That's very interesting and well-written.

I had no idea that the average age of the WTA top 100 was over 30 - that seems completely counter-intuitive when the received wisdom seems to be that if a girl needs to be a star by 18 to have any chance of making it to the top!

It says at the end "has this policy been decided because the head of women's tennis was a statistician and the head of men's tennis wasn't," but the problem doesn't seem to be that the head of women's tennis (whichever one they are referring to in that report - Maes?) used stats to inform their decisions, but that they interpreted them so badly.

For example, the 58:42 ratio of 2007 top 100 players being ranked top 50 in the ITF junior rankings before reaching the age of 16 isn't anywhere near a significant enough ratio to be basing fund / scrapheap decisions on, especially when you take into account the stats for current and recent GB top players in table 1.3.

Similarly, if it's true (I say this because I'm aware that we're only getting one side of the story in this report) big changes in the rules governing the bottom end of the WTA rankings in the last decade or so weren't taken into account in assessing how old successful players were when they first got ranked, well that's just silly!

There are few things more dangerous than a mass of statistics in the hands of people who have no idea how to interpret them properly or how to use them as a tool to decide where to dig deeper ... unfortunately that's what happens to most statistics though!

It was good to read reports this week of Steven Martens' apparently late conversion to the idea that players on the men's side shouldn't be written off too young any more and you'd really hope that this report and, even more prominently, Anne and Elena's recent success (as you can see at http://www.freewebs.com/britishtennis/stats/ag-f.htm?n=1, they're the 2nd and 3rd oldest Brits who still have a WTA singles ranking, the only older player being the GB doubles no. 1, which does make you wonder which other players might have been late bloomers had they not given up intheir early 20s) might lead to this dawning on those who run the women's game too.

I suppose it all depends on what you want - if you're only interested in producing top 5 players, then yes they probably do have to be stars at a very young age (though Henman wasn't) but top 5 players tend to be one-off phenomenons who don't get produced by any country's 'system' making it a pointless thing to try to control - the important thing there is to make tennis attractive enough to young people that the potential top 5 player does actually take the sport up in the first place then to tailor any support you give to them rather than trying to fit them into any kind of system.

It's the players who aren't prodigies like that but could with the right support get into the top 100/50/25, perhaps in their mid-late 20s, that a Federation can have more effect on - and it is worth having players like that, because the younger players are morel likely to feel that they can emulate them, less likely to feel that they should give up too young, and so on, all of which can only be positive and create the conditions for more top 100 players in the future.

-- Edited by steven on Thursday 15th of April 2010 11:01:34 AM

steven wrote:

---

There are few things more dangerous than a mass of statistics in the hands of people who have no idea how to interpret them properly or how to use them as a tool to decide where to dig deeper ... unfortunately that's what happens to most statistics though!

---

Couldn't agree more.  In this current age it is so much easier to gather loads of statistics on many things.  Managers ( and politicians ! ) and others in positions of influence then fancy themselves as statistical analysts and so often in many fields in life just draw the wrong conclusions, often in consequence wasting a great deal of time and money.

I spent 4 years at university, largely studying statistics during that time.  Through training I can be very wary of such as statistics being skewed by other causes ( and take steps to elimate other causal effects ), and will be more aware of what is statistically significant.  Where decisions can have very serious consequences, statistics in the hands of the these who have not at least got a good feel for numbers can be a very dangerous thing.  I am not saying it necessarily needs trained statisticians, though the more important and costly the consequences the better that would be, since at some levels just good common sense always helps.

All that said, having statistics interpreted correctly is one thing, there must always be a human empathy element too, have people involved who can relate to and understand the human consequences too.  Use statistics but don't be ruled by the statistics. So ideally you have properly interpreted statistics available and then they can form a part of the decision making process.

PS :  I have only briefly looked at the document here.  There seems a lot to it and I'd like to take time to look at it more clearly and digest what is being said.  So the above is all just talking generally re statistics and is not a comment on any statistics or inferences being drawn from either the LTA or the GBtennisgirls.

-- Edited by indiana on Thursday 15th of April 2010 02:15:49 PM

Yes, that average age of the top 100 in 2007 being 31 did seem intuitively wrong, in fact had to be wrong the more one thought about it.  Thanks to Ratty for basically showing that bit is crap ( statistical terminology  ).  Here's the y/e 2007 rankings if anyone wants to waste time trying to show somehow it was remarkably different in 2007 :  http://www.sonyericssonwtatour.com/SEWTATour-Archive/Rankings_Stats/Singles_Numeric_2007.pdf

The "average age of top 100 was 31" comment just seems relevant to then being used to show how long ago these players relevant junior age must have been and concluding it was generally before the "rankings rule change" that gave folk ranking points just for competing.

This represents very muddled thinking in presenting well what I do think in many ways is a good case.  The date players were first ranked may be effected by this, but that is not really the main issue here so it certainly wasn't worth using clearly wrong statistics re that point. 

The real focus, which do move onto and I too find tremendously worrying is this quote of what the LTA supposedly say :

"Also, the LTA say because 58% of players ranked top 50 in the ITF while under 16 that is the standard the LTA are aiming for with the GB players".

That is a ridiculously insignificant majority to be seemingly basing such vital funding and player participation policies on !  It would appear they are effectively possibly closing the door on a subsection that were able to produce 42% of the women's top 100 players.  That's crazy.

I do think it is true that if you look at the very top of the women's rankings, eg the top 10, you will find that they are almost exclusively very highly ranked young juniors or known to be exceptional young juniors if not actually ranked as such for some reasons.  There is another thread with stats that reinforces these thoughts.

But that is looking at the very very top.  As Steven suggests, you can't base policy on trying to only cater for the group that is by far more likely to produce top 10 players, you have to consider really trying to expand our player base of top 100ers, top 200ers.  The 58 : 42 ratio of top 100ers is I say again a ridiculously insignificant majority and so exclusive to so many potential top players to not be at all credible.  Indeed as the gbgirls point out, just look at the how basically hardly any of the current top ranked GB women were "top 50 in the ITF while under 16".

I do realise that some credibility is lost by the gbgirls by the dodgy ( nae, wrong ) average top 100 age stats that they bring up ( though to pursue an insignificant point about when earliest ranked ), that it is their quoting of what the LTA say, and there is some truth in the LTA's point about girls generally mature earlier than boys, but I would say that :

a)  you should not exclude / restrict compared to boys based on general earlier maturing of girls, because that is not uniformly true, both in sport and in life generally.

b)  the LTA are totally damned by their own stats if they really are trying to base so much on this "58:42" statistic, which would exclude by definition 42% of the women's top 100 and almost all GB's currently top ranked players.

-- Edited by indiana on Friday 16th of April 2010 02:11:10 PM