Watching a whole match from start to finish takes a long time and teaches less than you would expect. You notice your winners, wince at your errors and finish with a feeling instead of a plan. A short routine with a tally sheet gives you numbers you can compare with your next match.
You need a pen, the sheet below and a match filmed from behind the baseline, with both baselines and the alleys in the frame. If you have not filmed one yet, start with how to film your tennis.
The 20-minute match review routine
| Minutes | What to do | Why |
|---|---|---|
| 0 to 2 | Before you press play, write down three guesses, such as "my backhand let me down" or "I missed too many first serves". | Your memory of a match can be wrong. The gap between a guess and a count is often the best finding. |
| 2 to 3 | Choose your sample: about six games, three on your serve and three on your opponent's. If you remember where the set turned, use that stretch. | A sample keeps the job short and gives you a repeatable test. |
| 3 to 13 | Tally pass. Watch each point once, mark the sheet and skip the time between points. Do not stop to study technique yet. | Stopping breaks the count and turns 20 minutes into an hour. |
| 13 to 16 | Add up the counts and compare them with your guesses. Look for clusters: one side, one shot or one kind of miss. | A cluster is something you can practice. A single error is not. |
| 16 to 20 | Go back to your three most repeated errors. Pause at contact and look at your body, feet and racket. | Now you are studying technique with a reason. |
Ten minutes is usually enough for six games if you skip the gaps between points and the changeovers. If your sample runs longer, cut it to four games and keep the same number next time.
What to tally by hand: a sheet you can copy
Copy this table onto paper or into your notes. Keep the old sheets, because the last column is where the value is.
| What to count | How to count it | This match | Last match |
|---|---|---|---|
| First serves in | Mark each first serve in or out. Split deuce and ad sides if you can. | ___ of ___ | ___ of ___ |
| Double faults | Count each one. Add a dot if it came at 30-30 or later in a game. | ___ | ___ |
| Missed returns | Note whether it was a first or second serve. | ___ | ___ |
| Unforced errors, forehand | Use the test below. Write N for net and L for long or wide. | ___ | ___ |
| Unforced errors, backhand | Same marks. | ___ | ___ |
| Unforced errors, other | Volleys, overheads, slices and drop shots. | ___ | ___ |
| Net approaches | Each time you came forward on purpose. Write W for won and L for lost. | ___ | ___ |
| Short balls left alone | Balls you could have attacked, but you stayed back. | ___ | ___ |
People count unforced errors in different ways, so use a test you can repeat. Count an error as unforced if you had time to set your feet and the ball was no harder than a normal rally ball. If you are unsure, leave it out. The exact rule matters less than using the same rule every time.
Do not grade yourself against other people's numbers
SwingVision has published data from more than 250,000 sessions of its app. By SwingVision's count, first serves in stay between about 53% and 56% from NTRP 2.5 to 5.0, and unforced errors per match are about 8 at 2.5 and about 11 to 12 from 3.0 to 5.0. SwingVision's figures come from its own users, and it does not state how it assigned levels or classified errors, so your hand count will not match them. Still, from 3.0 up, those two counts barely separate the levels in SwingVision's data. Use your own last match as the benchmark.
What an app can do, and what stays on paper
Apps split this job in different ways, so check which part you still have to do yourself. Details of other apps are from their listings and sites as of November 2026.
| Job | By hand | TennisAI | Match-tracking apps |
|---|---|---|---|
| First serves in, errors, approaches | Yes, with the sheet above. | Not advertised on its listing. | SwingVision advertises automated stats. Baseline Vision's post-match summary covers points won, winners, aces and unforced errors. |
| The match score | Write it down. | Not advertised on its listing. | SwingVision and Baseline Vision advertise automated scoring. |
| Stroke technique | Yes, if you know what to look for. | Yes: scores for forehand, backhand and serve, plus feedback on swing path, timing, contact point and footwork. | SwingVision's site lists a technique review in its Pro plan. |
| What to do next | Your own judgment, or a coach. | AI Coach chat for drills and training questions. | SwingVision advertises coaching after each session. |
| Equipment | A phone on a mount. | A phone. No sensors or wearables. | SwingVision: an iPhone or iPad on a mount. Baseline Vision: its own net-post camera. |
If you want match numbers without a pen, SwingVision is the stronger tool, and our comparison with TennisAI sets out the differences. TennisAI's stroke feedback fits the second half of this routine, after your tally has told you which stroke to study.
How to turn three findings into next week's practice
Pick three findings from your sheet. As a simple rule, a finding must have a number and must have happened at least three times in your sample. Then use this table to go from the count to a target. The targets are suggestions, so adjust them to your level.
| If your sheet shows | Check on the video | Practice target |
|---|---|---|
| First serves mostly in the net | Toss position, and whether you reach up to the ball at contact. | Serve 30 first serves at about 80% effort. Count how many land in, then try to beat the count. |
| Double faults at 30-30 or later | Whether your arm slows down on the second serve. | Play a made-up score of 30-30 and hit 10 second serves to each side. Target: 7 of 10. The second serve guide has more. |
| Backhand errors mostly in the net | Late contact, or a swing that stays flat. | Cross-court backhands, aiming about 1 m (3 ft) above the net. Target: 15 in a row. |
| Forehand errors mostly long or wide | An open racket face, leaning back or swinging too hard. | Cross-court forehands at a controlled pace, landing past the service line. Target: 15 in a row. |
| Many missed returns | Whether you start your split-step around the moment the server hits the ball, as world-class returners do. | Return 20 serves from a partner with a short swing, aiming deep and down the middle. Target: 15 in play. |
| Few approaches, many short balls left alone | Where you stand when the short ball lands. | A partner feeds 20 short balls. Approach on each one, split-step and volley. Count the points you win. |
Write each finding as four lines, so you can act on it. The numbers in this example are made up to show the format.
One workable way to fit this in, as a suggestion only: give each finding about 20 minutes of one practice session, starting with the one that cost you the most points. After your next match, run the same routine with the same sheet. If the count improves, keep the target for another week or replace it. If it does not, go back to the video and check whether you picked the right cause.
Does reviewing match video actually help?
The evidence on video review in tennis is mixed. A 1989 randomized trial of 66 intermediate players found that video feedback was no better than traditional serve training. A 2024 study of 23 intermediate players found that those who chose when to get video feedback improved their tactics more than players on a fixed schedule. Neither study tested this routine. What video plainly gives you is a record of what happened, which you can compare with what you remember. It helps only if it changes what you practice.
Frequently asked questions
How do you analyze a tennis match?
Review a sample of about six games, tally first serves in, unforced errors by side and net approaches, then pause on the errors that repeat. Turn the top three findings into practice targets and repeat the same tally after your next match.
What stats should I track in a tennis match?
Start with first serves in, double faults, missed returns, unforced errors by side, and net approaches won and lost. They are simple to count by hand, and each one points to something you can practice.
What counts as an unforced error in tennis?
People count them in different ways. A practical test is that you had time to set your feet and the ball was no harder than a normal rally ball. Use the same test every time so that your counts can be compared.
Is there an app that tracks tennis match stats automatically?
Yes. As of November 2026, SwingVision (an iPhone or iPad on a mount) and Baseline Vision (its own net-post camera) both advertise automated scoring and match stats. TennisAI focuses on stroke feedback, and its listing does not advertise match stats.
Where should I put the camera to film a tennis match?
Behind the baseline and as high as you can, for example on a mount at the top of the fence. Both baselines and the alleys should stay in the frame.
See this in your own game
Upload a practice or match video to TennisAI and get feedback on your forehand, backhand and serve.
Sources
- SwingVision blog: The real data behind every NTRP level
- App Store: SwingVision: Tennis Pickleball
- SwingVision: official site (plans and equipment)
- Baseline Vision: official site
- Tennis.com: SwingVision delivers pro-level insights for recreational players (fence mount placement)
- PubMed: Mecheri et al. 2019, split-step timing and leg stiffness in world-class tennis players
- PubMed: Van Wieringen et al. 1989, video-feedback in learning the tennis service (intermediate players)
- PubMed: van der Meer et al. 2024, self-controlled video feedback and tactical skills in tennis
This article is general information about tennis training and does not replace a coach or medical advice. Stop and see a professional if something hurts. TennisAI is made by Pace, the publisher of this blog.
