Mens Tennis Forums banner

Is break point conversion a useless statistic in determining how a match went?

  • Yes, useless, many could come from only one game

    Votes: 4 15.4%
  • No, it shows the mental weakness of player X at key moments

    Votes: 10 38.5%
  • Not sure/it depends on the particular match

    Votes: 12 46.2%

  • Total voters
    26
1 - 20 of 27 Posts

·
Registered
Joined
·
2,501 Posts
Discussion Starter #1
People often go on about such and such a match where player X butchered the match because they converted only 3 from 17 break points or suchlike.

The problem with this analogy is that that statistic could be useless if there was a long back and forth deuce game or two where there were 9 or 10 break points squandered before player X actually won the game and broke serve. But in the overall match stats you would assume the player didn't break and missed a chance when really they didn't, or somehow a heavy total of break points somehow represented multiple chances to win sets or the match.

Agree?

People on this forum quote the final break point conversion from a match as proof that player X could have won the match or made it substantially closer, when sometimes there is no real proof of that at all.
 

·
Registered
Joined
·
1,111 Posts
I always thought this was misleading as well, and thought it'd be good to supplement the BP conversion stat with something like (x), indicating the total number of games in which the player had BPs.
Like, 3/17(4) - the player missed a lot of BPs, but out of 4 total games in which he had break points, he managed to successfully break in 3 of them.

Also, not really related, I always felt it seemed a bit 'unfair' in a way that if players converted on their 1st BP opportunity at 0-40, the stats only showed them getting 1 BP opportunity rather than 3. :D
 

·
Registered
Joined
·
11,734 Posts
You have a point, but at the end of the day it always comes to the mental strength/weakness.
 

·
Registered
Joined
·
3,864 Posts
Depends, as with many stats in sports. Can be very telling, or very misleading. The argument in the original post is solid. Even so, I don't think it's half as ridiculous as the obsession with winners on here.
 

·
Registered
Joined
·
20,731 Posts
People often go on about such and such a match where player X butchered the match because they converted only 3 from 17 break points or suchlike.
The thing is, people overestimate how far relatively routine events are from the usual - in other words, ascribe meaning to something that is just variance.


For example, if you're playing some big server who's winning 65% of their service points, and he performs the same way on bps, ye olde binomial says the chance of being broken >3 times from 17bps is 90%. With a weaker server (60%), it's about 95%. Hell, make your player serve at Volandri pace (say 55%), and it's still only 98%.


Given we have about 2800 matches a season, a 2% event will actually happen pretty regularly.
 

·
Registered
Joined
·
20,731 Posts
IMO, they should also have a Games broken/games with BP stat. That would be more telling.
Games broken % has more correlation with performance - (by definition) combines your effectiveness on bps with the number you're creating. You can see it here:

http://www.atpworldtour.com/Matchfacts/Matchfacts-List.aspx?c=9&s=0&y=0




Commentators (and probably fans) want all that narrative of "clutch", some sort of comic-book ability to master your fate by sheer force of will, but it has very limited actual meaning.
 

·
Registered
Joined
·
15,733 Posts
Stats are overrated in general.

Always use your eyes to know what's really going on.
agreed, stats can be incredibly misleading

BP conversion is one of the better stats imo but as the OP shows even that is dodgy to view

at the end of the day the only stat you need is the score :shrug:
 

·
Registered
Joined
·
27,181 Posts
Seconds taken between points is the only stat I care about.
 

·
Banned
Joined
·
9,580 Posts
Wasting 17 break points at '04 Wimbledon final wasn't significant?
How about 4 set points, to go up 2 sets and 0-40 chance in the 5th set of 2009 Wimbledon final wasted? No big deal?
 

·
Registered
Joined
·
3,640 Posts
Well it depends.
Generally speaking this stat is not very useful.

But for the purpose of answering a specific question - like "mental fragility" - this is the most important stat.

In that sense , its similar to H2H which specifically answers which player is better between 2 players.
 

·
Registered
Joined
·
3,971 Posts
I would say break point conversion is highly significant in about 50% of matches, relatively significant in 25% and insignificant in 25%.

That is just my gut feeling. True, there are many plausible scenarios where break points converted doesn't tell you much but I don't think each of them is very common.
 

·
Registered
Joined
·
6,721 Posts
I think people read too much into it sometimes. Often failing to convert BP's is due to someone saving them, rather than someone making silly errors when they have a chance. Of course that plays a part too, but failing to convert isn't solely a sign of mental fragility.
 

·
-Lifetime Scientist-
Joined
·
12,190 Posts
IMO, they should also have a Games broken/games with BP stat. That would be more telling.
This would be nice.

Was also thinking: average of BP conversion in each game over all games with BP.

(1/N)*sum_{i=1}^{N}[b_i/a_i],

where N are the games with BP, b_i is either 0 or 1 and a_i the number of BPs in each game.

Problem here is that when b_i is zero, it eats the information on BPs wasted.

Maybe divide the result of the formula above by the number of BPs wasted on games where breaking wasn't achieved.
 

·
Registered
Joined
·
21,308 Posts
There are situations where it can be misleading for sure, just like most stats.
Winners will be inflated against Karlovic/Raonic and deflated against the top defenders as well
 

·
Registered
Joined
·
9,116 Posts
It can be misleading for sure. A guy could go 2/17 in break points during a match and win 64 64, because he broke in one game after needing 10 BPs and another one needing 7.
 
1 - 20 of 27 Posts
Top