09-15 · 9 min read

The prompt

Let's get straight into what you're here for: the prompt. Simply copy everything below and paste it into your AI tool. Include your play data at the bottom of the message or attach it as a CSV or Excel file. If you have a video analysis tool like Hudl, the easiest way is to export the play-by-play breakdown data from there.

You are a football scouting analyst preparing a self-scout / opponent report for a team's coaches and players. You will be given raw play-by-play data covering one or more games. Produce a tendency report based ONLY on that data.

STRICT RULES — follow all of them:
1. Use only the data provided. Never invent, assume, or estimate a play, player number, formation name, or outcome that isn't in the data.
2. Every statistic must show its sample size in parentheses, e.g. "Ran left on 21 of 26 first-down carries (81%)" — never a bare percentage.
3. Minimum sample size is 8 plays for any category. If a category has fewer than 8 plays, do not call it a "tendency" — list it under section 9 as a small-sample note instead.
4. Round percentages to the nearest whole number.
5. If a column needed for a section isn't present in the data (e.g. no formation column), say so briefly in that section and skip it — do not guess or fill in placeholder data.
6. If a tendency reverses or changes across games (e.g. run-heavy in games 1–3, pass-heavy in games 4–5), report that shift explicitly instead of averaging it away.
7. Write in short, stat-driven sentences. No filler adjectives, no hype, no generic commentary.
8. Bold (using **text**) the single most exploitable tendency in each section.

OUTPUT FORMAT — use exactly these 9 sections, in exactly this order, every time, with these exact headers:

1. Snapshot Summary
   5–8 bullet points: the tendencies a player should know cold before kickoff. Each bullet must include the stat and sample size.

2. Tendencies by Down & Distance
   A table: Down | Distance range | Run% | Pass% | Sample size | Most common call.
   Group distance into short (1–3), medium (4–6), long (7+).

3. Formation & Personnel Tendencies
   A table: Formation/Personnel | Run% | Pass% | Sample size | Notes.
   Only include formations/groupings with 8+ snaps.

4. Directional Tendencies
   Run direction (left/middle/right) and pass direction breakdown, each with sample size. Note any formation-specific directional bias (e.g. "Ace formation runs are 90% to the left, 18 of 20").

5. Key Players
   Top ball carriers and top pass targets by usage %, with attempt/target counts. Note any single player who accounts for an outsized share.

6. Situational Tendencies
   Cover whichever of these the data supports: red zone, backed-up (own territory), 3rd/4th-and-short, 3rd/4th-and-long, two-minute drill, and any clear game-state pattern (e.g. leading vs. trailing). Skip any you don't have data for.

7. Explosive Plays
   List the plays that gained the most yards (use a reasonable threshold like 15+ yards if you have gain data). Identify what formation/play type/direction produced them, if a pattern exists.

8. Tendency-Breakers
   Anything that contradicts the main patterns above — situations where they clearly deviated from their normal tendency. This section exists to stop the reader from over-trusting the averages.

9. Confidence & Data Notes
   State total games and total plays analyzed. Note anything missing from the data that limited the analysis, and flag any tendency built on a borderline sample size (8–12 plays) as lower-confidence.

Formatting: use markdown tables for any comparison with 3+ rows. Keep prose between tables to 1–3 sentences. Do not add a "recommended game plan" section unless I ask for one separately — this report is descriptive, not prescriptive.

Here is the play-by-play data:
[PASTE OR ATTACH YOUR DATA HERE]

How it works

1. Log the plays.For every game, record each play as one row: down, distance, formation, personnel, play type, direction, ball carrier or target, and result. A spreadsheet or CSV is perfect. But of course, you have to make due with what you got. Try to get as much data and details in there as possible.
2. Combine your games.Stack all games into one file. More data = more reliable tendencies. Ideally you have over 100 plays of data for a single team.
3. Paste the prompt above into your AI tool, then attach or paste your data.The prompt tells the AI exactly what to look for, how to prove every claim with numbers, and what order to present it in.

What your data needs

You don't need all of the data below, but again: the more data the better. The prompt is built to skip anything you don't have rather than guess. But all of the data below will enhance the report:

The report structure

The base report consists of the same 9 sections every time. This consistency is what makes the report easily skimmable. Of course, by all means adapt the prompt to add sections you want or remove sections you don't care about.

Some ideas on how you can improve the report based on your preference:

Example report

1 Snapshot Summary

• Run-first on 1st down: 73% run (52 of 71 snaps), almost always Inside Zone out of Ace.
• 2nd-and-long is the one clear passing down: 70% pass (23 of 33 snaps), usually a quick slant to #83.
• Ace formation = run, almost every time: 84% run (54 of 64 snaps), and when they do run from it, 89% goes left.
• #24 is the offense: 76% of all carries (118 of 154 runs).
• #83 is the passing-down target: 35% of all targets (41 of 117), concentrated on 2nd/3rd-and-long.
• 3rd-and-long breaks their identity: despite being run-heavy everywhere else, they pass 89% (8 of 9) on 3rd-and-7+.
• Two-minute drill is nearly automatic pass: 91% (10 of 11 snaps), out of Shotgun Spread.
• Explosive plays (15+ yds) mostly come off play-action from I-Formation, not from their base Ace run game.

2 Tendencies by Down & Distance

4th-and-short (2 plays) excluded — below the 8-play minimum. See Section 9.

3 Formation & Personnel Tendencies

174 of 182 snaps had a formation tag; 8 snaps (2 games) were untagged in the source data and are excluded from this section.

4 Directional Tendencies

5 Key Players

#24 is the offense's identity on the ground; stopping him without extra numbers in the box is unlikely. #83 is the clear pass-down valve — see Sections 2 and 4.

6 Situational Tendencies

7 Explosive Plays (15+ yards)

21 explosive plays across 5 games. Breakdown of how they happened:

• 62% (13 of 21) came off play-action out of I-Formation, targeting #88 or #11 deep — not their base run look.
• 29% (6 of 21) were broken-tackle runs by #24 out of Ace, left side.
• 9% (2 of 21) were QB scramble plays off a collapsed pocket.

8 Tendency-Breakers

• Game 4, 2nd half (trailing the whole game): abandoned the run almost entirely — 81% of second-half snaps were Shotgun Spread, well above their 5-game average. A team that falls behind against them should expect this shift, not the run-heavy identity from Sections 2–3.
• 3rd-and-long breaks the run-first identity: 89% pass (8 of 9) — the opposite of their overall run-leaning profile. Don't play run-first on 3rd-and-7+.
• Ace isn't always a run tell on 3rd down: of their 10 Ace pass plays all season, 4 came on 3rd down as play-action deep shots. Small sample (4 plays) — flagged as a possible trap look rather than a confirmed pattern.

9 Confidence & Data Notes

5 games, 182 offensive snaps analyzed. Overall confidence is high for Sections 1, 2, 4, and 5, where sample sizes are large and consistent across games.

Descriptive report only — reflects what Central Ridge has done in the 5 logged games, not a prediction or a game plan.

Back to the app