Mens Tennis Forums banner

1 - 18 of 18 Posts

·
justice for all
Joined
·
13,110 Posts
Yeah, as opposed to common sense and gut feeling of some people it doesn't make much difference when you look at big numbers.
 

·
Registered
Joined
·
3 Posts
Well more often that not the player that wins a set does it by serving for it. That means that on average, the person who won the previous set will serve second in the following set. So from sets 2-5 the better player is serving second more often than not which skews the stats a bit. So only the data for the first set is really meaningful as whether the better player serves first or second is completely random.
 

·
Registered
Joined
·
3,781 Posts
Well more often that not the player that wins a set does it by serving for it. That means that on average, the person who won the previous set will serve second in the following set. So from sets 2-5 the better player is serving second more often than not which skews the stats a bit. So only the data for the first set is really meaningful as whether the better player serves first or second is completely random.
Always find it funny when people with a couple of posts over a span of years post an interesting post (like yours here). That other thread that you posted in where the two dunces were trying to argue that total points won isn’t a significant stat because it is only the important points won that matters is absolutely cringeworthy...
 

·
justice for all
Joined
·
13,110 Posts
Well more often that not the player that wins a set does it by serving for it. That means that on average, the person who won the previous set will serve second in the following set. So from sets 2-5 the better player is serving second more often than not which skews the stats a bit. So only the data for the first set is really meaningful as whether the better player serves first or second is completely random.
That's an interesting observation.
 

·
Registered
Joined
·
4,553 Posts
I always chose to serve first in the first set and won 80% of my opening service games this year (around 90% if we include practice sets) despite being the inferior player in most of them. Finally my brilliant choicemaking is backed up by scientific prove.
 

·
justice for all
Joined
·
13,110 Posts
I always chose to serve first in the first set and won 80% of my opening service games this year (around 90% if we include practice sets) despite being the inferior player in most of them. Finally my brilliant choicemaking is backed up by scientific prove.
Are you a servebot? :wink2:
 

·
Registered
Joined
·
44 Posts
Well more often that not the player that wins a set does it by serving for it. That means that on average, the person who won the previous set will serve second in the following set. So from sets 2-5 the better player is serving second more often than not which skews the stats a bit. So only the data for the first set is really meaningful as whether the better player serves first or second is completely random.
This is correct; it's also the same reason the score 6-3, 6-4 occurs much more often than 6-4, 6-3.
 

·
Registered
Joined
·
12,644 Posts
Funny. I can't remember but some troll here had a theory that Rafa lost the match because of the coin toss. He was an amazing troll, but he got banned :crying:. Help me out here, please :wink2:.

It was one of the funniest threads on MTF. I hope it still exists.
 

·
Registered
Joined
·
4,553 Posts

·
Registered
Joined
·
2,463 Posts
Funny. I can't remember but some troll here had a theory that Rafa lost the match because of the coin toss. He was an amazing troll, but he got banned :crying:. Help me out here, please :wink2:.

It was one of the funniest threads on MTF. I hope it still exists.
https://www.menstennisforums.com/2-general-messages/185899-coin-toss.html

This one? :p
 

·
Registered
Joined
·
14,076 Posts
I could bust that whole article right now, but in stead of a forum post I actually want to publish the research myself as well.


Believing it has no impact in a slam 5th set is fantasy/wish think.


You don't even need the numbers, should be obvious to anyone its easier to serve at 5-5 and not 5-6.
 

·
justice for all
Joined
·
13,110 Posts
I could bust that whole article right now, but in stead of a forum post I actually want to publish the research myself as well.


Believing it has no impact in a slam 5th set is fantasy/wish think.


You don't even need the numbers, should be obvious to anyone its easier to serve at 5-5 and not 5-6.
That's what common sense/ gut feeling says. But do figures support this theory? I remember a thread with some data published that denied it - don't have time to look for it now but I'm almost sure there was some data on it.

If you want to do it yourself, I think to have statistically relevant data, you would need to investigate at least 300 5-set matches.

Additional factors that would might be taken into account/ weighted in:

- momentum? (who won set no 4?)
- rank/ seed? (a higher ranked player has higher probability to win a set)
- experience? (a player with more 5-set matches under his belt has some advantage too)
- other?
 

·
Registered
Joined
·
2,366 Posts
Well more often that not the player that wins a set does it by serving for it. That means that on average, the person who won the previous set will serve second in the following set. So from sets 2-5 the better player is serving second more often than not which skews the stats a bit. So only the data for the first set is really meaningful as whether the better player serves first or second is completely random.
This is correct; it's also the same reason the score 6-3, 6-4 occurs much more often than 6-4, 6-3.
I could bust that whole article right now, but in stead of a forum post I actually want to publish the research myself as well.

Believing it has no impact in a slam 5th set is fantasy/wish think.

You don't even need the numbers, should be obvious to anyone its easier to serve at 5-5 and not 5-6.

[Citation needed]
 

·
Registered
Joined
·
10,400 Posts
At t 300 matches your error bars would be about 5-6 percent. You'd be better off with 1000 matches since the difference is likely lower than your error. That would give you an error of about 2 percent.
 
1 - 18 of 18 Posts
Top