Fed/Nadal rd 3

[maverick1]

I have no Tennis credentials, but I am going to play devil's advocate (some may prefer the term "arrogant idiot" ).

The concept of attaching a precise number to each player and calling it his "shot tolerance" doesn't seem useful at all.

Any player whose mental game is above 3.0 level should recognize the problem if his impatience is costing him matches and accordingly vary his "shot tolerance" depending on how a match is going. I am a lowly 3.5 and I do it. OK, "Recognize" may be a strong word for what I do, but I definitely guess that that may be the problem and I can change from trying end points quickly to waiting for opponent's errors, or vice versa.

If I can do it, I would expect that a pro would also vary his "shot tolerance" to suit his chances of winning.

If this concept were to be useful, the idea seems to be to use in a scenario like the following:
You lose to someone a couple of times in a row; then an astute observer who has been watching the matches tells you that when you kept the ball in play for 6 shots or more, you won 80% of the points. So you define the opponent's shot tolerance as 6, and the next time you play that player, you decide to not take any risks for the first 6 shots, and voila, you win!

Yeah, right. Why don't we first consider the obvious, that this opponent was born with a brain, and played the way he did because that style was winning him the match?

If he is losing to you the next time because of his unforced errors after long rallies, he would have to be pretty immature not to realize this by the middle of the first set.

It seems far-fetched to think that a professional tennis player "just doesn't want to play the point any more" after a certain number of shots. At best I can believe that a pro may calculate that the opponent is better than himself at the grinding game, so he will adopt a more aggressive strategy.

That is my 2 cents.
Now someone please prove me wrong, and bring some clarity to this topic.

[johnyandell]

There is a difference between intellectual analysis and reality.

You write:
It seems far-fetched to think that a professional tennis player "just doesn't want to play the point any more" after a certain number of shots.

Sure that seems to make sense, but it's not the reality of competitive tennis. It's sort like sports writers who have all the answers after their team loses. Looking from the outside they think it's all easy and obvious. And they are quick to demand perfection. But usually they are clueless about why what actually happened, happened.

Shot tolerance in tennis is something that is fluid but if you are an experienced competitor it's something you can feel in yourself and others.

As you compete at higher and higher levels, commit to winning, and then are actually able to win, the various emotional pressures and and limitations become all too real. To me it's all crystal clear in Elliot's article and that's why I published it.

It might be something some people can't really understand unless they were to experience it themselves.

[airforce1]

Sure you count, But...

But I have to disagree with counting each point and taking an average in a random match.

To be with the spirit of Elliot's article as I understand it and use this with my teams, you have to see a player in a match where he is tested by an opponent that will push him deep in a lot of points. Otherwise, you will never even see his shot tolerence because he will never get to it.

Once you view him in a match where he is forced to play out points, you have to look at the points where a player made an unforced error when not in an attacking position. If I serve and get a short ball reply, then hit an unforced error off a sitting duck, this tells you nothing about my shot tolerance. I was supposed to be aggressive with that ball.

So once you have the rally length on points ending in unforced errors and not in attacking position, you can look at the median or average length of said rallys to see which is more telling. Probably the median will be a better number, as it should take out much of the oddball points that are way longer or shorter than the others.

For example to use some numbers:
a player like Roddick hits 11 aces, 14 service winners and 10 rally point winners, for a total of 35 winners. He also has 25 unforced errors, with 5 of them coming on easy putaways missed, after a big serve.
That would give us 30 points ending in 1.5 strokes each and the other 10 winners in 3.5 strokes each.
Of the 20 unforced errors we have left after subtracting the 5 in attack position, 7 were on first 3 strokes and 3 were after 15 strokes.
taking these 10 out we are now left with 10 points played out to an unforced error where his shot tolerance was tested. Assuming that all of these ended at between 6-8 strokes,

we can see his shot tolerance is around 7 strokes (or 8 to really mark the outside limit, only 3 times in the match did he get past this limit)

On the other hand, if you just averaged all the rallys listed here, the tolerance would be a significantly lower number of only about 3.6
I think that you would agree that 3.6 would be pretty useless in this case, as making him hit only 4 balls each time you got into a rally, would not get you anything, in and of itself. BUT On the other hand, getting him to go past 8 balls in each rally would be a level that could really pay off big!

Granted, forced errors were left out of this, but that is because they normally tell you little about a players shot tolerance, as he was forced to miss before he reached his tolerance. (unless these points went as long or longer than the computed tolerance of 7-8, which from my charting is not the norm)
Granted also, I used a big server to exaggerate the difference to make the point, but there are many big servers out there and shot tolerance is very effective against this type player on the whole if you can get past his serve.
Just one more opinion

[johnyandell]

Hey, the only point I am making is this. No one has ever counted this out as far as I know and you might reevaluate it once you saw the raw numbers for all points. Or not.


I'm just curious about it. Curiosity is the start of everything new. It's driven all my work in the past and I'm not advocating some final solution. There are many ways to crunch numbers and maybe shot tolerance is only revealed in some of those. Or not. I find it interesting that a simple question/premise that would just yield some data has produced so many tortuous analytic/argumentative posts.

Rather than trying to play authority since it's my site, I was just being open about how I myself would approach trying to figure out more about a new topic. So let's all agree to disagree if that's what it takes! I'm not really interested in arguing about what the data might yield or not! I'm just interested in the data.

[jhm36]

Here's an idea. Instead of counting points from actual matches, how about Elliot or John or someone else who as a clearer idea than some of us on the thread of what shot tolerance is and how to work with it just generating some hypothetical numbers and showing us by made-up example what we're dealing with, kind of like airforce1 was doing? How about someone just assume numbers and explain what's going on and how shot tolerance was useful to tell us what happened and how we can apply that knowledge on the court with our own games?

Or, barring that, how about bringing up a few points in matches we've all seen recently in which hitting the shot tolerance wall caused someone, say Federer or Nadal, to lose or win a point? After all, we're talking about a very specific thing here, right, a person deciding to end a point because it's gone on too long and pushed him out of his shot tolerance comfort zone? There should be examples of hitting this wall in a match. What possibly were a few? Let's take the Federer-Nadal match at the Masters Cup, the most relevant match in this discussion. Any particular points stand out where Fed and Nadal may have hit their limits? Since we don't really have a range for those limits, I know we'd be working in the dark a bit, but why not give it a go?

[johnyandell]

You guys just don't want to do what I want you to do, do you? I'm joking, but seriously, every suggestion here is getting wackier. (Sorry Jim.) If I'd spent the time counting those matches instead of making these posts, I might my data by now.

Elliot went out and just counted shots. That's not a tough concept to grasp. It was good enough for him and it's good enough for me too.

When he says something like how many hard deep crosscourts can you hit and that's your shot tolerance, he doesn't (obviously) mean that's THE measure. What he means is that it's a good indication of what your max number is likely to be.

[lukman41985]

How hard is it to just count? This is frustrating to read...

[jhm36]

hmmm, I think John is trying to test our own shot tolerance by giving us grief for not understanding a stat that he won't even measure!

Now that's whacky.

[johnyandell]

Jim,

Don't think you understand my motives whatsoever. I have no interest in testing you or anyone else on anything.

I was actually hoping to get some help, and I thought it might be a fun project to get a few subscribers involved in the data development process, something that could lead to another article, and that's about it. You of course are welcome to your views, questions, ponderings etc. But that doesn't mean I share them. Better just to let this one drop. In the future if I decide I need to so something stats wise I'll do it myself.

John Yandell