I found 388 tiebreaks from the last eight ATP slams. For each one, I compared each player’s winning percentage on serve during the first 12 games of the set to his winning percentage on serve during the tiebreak. If players were robots, there might be a difference between the set and the tiebreak for any given match, but in general, the numbers should be the same.
But players aren’t robots. As it turns out, players win more return points than expected during tiebreaks.
So, to make a broad generalisation, serves are slightly LESS reliable in tie-breaks than usual. Whilst it's true that some players with extraordinary serves can use them well in these situations (Isner, Sampras, Federer), this is not a function of tie-breaks structurally favouring servers. It is a function of those players being far better than average at increasing their level in tie-breaks (and not only on the serve).
I have another explanation for the interesting observation the blogger has made comparing return points won in tiebreaks and in service games :
when a service game is easy, many returners just let it go, which articially inflates the average stats about points won on serve. In a tiebreak, players know that all points are important then returners are much more concentrated.
It's also one of the explanations imo why statistically, returners win more often (around 2 points difference according to my observation) the breakpoints than the servers comparing to other points (another reason is that usually when a returner has a breakpoint, it also means he has momentum on his side as it's on average relatively rare to have a breakpoint).
This stat could also be explained by "being harder to serve well on breakpoints" as you said but I have another explanation as I said.
Then I would very much like to know the magnitude of the blogger's observation (how many percent difference ?) but unless the magnitude is high, I don't think this stat shows that it's hard to serve on a breakpoint or during tiebreaks, as it could be explained by other reasons.
I don't have any statistical evidence for the opposite though, just an impression, the same kind of impression which makes you and me think that it's harder to build points under pressure as it needs more "brain action".
But I just say there may be other explanations for the statistical observations the blogger has made : that's the problem with stats, interpretations can be very diverse and quite often someone thinks or says he has this interpretation and then he says "the statistical observation proves that" whereas the explanation for his stat may be another one and actually he proved nothing (this is one of the explanations why stats are so often unfairly used by politicians or people who want to show something, and why a bloody politician -Disraeli if I remember- said one day "there are two ways to lie : lying and stats" - what an irony that many people say that statisticians are lyers and manipulators because of a politician's word
Thanks for the info. It shows how strange tie-breaks are. Without the information, just knowing a players circumstances, form, confidence etc... I think it would be pretty much impossible to successfully predict which players would have a good tie-break record and which ones would not.
the fact that tiebreaks go so much by streaks proves imo that it's a very confidence-oriented moment of the game : often players who are good at that are players who are very self-confident. Fed is one example. And I also think of Istomin : in tiebreaks, he displays to me a face of very positive confidence, and no doubt he has a good tiebreak stat. And of course having good stats in tiebreaks in the past helps to increase that confidence, hence the streaks.
To say another general thing I think about tiebreaks :
what I like about them is that usually in tennis, it's the better player who wins the tiebreaks, that is the player who won more % points on return than his opponent during the set, or sometimes, that's another scheme, who became superior and got the momentum in the end of the set.
It's the opposite from penalty-sessions in football (soccer) which, according to my observation, rather lead to the opposite result : teams which have dominated the match more often lose the penalty-session than the opposite, I think because for mental reasons teams which had been dominated are "happier" to be there and also their goalkeeper may have been more trained during the match. Which is something I really dislike about those sessions in football. Fortunately in tennis, it's rather the opposite