Rolling the Dice: The Case for Two First Serves

What’s life without a little risk?

It’s 11:00 pm on a Tuesday night in Paris. Out on Court Suzanne Lenglen, 17th seed Alexander Bublik finds himself a set up on local qualifier Gregoire Barrere in the first round of Roland Garros. The boisterous crowd are going nowhere. Bublik is serving at 4-4 30-30 and everyone is willing this match to go long into the early hours. His 1st serve finds the tape. Suddenly the thought of Barrere levelling the match is a real possibility. Bublik laces his 2nd serve down the T. Ace. 10 minutes later he’s up two sets to love.

. . .

It’s hard to overstate the importance of the 2nd serve. More than a third of all points on tour begin with one (generally on both ATP and WTA tours the average 1st serve % is in the low 60s). For a sport where tactics vary so much, pretty much every player has the same philosophy for the 2nd serve: don’t miss. While this might be obvious - a missed 2nd serve is of course a point lost - a poor 2nd serve isn’t much better. Common logic would say keep your 2nd serve away from the opponent’s strength and maybe throw a few in to that strength to keep them guessing. However, if a player is just rolling the 2nd serve in they are going to be starting most points on the back foot. Bublik is well aware of this. As a result, he’ll often prioritise pace over precision on the 2nd delivery and effectively hit another 1st serve. Just how wise is this?

The Benefits

• As mentioned, you’re not going to be starting any points on the back foot if your 2nd serve is a 1st serve. The opponent is going to have to change their return tactic and won’t be able get on the attack early in rallies.

• Unpredictability. With the opponent anticipating a 1st serve, this is the perfect opportunity to catch them off guard and throw in an underarm. Bublik is famous for this. In the lead up to Roland Garros this year he hit 6 underarm serves in the one game. There is a lot to be said for keeping an opponent on their toes. When they know there’s a chance of an underarm serve, they are never going to be fully comfortable in their return position.

The Drawback

You are going to double fault more - there are no two ways about it. Bublik, who doesn’t fully commit to two 1st serves, double faults 1 out of every 10 points. This is by far the most of the ATP top 50. If we assume every player in the top 50 adopted the two 1st serves approach (with their 2nd serve % the same as their 1st), players would on average double fault 13.8% of the time, or 1 out of every 7 points. A player like Zverev who has made 72.2% of his 1st serves over the last year would be expected to double fault just under 8% of the time, or 1 in 13 points. At the other end of the spectrum you have De Minaur who only makes 56.3% of his 1st serves. He would be expected to double fault 19.1% of the time, or 1 in every 5 points. 

The Numbers

Let’s imagine a player adopts the two first serves tactic. A simple way to estimate their new 2nd serve points won would be to multiply their 1st serve % by their 1st serve points won. This might sound familiar as it is how I calculated 1st serve effectiveness (1stEff) in this article on Hurkacz. Obviously this isn’t a perfect metric, but it should stand as a good enough indicator. In theory, the new 2nd serve % could be higher than expected as after missing a serve a player would know what slight adjustment to make for the 2nd serve. But then if they often attempt the same serve again their opponent will have a better idea of where the serve is going to go which would theoretically reduce their 2nd serve points won. Also, exclusively attempting 1st serves would in turn give the opponent more opportunity to get used to the 1st serve. Basically, there are a lot of factors in play here.

To see whether a player should consider the two 1st serves approach, we can compare their 1st serve effectiveness to their 2nd serve points won (2nd%). If the 1st serve effectiveness is higher than their 2nd serve points won, a player should definitely consider it as a tactic. If the numbers are similar, let’s say there’s less than a 2% difference then a player should maybe consider it. If the 1st serve effectiveness is notably lower then a player should probably keep doing what they are doing.

I’ve subtracted 2nd serve points won from 1st serve effectiveness for every player in the ATP and WTA top 50. Two players from the ATP and WTA fall into the should definitely consider category (1stEff > 2nd%). 6 ATP and 4 WTA players fall into the should maybe consider category (1stEff > 2nd% - 2%). Here is a table showing these players split by their categories:

The table is full of players that either have an effective 1st serve, a poor 2nd serve or both. The first name that sticks out is Bublik himself. Is this evidence that the two 1st serves tactic doesn’t work? Maybe, but Bublik doesn’t use the tactic exclusively. It could, on the other hand, be evidence that he should fully commit to it (Bublik having the lowest 2nd serve points in the ATP top 50 would throw cold water on that theory, however). For me, Bublik doesn’t make enough 1st serves to fully commit to the tactic yet. His 1st serve % (60.3%) is ranked 40th in the top 50. He does, however, have one of the fastest 1st serves in the top 50. Were he to take a little pace off it and improve the 1st serve %, the two 1st serves tactic could quickly become more fruitful.

Looking at the other names in the definitely category: Eubanks has a fantastic 1st serve, but struggles to back up his 2nd serve when he gets drawn into rallies. Kasatkina and Sorribes Tormo both have slow 2nd serves that regularly get taken advantage of by their opponents, while also having a very high 1st serve % (3rd and 4th in the WTA top 50 respectively). If your 2nd serve is getting eaten up it might be time to stop looking at it just as a means to start the point, and instead as a means to get on the front foot, even if that means giving up more double faults.

It’s a similar tale for the players in the maybe category. Zverev probably has the best 1st serve on tour at the moment. Despite his struggles in the last year, Auger-Aliassime’s 1st serve has remained solid. Medvedev’s 2nd serve points won currently sees him ranked 47th out of 50 (back in 2019 he was ranked 10th). It’s a worrying trend for him. Every year since 2019 he has won fewer 2nd serve points than the previous year. Hurkacz’s dominance on 1st serve points and lack thereof on 2nd was laid out in the last article. Keys has the highest 1st serve effectiveness in the top 50. Putintseva is 3rd and Azarenka is 10th. Alexandrova wins just 42.2% of her 2nd serve points (5th lowest in the top 50).

Picking Your Moments

So we’ve established that a few players could benefit from adopting the two 1st serves approach. But how about saving it for certain scenarios? For example, break points. Using 1st serve % we can estimate total serves in play (EstSIn) adopting the two 1st serves approach. Multiplying this number by 1st serve points won gives an estimate for service points won (EstSPW). We can now compare this number to a player’s break points saved (BPSvd%). However, this time I’m going to change the criteria slightly. On tour, the average break points saved is always lower than service points won (2% lower on average over the last 10 years). This makes sense. If a player is facing more break points than usual one day, their service points won will naturally be lower than usual. With lower service points won, you would expect lower break points saved. With that in mind, if the estimated service points won is more than 2% higher than the break points saved, a player should definitely consider the two 1st serves tactic. If the difference is in that 0-2% range, they should maybe consider it. If the estimated service points won is lower, there’s probably no need for a change.

Subtracting break points saved from estimated service points won gives some very interesting results*. 12 players from the ATP and 6 from the WTA fall into the should definitely consider category (EstSPW - 2% > BPSvd%). 12 ATP and 14 WTA players fall into the should maybe consider category (EstSPW > BPSvd%). Here is a table showing the players in the definitely category.

*Full table can be found at the end of the article

Some familiar names on the ATP table, however, none of the WTA players from the previous table make it to this one. Pretty much all of the players here underperform on break points. That is to say, they save fewer break points than would be expected based on their service points won numbers. The numbers should be very concerning for Eubanks and Auger-Aliassime especially. Both of them, along with the other players towards the top end of the spectrum, should seriously consider adopting the two 1st serves approach on break points.

There’s no set rule for why these players perform worse on break points, but looking at the numbers it’s fair to say that they tighten up more than most. Going all out on a 2nd serve would certainly alleviate the pressure of getting drawn into a long rally. Yes, you have more chance of double faulting. But if a player can step up to the line accepting that might be the case, it could well free them up. I don’t believe every player here should adopt the two 1st serves approach on break points or even in “big” points in general, but it’s certainly something to consider.

. . .

It’s 9:00 pm on a Thursday night in Paris. Out on Court 5, 17th seed Alexander Bublik finds himself two sets down to Jan-Lennard Struff in the second round of Roland Garros. The crowd haven’t seen much of a contest. Bublik is serving at 0-1 30-40 and hopes are not high for this match to go the distance. His 1st serve doesn’t find its target. Bublik laces the 2nd serve. Double fault. 20 minutes later Struff is through to the third round.

Appendix

Cover photo adapted from Hameltion, CC BY-SA 4.0, via Wikimedia Commons

The data used for this article was for the last 52 weeks (as of the start of Roland Garros) and only includes matches against top 100 opponents. As ever, data came from the gold mine that is tennis abstract.

Abbreviations

• 1stIn = 1st serve %

• 1st% = 1st serve points won

• 2nd% = 2nd serve points won

• 1stEff = 1st serve effectiveness

• SIn = Total serves in play

• EstSIn = Estimated total serves in play

• SPW = Service points won

• EstSPW = Estimated service points won

• BPSvd% = Break points saved

1st serve effectiveness

1stEff = 1stIn*1st%

Estimated total serves in play

EstSIn = 1stIn + 1stIn*(1 - 1stIn)

Estimated service points won

EstSPW = EstSIn*1st%

Tables

Estimated service points won - break points saved for ATP top 50. Rk is ATP rank at time of writing. Filters used: Tour-level matches, vs top 100 opponents.

Estimated service points won - break points saved for WTA top 50. Rk is WTA rank at time of writing. Filters used: Tour-level matches, vs top 100 opponents.