PDA

View Full Version : Losses: The Matches That Don't Count


cynicole
Dec 22nd, 2001, 03:34 AM
I crunched these numbers a month ago and tried posting them but couldn't get the table to work properly so I quit. But now that wins are being discussed I thought "what the heck!"

Please note that all ranking statistics are year-end numbers from the Monday after Munich.


-------------------------LD--JC—-VW--MH--KC--SW—-JH—-JD—-AM—-MS—
Losses-------------------09--14--05--15--18--07--18--23--11—-10-
Walkovers----------------01--00--01--00--01--00--00--00--01—-00-
2-Set Losses-------------05--03--04--09--11--03--06--14--09--06-
3-Set Losses-------------04--11--01--05--07--04--12--09--03—-04-
Tiebreakers Won in Ls----03--02--00--01--00--01--02--00--01--00-
TBs Lost in Losses-------01--03--02--02--01--02--05--03--01--04-
Losses to Top 10 Players-08--10--02--12--10--07--10--14--06--06-
Ls to Outside 25 Players-00--01--00--01--04--00--02--07--03--03-
Average Rank Lost To-----05--13--11--10--15--04--15--21--15--20-
Median Rank Lost To------05--08--10--06--07--03--10--05--06--08-
Readjusted Rank Lost To--05--08--10--07--13--03--12--17--10--14-


Now in detail...


-------------------------LD--JC—-VW--MH--KC--SW—-JH—-JD—-AM—-MS—
Losses-------------------09--14--05--15--18--07--18--23--11—-10-
Walkovers----------------01--00--01--00--01--00--00--00--01—-00-

Of course, the fewer tournaments you play the fewer opportunities you have to lose. Out of those who played 17 tournaments, only Davenport had fewer than 10 losses.


-------------------------LD--JC—-VW--MH--KC--SW—-JH—-JD—-AM—-MS—
2-Set Losses-------------05--03--04--09--11--03--06--14--09--06-
3-Set Losses-------------04--11--01--05--07--04--12--09--03—-04-
Tiebreakers Won in Ls----03--02--00--01--00--01--02--00--01--00-
TBs Lost in Losses-------01--03--02--02--01--02--05--03--01--04-

I include whether the match was a 2- or 3-setter and tiebreakers to give an idea of how close the matches may have been. Of course, it's hard to tell from the scores if a match really was close or not but this was the least hectic way for me to gather data.

I don’t know what you get walking away from these numbers but after seeing this I actually appreciate Capriati’s efforts a bit more. She lost in straight sets just 3 times (fewer than Venus!) in 2001 compared to 11 3-set losses. Though she wasn’t great towards the end of the year it wasn’t all that easy to take her down.

(If you notice the disparity in Hingis’ numbers, it’s because she pulled out in the first set against Davenport at Filderstadt.)

<br />Now comes the more confusing stuff. I include these only because I bothered to do the math...


-------------------------LD--JC—-VW--MH--KC--SW—-JH—-JD—-AM—-MS—
Losses to Top 10 Players-08--10--02--12--10--07--10--14--06--06-
Ls to Outside 25 Players-00--01--00--01--04--00--02--07--03--03-

It depends on where you want to draw the line. I chose 25 but if I pick 20 or 30 some of the numbers do change significantly. The strangest exception is Dokic. She didn’t lose to anyone ranked between (and including) 20 and 37.

For the following, I counted the walkovers (simply because I was too lazy to reformulate my excel sheet). Counting the walkovers actually helped everyone involved except Davenport.


-------------------------LD--JC—-VW--MH--KC--SW—-JH—-JD—-AM—-MS—
Average Rank Lost To-----05--13--11--10--15--04--15--21--15--20-
Readjusted Rank Lost To--05--08--10--07--13--03--12--17--10--14-

The "Average Rank Lost To" is the total of all the ranks of all the players divided by the number of losses.

The "Readjusted Rank Lost To" is basically the same except that I knocked off both the highest and lowest rank from the total. It helped everyone but Davenport (who remains the same yet again).


-------------------------LD--JC—-VW--MH--KC--SW—-JH—-JD—-AM—-MS—
Median Rank Lost To------05--08--10--06--07--03--10--05--06--08-

This one is the dead center score... or at least what Excel told me was the median. Quite deceptive. Dokic and Davenport are tied with 5 but part of the reason why Dokic’s is so low is because she lost to Davenport 5 times and Capriati thrice. Venus and Henin are tied with the worst number. If you could see a correlation between them, tell me.

[ December 21, 2001: Message edited by: cynicole ]

[ December 22, 2001: Message edited by: cynicole ]</p>

TheBoiledEgg
Dec 22nd, 2001, 04:15 AM
thanx Nicole <img src="graemlins/kiss.gif" border="0" alt="[Kiss]" />

pretty good stuff there.

I guess everyone's a bit bored without tennis <img src="wink.gif" border="0">

Robbie.
Dec 22nd, 2001, 04:22 AM
Some very interesting results here...a deep and very thourough analysis. <img src="biggrin.gif" border="0">

Except doesn't Serena have the best median score (3)? <img src="wink.gif" border="0">

What I find most intersting is the average rank(or readjusted rank)lost to. I find readjusted better because it erases any anomalies in a sample area. That stat really shows that these players generally lost to quite highly quality players (although I think Monica's is probably a little biased because of her losses to Garbin and De Los Rios when she was injured). It also makes sense the Dokic is most liable to be upset by players outside the top 25.

As you said it is a testimony to capriati's toughness that she was beaten in straight sets only three times all year (twice to Venus, if I remember correctly). The most startling of those statistics is that Martina was often beaten quite easily in straight sets.

Anyways I never got medians <img src="redface.gif" border="0"> <img src="wink.gif" border="0"> . Averages give a much better indication of a statistics general performance.

[ December 21, 2001: Message edited by: tennisaddict ]</p>

cynicole
Dec 22nd, 2001, 07:08 AM
Oops <img src="redface.gif" border="0">

Thanks tennisaddict! I'm making corrections now.

Medians suck. I was experimenting with them...but it turns out it's trash.

Monica_Rules
Dec 22nd, 2001, 01:05 PM
Interesting fact here!

veryborednow
Dec 22nd, 2001, 02:05 PM
Jen lost quite a few matches compared to the other top players <img src="frown.gif" border="0">

Brian Stewart
Dec 22nd, 2001, 05:25 PM
Did you use the rankings at the time they played, or the final ranking?

cynicole
Dec 22nd, 2001, 09:38 PM
Brian Stewart - I used the rankings from the week after Munich (which was when I crunched the numbers).

There's a trade-off really. Venus' numbers get helped because three of the people that beat her moved up in ranking (Maleeva, Henin and Shaughnessy were much lower then than they are now). Whereas Seles and Henin take a hit because Tathiana Garbin became a non-entity around mid-year, got injured and slid down the rankings. (Well, I think she slid. I don't really know where she was before but she hasn't done anything memorable for months.)

OTOH, some players could get tons of quality points by beating them and that boosts their ranking.

Dokic's numbers might change like crazy if I use the rankings from the last week of this month because some of those players are in challenger territory. I think everything else would be about the same. I can't imagine how much worse they would be for Dokic if I used the rankings "at the time of the defeat."

In the end, I just determined it was too much work to figure out what everyone was ranked at the time...so I used the post-Munich rankings.

I still think those rankings-based stats are fuzzy stats...which is why I called them "confusing" in the first post. I just wouldn't read much into them.

The bit that blew me away was the 2- vs 3-set loss stat. I didn't realize that Capriati hung in there as much as she did. And only two other players had more 3-set losses than 2-set losses (Henin and Serena). Henin actually has more 3-set losses than Capriati...but she has more 2-set losses too.

Had I known as much, I wouldn't have taken so many cheap shots at Capriati for being a "weak no.1" or "lucky." Looks like she fought very hard.

Brian Stewart
Dec 23rd, 2001, 03:50 PM
I was just curious which rankings were used. If you don't have the info recorded somewhere, it would be a pain to try to figure out. Fortunately, I have the singles, doubles, and qualifying draws for each WTA tourney, along with rankings at the time. (I need this info for my fantasy league.) If I get the time, (and motivation), I might check the player records at Yahoo, find out where their losses occured, and retrieve the info from my sheets. Then I'll post it here, and you can modify the results if you wish.

(As long as I don't have to do the work... <img src="smile.gif" border="0"> )

cynicole
Dec 23rd, 2001, 08:14 PM
Brian - "motivation" is the key word. I crunched these numbers weeks ago but had problems getting the table to work properly. Excel likes to code things it's own way. And then when I turned it into an html table I got too many blank spaces that I gave up trying to post these stats. I solved my problem this time with dashes and the Courier New font...but it took several posts on "wins" to make me motivated enough to do it.

So if you're motivated enough to figure out the rankings of the vanquishers all I have to do is plug the numbers into the existing Excel sheet (it's in formulas so the numbers would change automatically).

Then again, I think Robert Waltz puts out this stuff in his Statistical Abstract...so maybe you wouldn't have to go through the labor of looking everything up.

A4
Dec 23rd, 2001, 08:37 PM
Great stuff cynicole. Tennisaddict, did Martina really have a better season than Serena?<br />Well, I'll be interested in checking if the win-loss records tally, i.e. the player with the most winning percentage should have the least losing percentage, which could be another yardstick to determine the season's most formidable player. I admit it will be only a rough estimate since the number of tournaments played isn't even.

Robbie.
Dec 23rd, 2001, 10:43 PM
A4, please don't get me inot this argument again <img src="wink.gif" border="0"> , I did it to death on anther thread and got really frustrated, so lets just drop it <img src="biggrin.gif" border="0"> <img src="redface.gif" border="0">

Anyways, Serena definitely has the better winning percentage (84% of matches) to Martina's 80%, but they are both excellent - remember Serena played about half the tournaments Martina did so she hadfar less oppurtunityu to lose.

Either way, I'm not getting into this again, if you want my opinion read the other thread <img src="wink.gif" border="0"> . You of course are entitled to your opinion and me to mine <img src="smile.gif" border="0"> .

Cheers.

A4
Dec 24th, 2001, 09:06 AM
tennisaddict, sure I understand. I read most of your opinions on the other thread and had come to the same conclusion as you did, that we may certainly have different opinions. I just couldn't resist one last dig though. Cheers right back to you too!

Brian Stewart
Dec 25th, 2001, 01:54 PM
Here's this list of rankings at times of their losses. I grabbed the list of each player's losses from Yahoo (except Henin, for whom there is none. I had to go to the ITF database, and they're missing some results for her.)

Davenport- 1, 14, 10, 11, 2, 6, 3, 4, 10, 10X(Walkover)

Capriati- 29, 4, 2, 3, 9, 59, 9, 10, 10, 4, 4, 15, 3, 14 (who knew that loss to Kuti Kis would cost her the year-end #1?)

V.Williams- 1, 20, 10X, 18, 25, 15 (weird- no losses to 2-14)

Hingis- 14, 2, 19, 3, 11, 5, 9, 6, 4, 83, 10, 10, 10, 13, 3

Clijsters- 1, 2, 27, 10, 7, 67, 30, 112, 4, 9, 3, 34, 48, 9, 3X, 4, 11, 20, 2

S.Williams- 1, 1, 5, 4, 4, 10, 4

Henin- 4, 20, 30, 19, 50, 51, 8, 4, 14, 2, 8, 4, 10, 17, 19, 3, 21, 10

Dokic- 2, 3, 10, 42, 2, 4, 131, 92, 7, 3, 169, 43, 66, 2, 9, 8, 2, 1, 9, 57, 3, 3, 2

Mauresmo- 2X, 3, 1, 23, 56, 31, 21, 3, 2, 28, 3, 14

Seles- 11, 14, 27, 54, 72, 4, 3, 3, 10, 37

cynicole
Dec 26th, 2001, 04:47 AM
Thanks for the rankings, Brian. Here are the new numbers...

This time, I dropped the walkovers from the "Losses To" categories but I kept them in the averages and medians (because of pure laziness).

<font face="Courier New">-------------------------LD--JC—-VW--MH--KC--SW—-JH—-JD—-AM—-MS—<br />Losses-------------------09--14--05--15--18--07--18--23--11—-10-<br />Walkovers----------------01--00--01--00--01--00--00--00--01—-00-<br />2-Set Losses-------------05--03--04--09--11--03--06--14--09--06-<br />3-Set Losses-------------04--11--01--05--07--04--12--09--03—-04-<br />Tiebreakers Won in Ls----03--02--00--01--00--01--02--00--01--00-<br />TBs Lost in Losses-------01--03--02--02--01--02--05--03--01--04-

RANKINGS AT THE TIME OF LOSS<br />Losses to Top 10 Players-07--10--01--10--10--07--09--16--05--04-<br />Ls to Outside 25 Players-00--02--00--01--06--00--03--07--03--04-<br />Average Rank Lost To-----07--13--15--13--21--04--16--29--16--24-<br />Median Rank Lost To------08--09--17--10--09--04--12--07--09--13-<br />Readjusted Rank Lost To--07--10--16--09--17--04--15--24--13--20-

YEAR-END RANKINGS<br />Losses to Top 10 Players-08--10--02--12--10--07--10--14--06--06-<br />Ls to Outside 25 Players-00--01--00--01--04--00--02--07--03--03-<br />Average Rank Lost To-----05--13--11--10--15--04--15--21--15--20-<br />Median Rank Lost To------05--08--10--06--07--03--10--05--06--08-<br />Readjusted Rank Lost To--05--08--10--07--13--03--12--17--10--14-

And in case you want to compare both, here are both (I list the "Time of Loss" numbers first).<br />-------------------------LD--JC—-VW--MH--KC--SW—-JH—-JD—-AM—-MS—<br />Average Rank Lost To-----07--13--15--13--21--04--16--29--16--24-<br />Average Rank Lost To-----05--13--11--10--15--04--15--21--15--20-

Readjusted Rank Lost To--07--10--16--09--17--04--15--24--13--20-<br />Readjusted Rank Lost To--05--08--10--07--13--03--12--17--10--14-

Median Rank Lost To------08--09--17--10--09--04--12--07--09--13-<br />Median Rank Lost To------05--08--10--06--07--03--10--05--06--08-</font>

Of these, the stats I prefer are the readjusted rankings because they eliminate the extreme numbers involved. A fluke loss to a low-ranked player (or a fluke win over a high-ranked one) is removed.

Time of Loss Adjusted Rank Order (Best to Worst)<br />04 S Williams<br />07 Davenport<br />09 Hingis<br />10 Capriati<br />13 Mauresmo<br />15 Henin<br />16 V Williams<br />17 Clijsters<br />20 Seles<br />24 Dokic