The Official Fifth-Set Tiebreak Thread

Should all of the Slams have a tiebreaker in the fifth set?

  • Yes

    Votes: 9 25.7%
  • No

    Votes: 26 74.3%

  • Total voters
    35
  • This poll will close: .

skip1969

G.O.A.T.
With the 2017 US Open right around the corner, I thought the time was right to introduce a statistical thread about best-of-five set matches. Part of the reason is to have a thread that can be bumped every time the topic comes up . . . and it does come up. If we don’t bring it up here on the board, rest assured that John McEnroe will bring it up. He has an obsession about the fifth-set tiebreaker, and cannot understand why it’s not used at The Australian, Roland Garros, or Wimbledon. Since he likes to give his opinions as if his opinion are facts, I thought I’d start a thread with actual facts in it.

I’m not normally a numbers cruncher, but I’ve compiled some data on five-set matches over the last 10 years (with 2017 being year 11). Why ten years? No reason. I wanted enough data to post, but I was too tired to go further back in time.

McEnroe (and others) assert that every Slam should have a fifth-set tiebreaker like the US Open. My counter argument has always been that matches that go beyond 6-games-all in the fifth are so few and far between, that I don’t think the rules need to be changed. As a sidenote, I also like the fact that the Slams have their own autonomy and can keep their unique traditions. The tour is homogenized enough.

But that is my subjective opinion, and this is a thread about facts and numbers, not opinion. Now of course, a ridiculously long match (like the infamous Isner-Mahut 70-68, 3-day marathon match from Wimbledon 2010) makes some tennis fans clamor for tiebreaks at all the Slams. But a match like that is once in a lifetime. The stats for the past 10 years show that five-set matches make up a small percentage of all matches played. And those that go past 6-all are super rare. In this data collection, I will refer to those matches as overtime (OT) matches from The Australian, Roland Garros, and Wimbledon, versus those with the final-set tiebreaker (TB) at the US Open.
 

skip1969

G.O.A.T.
The Australian Open/Roland Garros/Wimbledon
2007-2016
The ten-year trend


3,810 total matches
3 set matches: 1,903 = 50% of all matches
4 set matches: 1,065 = 28%
5 set matches: 681 = 18%
Retirements & walkovers: 161 = 4%
OT matches: 116 = 3%

The US Open
2007-2016
The ten-year trend


1,270 total matches
3 set matches: 601 = 47% of all matches
4 set matches: 368 = 29%
5 set matches: 221 = 17%
Retirements & walkovers: 80 = 6%
TB matches: 37 = 2.9%
 
Last edited:

skip1969

G.O.A.T.
The Australian Open/Roland Garros/Wimbledon
2012-2016
The five-year trend


1,905 total matches
3 set matches: 946 = 50% of all matches
4 set matches: 528 = 28%
5 set matches: 351 = 18%
Retirements & walkovers: 80 = 4%
OT matches: 61 = 3.2%

The US Open
2012-2016
The five-year trend


635 total matches
3 set matches: 288 = 45% of all matches
4 set matches: 188 = 30%
5 set matches: 116 = 18%
Retirements & walkovers: 43 = 7%
TB matches: 18 = 2.8%
 
Last edited:

skip1969

G.O.A.T.
The Australian Open/Roland Garros/Wimbledon
2007-2011
The five-year trend


1,905 total matches
3 set matches: 957 = 50% of all matches
4 set matches: 537 = 28%
5 set matches: 330 = 17%
Retirements & walkovers: 81 = 4%
OT matches: 55 = 2.8%


The US Open
2007-2011
The five-year trend


635 total matches
3 set matches: 313 = 49% of all matches
4 set matches: 180 = 28%
5 set matches: 105 = 17%
Retirements & walkovers: 37 = 6%
TB matches: 19 = 2.9%
 
Last edited:

tacou

G.O.A.T.
Great stats! Seems like 3% roughly.
I hope I do not sound lazy/ungrateful, but you should check how many of the overtime matches went to, say, 12-10 or higher, and what are the typical results of a player who wins the overtime match in the following around. I think these would be helpful.

I believe at 12-12 there should be a TB. Just my opinion.
 

skip1969

G.O.A.T.
Great stats! Seems like 3% roughly.
I hope I do not sound lazy/ungrateful, but you should check how many of the overtime matches went to, say, 12-10 or higher, and what are the typical results of a player who wins the overtime match in the following around. I think these would be helpful.

I believe at 12-12 there should be a TB. Just my opinion.
LOL. You don't sound ungrateful at all. I have a pile of papers that I scribbled the stats on. I didn't really think I'd get that deep into the stats. But I did more than I first planned on doing.

I didn't go specifically into the scores of the five-set matches. I guess I'm not neurotic enough. Like I said, I'm no numbers guy. ;) But I did do some winning percentages. Gimme a sec here.
 

skip1969

G.O.A.T.
So, to partly answer @tacou , here are the winning percentages into the next round for those players who won in five sets, whether they won in five sets with the fifth set going going the regular distance (R), five sets with the fifth set going past 6-games all (OT), or five sets with a tiebreaker in the fifth (TB).

The Australian Open/Roland Garros/Wimbledon
2007-2016
The ten-year trend

Winners in five regular (R) sets went 226-334 in next round = 40% winning percentage
Winners in five overtime sets (OT) went 30-85 in next round = 26% winning percentage

2007-2011
The five-year trend

Winners in five regular (R) sets went 116 -157 in next round = 42% winning percentage
Winners in five overtime sets (OT) went 10-43 in next round = 19% winning percentage


2012-2016
The five-year trend

Winners in five regular (R) sets went 110-177 in next round = 38% winning percentage
Winners in five overtime sets (OT) went 20-42 in next round = 32% winning percentage



The US Open
2007-2016
The ten-year trend

Winners in five regular (R) sets went 82-100 in next round = 45% winning percentage
Winners in fifth-set tiebreak (TB) went 7-30 in next round = 19% winning percentage


2007-2011
The five-year trend

Winners in five regular (R) sets went 43-42 in next round = 50% winning percentage
Winners in fifth-set tiebreak (TB) went 4-15 in next round = 21% winning percentage


2012-2016
The five-year trend

Winners in five regular (R) sets went 39-58 in next round = 40% winning percentage
Winners in fifth-set tiebreak (TB) went 3-15 in next round = 17% winning percentage
 
Last edited:

tacou

G.O.A.T.
Whoa. Interesting. Obviously the 5th set TB sample pool is smaller, but winners there seem to fair even worse in the following round. Unsurprisingly, OT winners don't exactly do great either.

I need to meditate on this a bit.
 

Fedeonic

Hall of Fame
Whoa. Interesting. Obviously the 5th set TB sample pool is smaller, but winners there seem to fair even worse in the following round. Unsurprisingly, OT winners don't exactly do great either.

I need to meditate on this a bit.
Maybe the players who win in a 5th set TB spend more mental energy than the players who win in the Adv. 5th set. Understandable because in a 5th set TB you literally have no margin of error in ANY point, while you can spend several return games in the Adv. set doing nothing if you still keep serving well enough.
 

Tornes

Semi-Pro
Maybe the players who win in a 5th set TB spend more mental energy than the players who win in the Adv. 5th set. Understandable because in a 5th set TB you literally have no margin of error in ANY point, while you can spend several return games in the Adv. set doing nothing if you still keep serving well enough.
Maybe too. However I think at least part of the reason could be that Adv. 5th set is better in deciding who is the better player and therefore who has better chances in next round (eventhough he spend little more energy in this one).
 

reaper

Legend
I would go tie break at 18-18 in the 5th...would be very rarely used but avoids the Isner-Mahut situation where the winner has absolutely chance in the next match and the schedule is completely thrown.
 

skip1969

G.O.A.T.
I think one of my big takeaways is that just about 50% of all matches are straight-setters, which seems to me to be more of a 'problem' than the matches that go beyond 6-all in the fifth. Extended/OT/TB matches make up 3% of all matches.
 

tacou

G.O.A.T.
I would go tie break at 18-18 in the 5th...would be very rarely used but avoids the Isner-Mahut situation where the winner has absolutely chance in the next match and the schedule is completely thrown.
interesting… This would not seem to address the two main benefits of a deciding set TB though; healthier players for the following round and more manageable match times. that said my idea for one at 12 all might have the same issue.
 

Steve0904

Talk Tennis Guru
I believe I've changed my position on this since I first joined here, but now I think it's time for 5th set TBs. I'm tired of overly long marathons. I'd take a TB at 9-9 or something like that as a concession.
 

reaper

Legend
interesting… This would not seem to address the two main benefits of a deciding set TB though; healthier players for the following round and more manageable match times. that said my idea for one at 12 all might have the same issue.
I think there should be attrition in slams, so I think a 5th set tie break should only be played after a very long 5th set. It should be very difficult for that player to win their next match. If you play the 5th set tie break at 6-6, It's actually a "softer" situation than a player might encounter in a regular tour event. If a player plays consecutive 3 setters on the regular tour, that's 6 sets in two days, then they play the next day. A 12-12 tie breaker at least replicates 6 sets, 18-18 makes it 7 sets. 5th set tie breaks are needed to prevent the ridiculous Isner-Mahut situation, but I'd be much more comfortable with 12-12, 15-15 or 18-18 than at 6-6.
 

tennisaddict

Bionic Poster
While 3% may seem less for OT matches at first look , that is like 5 matches out of a total 127 matches. That is by no means a small amount given that these players barely even have a chance to win next round even with a 5th set TB

The scheduling nightmare it causes is bad. Sometimes this results in bye , retirement or bad schedule for the next round opponent and throws them in a disarray

We need a TB just like USO, Except for the SF and Finals .
 
Last edited:

tacou

G.O.A.T.
I think there should be attrition in slams, so I think a 5th set tie break should only be played after a very long 5th set. It should be very difficult for that player to win their next match. If you play the 5th set tie break at 6-6, It's actually a "softer" situation than a player might encounter in a regular tour event. If a player plays consecutive 3 setters on the regular tour, that's 6 sets in two days, then they play the next day. A 12-12 tie breaker at least replicates 6 sets, 18-18 makes it 7 sets. 5th set tie breaks are needed to prevent the ridiculous Isner-Mahut situation, but I'd be much more comfortable with 12-12, 15-15 or 18-18 than at 6-6.
Yeah this is how I view it too. For me, once you have played enough games in one set to arrive at a TB for a second time (24 games), both players have proven themselves. Time to pick a winner.
 

skip1969

G.O.A.T.
While 3% may seem less for OT matches at first look , that is like 11-12 matches out of a total 127 matches. That is by no means a small amount given that these players barely even have a chance to win next round even with a 5th set TB . . .
Did you look at the winning percentages? As it stands, the chances of winning your next round after you go five are already down to 40%. They go down to 26% if you go OT. The obvious takeaway is the same any way you look at it (tiebreak or no tiebreak): Your chances of winning the next round probably increase by winning in 3 or 4.

The scheduling nightmare it causes is bad. Sometimes this results in bye , retirement or bad schedule for the next round opponent and throws them in a disarray . . .
Every match that went past 6-all was factored into the totals, no matter if the fifth set went 8-6 or 70-68. So it doesn't necessarily mean that just cos a match went OT, it caused any scheduling nightmares, byes or retirements. Just sayin'. ;)
 

tennisaddict

Bionic Poster
Did you look at the winning percentages? As it stands, the chances of winning your next round after you go five are already down to 40%. They go down to 26% if you go OT. The obvious takeaway is the same any way you look at it (tiebreak or no tiebreak): Your chances of winning the next round probably increase by winning in 3 or 4.


Every match that went past 6-all was factored into the totals, no matter if the fifth set went 8-6 or 70-68. So it doesn't necessarily mean that just cos a match went OT, it caused any scheduling nightmares, byes or retirements. Just sayin'. ;)
I also think players could plan their matches differently if they knew going into the 5th set , there is a TB. You have a match against ISner , Karlovic or matches between two serve bots like Querrey and Anderson, it can go endlessly . It is rare that matches between defenders go long .

Tsonga lost two matches to Querrey recently in 5th set , all out of nowhere .

Even if there is no TB, they need to alternate serve order after 6-6

It is extremely loaded in favor of player serving first
 

Al Czervik

Hall of Fame
I used to think fifth set tiebreaks are stupid, but I am all in now. If a guy can't break through after 12 games, finish the match. Or, maybe play it at 10 all.
 

ledwix

Hall of Fame
Thanks for the data. Among overtime records and fifth-set breaker records, the difference in results isn't quite statistically significant, since 37 is actually a pretty small sample size of US Open overtime matches.

Code:
http://www.wolframalpha.com/input/?i=proportion+test&rawformassumption=%22FSelect%22+-%3E+%7B%7B%22ProportionDifferenceTest%22%7D%7D&rawformassumption=%7B%22F%22,+%22ProportionDifferenceTest%22,+%22p0%22%7D+-%3E%220%22&rawformassumption=%7B%22F%22,+%22ProportionDifferenceTest%22,+%22n1%22%7D+-%3E%22115%22&rawformassumption=%7B%22F%22,+%22ProportionDifferenceTest%22,+%22phat1%22%7D+-%3E%2230%2F115%22&rawformassumption=%7B%22F%22,+%22ProportionDifferenceTest%22,+%22n2%22%7D+-%3E%2237%22&rawformassumption=%7B%22F%22,+%22ProportionDifferenceTest%22,+%22phat2%22%7D+-%3E%227%2F37%22
At a significance level of 10%, which is not very strict at all, we still cannot conclude that fifth-set tiebreaks are worse for players than overtime matches or vice versa.

A 12-all tiebreak does sound fine to me. Six sets of tennis is enough tennis. But in the final, semi-traditional rules would also be nice.
 

skip1969

G.O.A.T.
Thanks for the data. Among overtime records and fifth-set breaker records, the difference in results isn't quite statistically significant, since 37 is actually a pretty small sample size of US Open overtime matches . . .
You're saying the sample size is too small to form any kind of real conclusions about a player's chances after a fifth set tiebreak, right? Doesn't that just go to prove my original point, which is that matches that go to 6-all in the fifth are so rare anyway, that it's silly to fret about the three Slams that don't have a tiebreaker in the fifth?
 

West Coast Ace

G.O.A.T.
You're saying the sample size is too small to form any kind of real conclusions about a player's chances after a fifth set tiebreak, right? Doesn't that just go to prove my original point, which is that matches that go to 6-all in the fifth are so rare anyway, that it's silly to fret about the three Slams that don't have a tiebreaker in the fifth?
Yeah, the sample size is too small. Stats 101 was too many yrs ago to remember.

The next round opponents would be interesting to review; could just be a higher ranked player or someone they had a bad H2H with.

But great thread and miles ahead of the JMac, BG, Navratilova, Shriverrants about rules changes. At least some level of ananlysis not their silliness.
 

skip1969

G.O.A.T.
@ledwix and @West Coast Ace

Again, I am NO math major. I went to art school. :D

But 10 more years of US Opens only gives you a sample size of 73 matches that went to a fifth set tiebreak.

1997-2006
The ten-year trend


1,270 total matches
3 set matches: 586 = 46% of all matches
4 set matches: 399 = 31%
5 set matches: 195 = 15%
Retirements & walkovers: 53 = 4%
TB matches: 36 = 2.8%

2007-2016
The ten-year trend


1,270 total matches
3 set matches: 601 = 47% of all matches
4 set matches: 368 = 29%
5 set matches: 221 = 17%
Retirements & walkovers: 80 = 6%
TB matches: 37 = 2.9%
 
Last edited:

skip1969

G.O.A.T.
Yeah, the sample size is too small. Stats 101 was too many yrs ago to remember.
I feel you! LOL

The next round opponents would be interesting to review; could just be a higher ranked player or someone they had a bad H2H with.
Exactly. Because there are a million variables that could factor into a player going five sets one round and losing in the next. It's not simply a question of them being 'tired' from a long match. I post the winning percentages just as an aside, because I think it strengthens the case for any player increasing his odds of winning the next round if he avoids going five, period.

1997-2016
The US Open
The twenty-year trend

Winners in five regular (R) sets went 167-210 in next round = 44% winning percentage
Winners in fifth-set tiebreak (TB) went 22-51 in next round = 30% winning percentage

1997-2006
The ten-year trend

Winners in five regular (R) sets went 85-110 in next round = 44% winning percentage
Winners in fifth-set tiebreak (TB) went 15-21 in next round = 42% winning percentage

2007-2016
The ten-year trend

Winners in five regular (R) sets went 82-100 in next round = 45% winning percentage
Winners in fifth-set tiebreak (TB) went 7-30 in next round = 19% winning percentage
 
Last edited:

abmk

Bionic Poster
Make it till 12-12 for the 5th set before a final TB - for matches before the final.

Play it out for the final.
 

Mr.Lob

Legend
Tiebreakers up until the finals. Then if you want the occasional 90-88 set...go for it. And the "it's up to them to break serve" argument doesn't cut the mustard in my book. It's expected to hold serve.
 

Raindogs

Hall of Fame
Although I hate the idea of a fifth set tiebreak, at this point I'm even more bothered by Mac's incessant droning on about it.
 

HAMPER7777777

New User
5th set TB. I don't Know about some people, but I have a job, and things to do, and just cannot spend 5+ Hours Watching 1 match. As usual, we do it best here in 'MURICA. 5th set TB, done deal, and move on. 1 of many reasons why the USO is my favorite slam! I don't have All Day to Watch John Isner (who I like BTW), going to 70-68, where 90% of games were either 40-0, or 40-15 (due to a double fault) anyway. Get it done, and move on. TB's are exciting, anyway.
 

ledwix

Hall of Fame
You're saying the sample size is too small to form any kind of real conclusions about a player's chances after a fifth set tiebreak, right? Doesn't that just go to prove my original point, which is that matches that go to 6-all in the fifth are so rare anyway, that it's silly to fret about the three Slams that don't have a tiebreaker in the fifth?
Yes. And even with the additional data you provided from the 1997-2006 period, there is still not enough evidence to say which format is better for a player's chances of advancing further, let alone that fifth-set tiebreaks are better.
 

Tornes

Semi-Pro
Yes. And even with the additional data you provided from the 1997-2006 period, there is still not enough evidence to say which format is better for a player's chances of advancing further, let alone that fifth-set tiebreaks are better.
All the evidence say that chances player would win are about the same for 5th set TBs and OT matches. So no true reason to change it. As Jeff Sackmann concludes in his older article at the same theme - Changing format to fifth-set tiebreaks would have little effect on future outcomes–it would just make those matches a bit more dependent on a lucky bounce.

Here is the article - maybe even little bit better analysis, using also factor of "expected wins" for the player who play OT and 5th set TB matches. www.tennisabstract.com/blog/2012/08/09/the-hangover-effect-of-a-marathon-fifth-set/
 

Tornes

Semi-Pro
All the evidence say that chances player would win are about the same for 5th set TBs and OT matches. So no true reason to change it. As Jeff Sackmann concludes in his older article at the same theme - Changing format to fifth-set tiebreaks would have little effect on future outcomes–it would just make those matches a bit more dependent on a lucky bounce.

Here is the article - maybe even little bit better analysis, using also factor of "expected wins" for the player who play OT and 5th set TB matches. www.tennisabstract.com/blog/2012/08/09/the-hangover-effect-of-a-marathon-fifth-set/
And for me the most interesting part - winners of OT matches were expected to win 43,4 % of their next matches, winners of TB matches only 38 %. It is for me pretty significant different and great argument for OT matches - the better players more often wins them.
 

skip1969

G.O.A.T.
Only 2 matches this year have gone to a 5th set tb. Wonder if that is a tournament low.
When the tourney finishes, I'll do the totals for 2017 and post them. And while I'm at it, I'll look to see what the low is. I've only done the past twenty years, but off the top of my head, yes, I think 2 ties for the lowest.
 
N

Nashvegas

Guest
There's some consensus being reached here along the lines of tie-breaks until the final, then no tie-break, and in the case of tie-breaks possibly going to a score like 12-12. Seems like maybe something most could support.

This idea will never happen, but what if they went by the clock to determine whether a tie-break would be used? Say the US Open limit is set at 4 hours. If it's 6-6 in the 5th set and the match is > 4 hours, go to a tie-break. If under 4 hours, keep going and play a tie-break after the match reaches 4 hrs.

The "problem" to be addressed is the unending match that wears out the players and screws up the schedule, and that's mostly a factor of time. So why play a tie-break to save time if a match gets to 5 sets in just 2 hours?
 
Top