Everyone who has sent a message of encouragement has had one land badly. The words were fine. The timing was not. Sent once a situation is already decided, encouragement reads as pity. Sent too early, it reads as doubt.
Sport puts numbers on that. Tennis puts unusually precise ones on it, because the state of a match is recorded point by point.
Take a player who has lost the first set. Across 116,382 completed best of three matches, that player goes on to win 16.52% of the time. So the starting position for anything you might say to them is about one in six.
What changes it is the next twenty minutes.
The same deficit, very different situations
Follow that player into the second set and the figure travels a long way.
Having levelled the match by taking the second set 6-1, the chaser wins 58.53% of matches, over 3,682 of them. Taking it 6-2 gives 56.38%. Taking it 6-3 gives 52.19%. Taking it 6-4 gives 50.16% over 8,132 matches, which is a coin toss.
Now the other direction. Trailing the second set 4-5, the chaser wins 8.17% across 26,514 matches. At 3-4 it is 9.51%. At 0-1 it is 10.36%.
Same player, same lost set, and a spread from 58.53% down to 8.17%.
Why “still in it” is usually wrong
The phrase people reach for is that someone is still in it. This data says the phrase is accurate in a narrow band and misleading on both sides of it.
At one game all in the second set the chaser wins 17.66% of the time, over 63,405 matches. That is the most common state in the whole dataset. It is also barely different from the 16.52% they carried into the set. Nothing has happened yet, so encouragement there is premature.
Three games in at 3-0 up, it is 45.22% over 8,755 matches. That is where words would mean something, and almost nobody sends them, because the scoreboard still shows a set down.
The trap in reading a score
One warning, because it catches careful people.
A score of one all can mean one game all or one set all, and the two are nowhere near each other. One game all in the second set gives the chaser 17.66%. One set all gives 50.16%. Both get typed as one all in a message. Thirty two points separate them.
What to take from it
The useful part is not about tennis.
A message is read against the state of the thing it is about. The same sentence is kind at one moment and hollow at another, and none of that is in the words.
So ask what the situation actually is, instead of what it looks like from outside. A set down and 3-0 up looks worse than one game all, and it is nearly three times better.
The tennis figures here are published as structured data by Live Tennis API under CC BY 4.0. Every percentage carries its own match count, so each one can be checked.