This has the answer if you are bothered to read it. http://www1.fee.uva.nl/pp/klaassen/index_files/service_statistician_web.pdf If you're not bothered, here is the final paragraph "It is also commonly believed that serving first in a set provides an advantage. This is not true, except in the first set. The reason for this advantage in the first set can be completely accounted for by the "first game effect": fewer breaks occur in the very first game of the match. Because of the "first game effect" it is advisable for most players to elect to serve when they win the toss, not to receive."

On the other hand some players elect to receive at the start because they want to settle in the match and see if they can get an early break before the oppononent settles into his serving. The last thing you want as a server is to get broken in the first game.

When you start serving for a set, you automatically are in the lead. 5-4, after you held your serve you only need to break your opponent to win the set. 4-5, you can only hope to break your opponent's serve to take the lead and possibly serve for the set.

The only advantage is the mental aspect. If you serve first and get broken, it's bad, but there's a chance for you to break back in the next game. If you serve second, and are serving to stay the match, you cannot lose serve or the match is over.

Serving 2nd means that later on in the set a break point can also be a set point (or a match point). So if you aren't very good at dealing with big points then you're at a disadvantage because it's even more of a pressure point, you might be even more liable to 'choke' the break point.

Strictly speaking, though, there is no advantage in serving first. If each player holds serve 6 times each, there's a tiebreak, or if it's a fifth set, then they carry on until either somebody breaks for the match or breaks and then serves out the match. The advantage bit is all in the mind. If it's on serve at say 4-3, the person who's 4-3 up has already held serve an extra time so the current server feels under pressure to match that. The receiver, on the other hand, will take some chances on the points.

That depends. If the first set ends 6-0, 6-2, 6-4 or 7-5, the same player who served first in the first set will also serve first in the second set. If the first set ends 6-1, 6-3 or 7-6, the player who served first in the first set will receive first in the second set. The same process carries on for the other sets.

Mental is a big thing is it not ? Judging by how many top players are routinely called mental midgets/chokers etc.

Of course, the only possible advantage is "mental". The rules are not bent to favor the first server. My question was, if such "perceived" mental advantage is born out by the number, or it's just "perceived", not actual, advantage. The Magnus & Klaassen paper posted above is really interesting. The authors were correct in their premise that the player who starts serving in a non-first set is usually the weaker player. (The raw percentages show that the player who serves first in a non-first set is more likely to lose!) So they use a Bayesian model to compare the probability of a player who won the previous set to win the following set (a) when the serve vs. (b) when they return. They conclude that there is no significant difference between (a) and (b), thus no actual statistical basis for the popular belief. What is interesting is that, between seeded players, at Wimbledon (because the raw data came from Wimbledon matches), if you won the third set or fourth set, your chance of winning the fourth set or the fifth set, respectively, is really abysmal. Especially if you won the fourth set, your chance of winning the fifth is 0% when you serve and 11.9% when you receive. I'd have thought that the player winning the 4th set has the momentum going in the 5th as well. But between non-seeded players, the player winning the 4th set has a slightly higher chance of winning the final set. I also wonder what the numbers would have been like for the player who lost the previous set.

The article is surprisingly ignorant in its analysis. It would be easy to prove the magnitude of the advantage in serving first, excluding mental and physical considerations. One simply needs to perform a Monte Carlo simulation (this was glaringly absent from the article). If the probability of winning any given game is, say, 60% for the server, then clearly the player serving first is statistically more likely to win the first set. This is because almost half of all sets end in an odd number of games. Any set ending in an odd number of games is more likely to be won by the player who served first in the set. Thus, it's no mystery why it's advantageous for winning the first set to serve first. For the remaining sets, the advantageous is less obvious. However, it is obvious that winning the first set makes a player statistically more likely to win the match. Therefore, it is clearly advantageous to serve first to improve your chances of winning the match.