First serve percentage is easy to quote and easy to misread. A higher number can mean a better serve, or only a softer one. This guide shows what the published numbers say, why they do not give you a target on their own, and how to use your own video to decide whether your first serve should be harder or safer.
Is there an official good first serve percentage?
We found no benchmark from the USTA or the ITF (now World Tennis). The USTA's NTRP level descriptions talk about the serve in words. At 4.5, for example, a player "can hit the first serve with power and accuracy and can place the second serve." They give no percentage. The numbers people quote come from elsewhere, and each has limits.
| Source | Who it covers | First serves in |
|---|---|---|
| SwingVision (vendor data) | Its own app users, under NTRP labels 2.5 to 5.0 | About 53% to 56% at every label |
| Kovalchik and Reid, 2017 Australian Open | Professional men and women, match medians | 59.6% and 60.3% |
| ATP Tour analysis, 52 weeks to August 2024 | Top 10 men | 63.8% average, from 55.2% to 72.4% |
Read the SwingVision figures with care. They come from sessions logged in its own app, and it does not say how the NTRP labels were assigned. Even so, the pattern is interesting. Its first serve percentage barely moves across labels, while the first serve speed it reports rises from 51.0 mph at the 2.5 label to 74.8 mph at 5.0, on SwingVision's own average-flight measure and not radar. So in that data, higher labels do not land more first serves. They land about the same share of faster ones.
Use 50% to 65% as a quick check, and treat it as our suggestion. Below 50%, your second serve has to do a lot of work. Above 65% is fine if you still win those points, and a warning if you only hit the ball softly. The rest of this guide shows how to tell which one you are.
Why points won behind your first serve matter more
Look at one peer-reviewed dataset: match medians from the 2017 Australian Open.
| Group | First serves in | Points won when the first serve is in | Second serve points won |
|---|---|---|---|
| Professional men | 59.6% | 72.1% | 50.4% |
| Junior boys | 61.6% | 66.7% | 47.3% |
| Professional women | 60.3% | 64.2% | 45.6% |
| Junior girls | 58.0% | 64.1% | 44.3% |
Junior boys landed more first serves than professional men, 61.6% against 59.6%, and won fewer points behind them, 66.7% against 72.1%. The paper's table calls the first column "First Service Points", and we read that as first serves in. These are tournament players, so use the pattern and not the numbers.
An ATP Tour analysis makes the same point at the professional level. It multiplies first serves made by points won behind them into a "First-Serve Rating". In its Top 10 list, Alexander Zverev made 72.4% and won 75.2% of those points. Alex de Minaur, in the same Top 10, made 55.2% and won 72.7%.
How a lower first serve percentage can win more points
Your service points won come from two parts: the points after a first serve lands, and the points after it misses.
Here is one invented server who wins 40% of second serve points. These numbers show the arithmetic and are not a benchmark.
| Version of the first serve | First serves in | Won when in | Service points won |
|---|---|---|---|
| Today | 50% | 70% | 55.0% |
| Harder, lands less often | 45% | 75% | 55.8% |
| A little slower, still dangerous | 60% | 67% | 56.2% |
| Much slower and soft | 70% | 52% | 48.4% |
The version with the highest percentage has the worst result. The table also gives you a break-even: to go from 50% to 60% in, this server can afford to fall from 70% to 65% won when in. Any lower and the extra first serves cost points.
The gap between your two win rates decides how much an extra first serve is worth. Here it is 30 percentage points (70% against 40%), so 10 more first serves per 100 are worth about 3 more points won per 100, if the win rates hold. With a small gap, a first serve miss costs you very little.
How to count your serve stats from video
Any match video works. If you would rather not count by hand, SwingVision's App Store listing advertises automated stats as of November 2026, and that is a real advantage. Our TennisAI and SwingVision comparison covers the difference. To do it yourself:
- Film a match or some practice sets. Our guide to reviewing match video covers the setup and how to get through the footage.
- Go through your service games only. Give each service point one code from the table below. A let is replayed, so it gets no code.
- Add up the codes and work out the four numbers.
- Repeat over several matches until you have at least 100 first serves. That number is our suggestion.
| Code | What happened | Example count |
|---|---|---|
| 1W | First serve in, you won the point | 18 |
| 1L | First serve in, you lost the point | 9 |
| 2W | Second serve in, you won the point | 9 |
| 2L | Second serve in, you lost the point | 10 |
| DF | Double fault | 4 |
| Number | How to work it out | From the example |
|---|---|---|
| First serves in | (1W + 1L) divided by all points | 27 of 50 = 54% |
| Won when first serve is in | 1W divided by (1W + 1L) | 18 of 27 = 67% |
| Second serve points won | 2W divided by (2W + 2L + DF) | 9 of 23 = 39% |
| Service points won | (1W + 2W) divided by all points | 27 of 50 = 54% |
Small counts move a lot by chance. If your true rate is 60%, ordinary sampling arithmetic gives a 95% range of about 14 percentage points either way over 50 first serves, 10 over 100 and 7 over 200. So a move from 52% to 60% between two matches may mean nothing at all.
What to change: a decision table
Compare "won when first serve is in" with "second serve points won". The cut-offs are our suggestions.
| What your tally shows | What it suggests | What to try |
|---|---|---|
| Gap under 10 percentage points | Your first serve gives you little advantage | Add pace or better placement, because a miss costs little |
| Gap of 20 percentage points or more, under 50% in | A strong first serve that lands too rarely | Take a little pace off or add spin, and check the break-even |
| Gap of 20 percentage points or more, over 60% in | A good balance | Leave the first serve alone and work on the second |
| Second serve points won under 35%, or many double faults | Your second serve is the weak part | Our second serve guide, then a kick serve if you want a higher bounce |
| Anything in between | The tally cannot decide | The on-court test below |
More pace is not the only way to win more points behind the serve. In a study of 25,680 first serves by 151 professional men at the Australian Open, hitting aces depended more on accuracy than on speed. That study is about aces by professionals, but it is a good reason to try better placement before you swing harder.
An on-court test: harder or safer first serve?
You need a partner who will play the points out, and the same tally.
- Play 20 points using your full-pace first serve, half from each side. Use your normal second serve after a miss.
- Play 20 points using a first serve at about three-quarter pace.
- Repeat over three sessions, so each version has 60 points.
- Compare service points won. Use the version that wins more. Treat a gap of fewer than 6 of 60 points as a tie, because a gap that small can easily be chance, and keep the serve you trust under pressure.
When the numbers point at one serve, film it. TennisAI scores your serve from an uploaded clip and comments on timing, contact point and balance, which can help you find a cause. The score is an automated estimate that can be wrong, and it varies with camera angle and lighting. It rates technique, so it does not replace your count of points won. Keep both records.
Frequently asked questions
What is a good first serve percentage in tennis?
There is no official figure. We suggest 50% to 65% as a working range, based on vendor data from SwingVision and on professional match data. Points won behind the serve tell you more than the percentage.
What is a good first serve percentage for a recreational player?
No governing body publishes one. SwingVision's data for its own app users shows about 53% to 56% across NTRP labels 2.5 to 5.0, though it does not say how those labels were assigned, so treat it as a rough guide.
What is the first serve percentage on the ATP Tour?
An ATP Tour analysis put the Top 10 average at 63.8% over the 52 weeks to August 2024. At the 2017 Australian Open, the median for professional men was 59.6%.
Should I hit my first serve harder or get more in?
It depends on how many points you win when it lands. Compare service points won for a full-pace and a three-quarter-pace first serve over at least 60 points each, and use the one that wins more.
How do you calculate first serve percentage?
Divide first serves in by all service points. A double fault counts as a first serve missed. A let is replayed, so it does not count as a serve.
See this in your own game
Upload a practice or match video to TennisAI and get feedback on your forehand, backhand and serve.
Sources
- USTA League support (Active Network): League NTRP rating information
- SwingVision blog: The real data behind every NTRP level (vendor data)
- Kovalchik and Reid, Journal of Sports Science and Medicine (2017): Match statistics of junior and professional players at the 2017 Australian Open
- ATP Tour, Infosys ATP Beyond The Numbers (August 2024): Does this blended stat make Zverev the game's most effective server?
- Whiteside and Reid, Journal of Sports Sciences (2017): Spatial characteristics of professional tennis serves
- App Store: SwingVision listing
- ITF (now World Tennis): Rules of Tennis 2026
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.
