The Second Serve — Where Big Hard-Court Matches Are Actually Decided

THE SECOND SERVE — WHERE BIG HARD-COURT MATCHES ARE ACTUALLY DECIDED

A practice session and twelve balls

August 18, 2026, Court 4 at the Billie Jean King National Tennis Center in Flushing Meadows, New York. Four in the afternoon. Thirty-one degrees Celsius, 58 percent humidity, wind from the east cutting across the court at around 14 km/h. Alex de Minaur stands behind the baseline and hits twelve second serves in a row. I start my stopwatch and write each one into a small notebook.

Seven went in. Five missed. Of the seven that landed, four were returned directly by his practice partner with a cross-court forehand, and the remaining three forced de Minaur to run to his left immediately after contact. It was only a practice session. No crowd, no umpire, no scoreboard. But there was one thing I managed to count, and it matched what I had been logging for twenty months.

Australia's number one does not lose big hard-court matches on his first serve. He loses on his second serve.

I wrote that line down, underlined it, and then sat still for another ten minutes watching him continue. Some things only appear when you are willing to sit quietly for longer than one set.

Context: a decade of hard court compressed

To understand why the second serve has become the focal point, it needs to be placed inside the current context of men's tennis.

From 2026 to now, the average speed of hard courts at major events has shifted in two opposite directions. North American events — especially Cincinnati and the US Open — were slowed considerably compared with the 2010s, while several European and Middle Eastern events were sped up. The result is that the gap between surfaces has narrowed but has not disappeared. Players still win in different ways in Melbourne, Indian Wells, Miami, Cincinnati and New York.

More important: the average height of the top ten has increased. Many players in that group stand above 1.88 metres and can serve at 210 km/h or more with consistency. That has made the first serve an almost default weapon — everyone has one. When everyone has the same weapon, that weapon gradually loses its power to separate.

Over the same period, the number of events on the calendar has not decreased. If anything, it has risen. The hard-court season before the US Open runs from late July to late August with three Masters 1000 events and several ATP 500 and ATP 250 tournaments. Add qualifying, training blocks, travel across time zones and media obligations, and a player inside the top group can play twelve to fifteen competitive matches in five weeks. This is a point I will return to at the end.

I began systematic note-taking in January 2026, after a long stretch of writing purely by feel. My method is simple: for every hard-court match by a player on my watch list, I record four sets of numbers. First, first-serve percentage. Second, points won on first serve. Third, points won on second serve. Fourth, points won in games where that player is a break down, or level at 4-4 or 5-5, in the third and fourth sets.

The first three are publicly available. The fourth is not, because it requires rewatching footage and counting by hand. That is why it takes most of my time.

The first serve: the number that gets talked about too much

In the press room, the first question after a defeat is usually about the first serve. What was the percentage, what was the average speed, was the court too slow. Those questions are easy to answer, and because they are easy to answer, they get asked most.

But my data shows that first-serve percentage, across a large sample, correlates weakly with the final result in hard-court matches between two players in the top twenty. In other words, a player landing 62 percent of first serves instead of 68 percent does not automatically lose. The difference lies in what he does with the remaining 32 percent.

Here is a specific example from my notes. In a 2026 hard-court quarterfinal I watched live, the winner landed only 58 percent of first serves, seven percentage points below his opponent. But he won 61 percent of second-serve points, while his opponent won only 44 percent. The gap on second serve was seventeen percentage points — more than double the gap on first serve. The final scoreline shows none of this.

This is the crux: traditional statistics place the first serve at the centre, while most elite hard-court matches are decided by the second serve. Numbers do not lie. It is just that we have to ask the right question.

There is a technical reason for this. When a player has a consistently fast first serve, the opponent is forced to stand several centimetres further back to buy reaction time. That position directly affects the second-serve return. If the second serve is good enough, the server can exploit the space immediately in front of him. If not, he faces a return at chest height, and from there everything becomes difficult.

In other words, the first serve and the second serve are not two separate things. They are one system. And that system, for most players in the top ten, is losing balance in the same direction.

The second serve: the blind spot of every scoreboard

Second-serve points won is one of the least discussed metrics in broadcast analysis, and also one of the most discriminating.

In my notes covering the 2026 hard-court season and the first half of 2026, players inside the top eight seeds at Masters 1000 events recorded second-serve points won between roughly 55 and 62 percent. Players ranked 30 to 60 recorded between 48 and 53 percent. The gap between the two groups is only about seven to nine percentage points — which sounds small.

But multiply it. A player serves roughly 60 to 70 times in a three-set match, and 90 to 110 times in a five-set match at a major. About a third of those are second serves. That means a five-set match contains roughly 30 to 37 second-serve points. An eight-point gap across 35 points equals nearly three points per match. Three points in a match where the total often lands between 180 and 220 is not a small number. It usually equals one break.

And in modern men's tennis, one break in a deciding set is usually the whole match.

I want to stress one thing about reading this metric. A high second-serve points-won rate does not mean the player has a better second serve. It could mean he chooses better locations, or that his second serve is designed to create a favourable next shot, or that he accepts more risk on the first serve to reduce how often he has to hit a second. There are at least four different ways to reach the same number, and they are not tactically equivalent.

That is why I always add a secondary column: after the second serve, did the player win the point with the third shot, or with an opponent error. If most of the winning points come from opponent errors, the metric is far more fragile than it looks. If most come from an aggressive third shot, it is a more durable signal.

Deciding games and the break points nobody remembers

The most interesting section of my notebook is a column I named "pressure". It is a player's points-won rate in games where he is a break down, or level at 4-4 or 5-5, or serving to hold the set in the third and fourth sets.

I chose these situations because they share one quality: there is no time to think. The ball arrives and the player must decide within seven to ten seconds. This is where technique trained over thousands of hours is either good enough or not.

In my sample, the pressure gap between the top group and the middle group is considerably larger than the gap in any other metric — roughly ten to twelve percentage points, compared with seven to nine on second serve and four to five on first serve.

I do not remember what I wrote. I remember what I counted. And what I counted is this: in the third set of hard-court matches between top-twenty players, the second serve carries a markedly higher share than in the first two sets.

There is a reasonable explanation. In the first two sets, both players are relatively fresh. They can run more to compensate for a weak second serve. By the third set, that ability to run and cover declines. The true quality of the second serve is exposed, and it is exposed exactly where it cannot be hidden.

This sounds obvious, but it has a consequence that draws little attention. If most big matches are decided in the third and fourth sets, and if most points in the third and fourth sets are decided by the second serve, then judging a player by his first serve is a systematic error.

I tested this with my own data across several events. In most cases, the player with the better second-serve points-won rate in the first two sets was also the winner, regardless of first-serve numbers.

Surfaces do not speak the same language

One caution: hard court is not a single surface. This is where much broadcast analysis makes a mistake.

Melbourne Park and Flushing Meadows are both hard courts. But the ball bounces differently, the bounce height differs, and the way the ball reacts to spin differs. In Melbourne, the surface tends to bounce lower and the ball travels slightly faster in daytime conditions. In New York, the bounce is higher, and evening conditions can completely change the perception of speed.

For the second serve, this difference matters. The second serve usually relies on spin and height to create space. On a low-bouncing surface, the same second serve may arrive at the opponent's hip rather than shoulder, turning a neutral shot into an attacking one.

So when I compare second-serve metrics across events, I always split by event rather than merging them into "hard court". People often cite a single number for the whole hard-court season, and that number hides the biggest difference.

One example from my notes: the same player, in the same season, showed a second-serve points-won rate differing by roughly six to eight percentage points between a European event and a North American one. He did not change his technique. The surface changed.

Alex de Minaur: a twenty-month file

Back to the specific case.

De Minaur was born on February 17, 2026, in Sydney. He turned professional in 2026, won his first ATP title in Sydney in 2026, and first entered the world's top ten in 2026. He reached a career-high ranking of world number six in July 2026, and from 2026 into 2026 he spent most of that period inside the top ten.

His strengths are widely known: speed, defensive coverage, the ability to turn defence into attack, and physical consistency. His most discussed weakness: modest height relative to the top group, and therefore serve speed.

From January 2026 to August 2026, I tracked and logged his hard-court matches in Melbourne, Indian Wells, Miami, Cincinnati and New York, plus several ATP 500 events. My sample covers roughly thirty matches — large enough to talk about a trend, not large enough for a firm conclusion.

In that sample, his first-serve percentage hovered around 62 to 66 percent, broadly in line with the top-twenty average. His points won on first serve sat at 72 to 76 percent, about two to four percentage points below the top eight seeds. This is the number the media usually uses to argue he lacks a big enough serve.

But the third metric is the telling one. His second-serve points-won rate in this sample hovered around 51 to 55 percent. The top eight seeds average 55 to 62 percent. That gap — roughly four to seven percentage points — is larger than the gap on first serve.

And the fourth metric, the pressure metric, is where the gap is widest. In games at 4-4 or 5-5 in the third and fourth sets, his points-won rate in my sample drops noticeably below his own average in the first two sets. That drop is larger than the average drop for the top eight seeds.

I need to be clear: this is a small sample, collected by hand, and it may contain errors. Nothing can be firmly concluded from thirty matches. But the direction of the data is fairly clear, and it differs from the story the media tells.

The media story is: de Minaur needs a bigger serve. The data story is: de Minaur needs a better second serve, and better at exactly the moments when he cannot run to cover.

These are two different stories, and they lead to two different training programmes.

What the top five say

To test the direction of this analysis, I compared with the leading players over the same period.

The common thread among them is not the biggest first serve. Some in that group genuinely have huge first serves; others do not. The real common thread is a high and stable second-serve points-won rate — and more importantly, a rate that does not fall much when the match reaches the third and fourth sets.

In other words, what separates the top group from the rest is not the peak of the first serve. It is the stability of the second serve under pressure.

I noticed this when I placed two columns side by side in the spreadsheet. The first column was second-serve points won in the first two sets. The second was the same rate in the third and fourth sets. For the top eight seeds, the difference between the two columns is usually only two to four percentage points. For players ranked 20 to 50, the difference is usually six to ten percentage points.

That decline, not serve speed, is the clearest sign of a player not yet ready for the biggest matches.

The beat keeper does not compose the music, but without him everything falls out of rhythm. The second serve is that beat keeper.

Schedule density: the forgotten culprit

So far I have discussed technique and data. But there is another variable that needs to be on the table, and it is often ignored because it appears in no scoreboard.

That variable is schedule density.

Between July and September, a player inside the world's top thirty can play three Masters 1000 events, one or two ATP 500s, and one Grand Slam across roughly eight weeks. Add qualifying at some events, mandatory practice blocks, media and sponsor commitments, and travel across three different time zones.

I do not have medical data, and I will not speculate about specific cases. But I have my observations, and they are fairly consistent: in a player's third or fourth match within twenty days, his second-serve points-won rate usually drops. Not his first-serve percentage — that often holds. His second-serve points won.

There is a simple physical explanation. The second serve demands more than the arm. It demands the legs, the hips, and the ability to rotate stably. When the body tires, the second serve is the first shot to lose quality. Nobody notices, because it does not break down entirely — it simply loses five to eight percent of effectiveness. But five to eight percent across 35 points is enough to lose a set.

Schedule density is the single biggest cause of injuries, and no medical team can save a player from two matches a week for eight straight weeks. I have observed this across many seasons and I will keep returning to it.

The counter-intuitive angle: the trap of the big serve

Here I want to say what most broadcast analysis does not.

When a player loses a big match, the default reaction is: he needs to serve bigger. This is an attractive conclusion because it is simple, easy to picture, and seems logical. If you serve harder, you win more points. Anyone can follow that.

But the data does not support that direct relationship.

I looked at players inside the top thirty who significantly increased their first-serve speed over roughly two seasons. For most of them, the speed increase did not produce a corresponding improvement in big hard-court results. In some cases, it came with a decline in first-serve percentage, and therefore more second serves.

This is the trap. A faster first serve produces more direct winners. But it also produces more second-serve situations, because the percentage drops. And if the second serve has not improved, the player has traded one winner for two points of risk.

The match I watched most closely this season illustrates it. The losing player's average first-serve speed was about 8 km/h higher than his opponent's. But he hit 34 second serves in the match, against 21 for his opponent. Of those 34, he won only 15 points. His opponent, with 21, won 13. Looking at speed, anyone would assume the loser had the better weapon. Looking at the score, the opposite is true.

There is another interpretation I find more plausible: the problem is not speed, it is structure. A good second serve is a second serve with a plan — knowing where the third shot goes, knowing which direction the return will come back if the opponent gets to it. That is a question of choice and discipline, not power.

And discipline, unfortunately, is far harder to coach than power.

I also want to be clear that this does not mean a big serve is useless. At the elite level, it remains an advantage. But it is a conditional advantage, and the condition is a second serve good enough not to give back what the first serve earned.

I keep a clear distinction between "not enough data to conclude" and "the data shows the opposite". Here I am in the second state, with moderate confidence. I am not saying the first serve does not matter. I am saying the order of priorities is wrong.

What would change my mind

A useful analysis must state what would prove it wrong.

If, across the next twenty hard-court matches I track, the player with the highest first-serve points-won rate in the top thirty is also the one winning most matches between top-twenty players, regardless of second-serve numbers, I will have to revisit this whole direction. If the pressure metric gap disappears as the sample grows, I will revisit it too. And if players who raised their first-serve speed last season improve their results accordingly, that will be a clear counter-signal.

Until then, I keep counting.

Internal signals to watch

As the hard-court season moves into its final stretch in New York, there are three things I will track and write down.

First, the second-serve points-won rate of top-ten players, split between the first two sets and the third and fourth. If a player has a high rate in the first two sets but a sharp drop in the third, that is a signal worth noting, whatever the result of that match.

Second, the number of second serves in games at 4-4 and 5-5. If a player hits notably more second serves in those games than his own average, it suggests he is changing his approach under pressure — and usually not for the better.

Third, the number of rest days between matches. This is the least discussed variable and possibly the most important. If a player plays his fourth match in eight days and his second-serve points-won rate drops, I will not call it a technical problem. I will call it a scheduling problem.

Fans have the right to live in emotion; I have the duty to live in data. But data is only useful if you know what you are looking for. For twenty months, I looked in the wrong place. Now I have corrected the question, and the answer — still incomplete — is beginning to surface exactly where it always was: the second serve.