Wisden Ranks All-Time Best

[marc71178]

Originally posted by Tim
How can Agarkar be ranked 25th of all time when he is being questioned by some commentators as no more than a club bowler?

Oi, leave him alone, he tries his best!

 

[Bazza]

I really don't see what the problems are with these lists. They are very comprehensive and must have taken alot of work to invent and calculate.

If anyone wants to suggest individual problems I will try and challenge them, but certainly arguments like Ponsford having two FC 400s are entirely flawed, given that this is based on test match and one day international cricket.

 

[Mr Mxyzptlk]

Originally posted by hourn
Test Batting
Too High - Brian Lara (10th) probaly should've been around 20th.
[Edited on 16/12/2002 by hourn]

Brian Lara has played some of the most awesome and amazing innings for the West Indies in the past few years. He's had to score runs when nobody else was (which was very regularly). Innings like 213 v Australia coming back from 51 all out, 153* to beat Australia by 1 wicket. The man scored nearly 700 runs in a 3 match series on the subcontinent!!! It included 3 big hundreds and bar a couple of poor decisions he may still have been batting today.

There has to be something great about this man.

 

[Mr Mxyzptlk]

Originally posted by hourn
Test Batting
Muttiah Murilitharan at number 1 is a controversial call. I thought Marshall should've been in the top 3 at worst, while Ambrose should've got a guersney in the top 5 or so bowlers.
[Edited on 16/12/2002 by hourn]

I echo these sentiments completely. I believe that if Murali has a medical condition and can't straighten his arm then he should not be a bowler. If he is really in love with cricket, he should learn to bat. These comments sound harsh but I think that the ICC have to be stricter to keep discipline in the sport, they can't make exceptions. What's to prevent me from having a car accident and some surgery or something and claiming that I can't straighten my arm or even worse yet, that I can't bowl overhand?

Regarding Ambrose: Ambrose was one of the most awesome bowlers I have ever had the privelige to see. He bowled some of the most devastating spells in history!

 

[Top_Cat]

I really don't see what the problems are with these lists. They are very comprehensive and must have taken alot of work to invent and calculate.

The amount of work put into something is moot if the result is unsatisfactory.

but certainly arguments like Ponsford having two FC 400s are entirely flawed, given that this is based on test match and one day international cricket.[/b]

Yeh, I meant that as ASIDE from his Test record, he has two FC 400's, the only player to do so. He actually averaged almost 50 at Test level and was considered as one of the best players of his time alongside Wally Hammond and the Don. To put him as 99th right next to Woodfull is to forget just how good a player he was.

Erm, Walsh had an outstandingly long career and has more wickets than anyone else - can't be all that bad.

The length of his career is irrelevent in this case. There's no doubt Walsh was a fine bowler and in the top 20 at least but there's no way I'd then go and rate him above the guys I mentioned. He was never as destructive as those guys, particularly Marshall and Holding.

 

[masterblaster]

.......thank god not many people have made a fuss about Sachin Tendulkar being 2nd in Tests and ODI's

 

[royGilchrist]

Im a bit surprised that not many people have made a fuss about Sachin not being number 1 in ODIs (and tests)

 

[age_master]

me too

 

[Bazza]

Originally posted by Top_Cat
The amount of work put into something is moot if the result is unsatisfactory.

Agreed but it's not like they just based it on the opinion of a few people who work there. It seems well thought out and generally the criteria are very suitable.

Yeh, I meant that as ASIDE from his Test record, he has two FC 400's, the only player to do so. He actually averaged almost 50 at Test level and was considered as one of the best players of his time alongside Wally Hammond and the Don. To put him as 99th right next to Woodfull is to forget just how good a player he was.

I checked out Ponsford and his record looks pretty impressive, but he was very inconsistent. He made some big scores, then struggled for a few years and was in and out of the side, hence only playing 29 tests in 10 years.

He also made some big scores against England when Bradman was no doubt showing how it should be done, and these might not have counted for much in high scoring games.

For example on debut he scored 110 against England. Good effort. But in this timeless match there were 6 centuries and an average of more than 40 runs per wicket, so that puts it into perspective a bit. In his next match he got 128, but that was out of 600, and again over 40 runs per wicket on average, plus another 4 centuries, 3 bigger than his. He scored 110 against England a few years later (out of 695, Bradman 232), and Sutcliffe got 161 for England.

His career best 266 came in an innings of 701 (Bradman 244). His next best 181 came in an innings of 584 (Bradman 304). He got 109 against WI out of 558 (Bradman 223). Australia won this game by an innings and 217 runs, so this would have been the equivalent of Bangladesh today - even big scores don't seem so impressive. Would you count Graeme Smith's recent 200 as a great innings? Me neither.He also did nothing against SA.

I think you get what I'm trying to say - averaging 48 is good but scores were higher in those days and all his big innings came against weak opponents or were overshadowed by better knocks, hence don't count for as much.

The length of his career is irrelevent in this case. There's no doubt Walsh was a fine bowler and in the top 20 at least but there's no way I'd then go and rate him above the guys I mentioned. He was never as destructive as those guys, particularly Marshall and Holding.

Maybe not but he was arguable the most accurate and consistent bowler imaginable. He had both sides of things - playing in a strong side sharing wickets with other big names, and later carrying a weak team on his shoulders (at times along with Lara and maybe Ambrose).

Looking at his career break down, he averaged under 30 nearly every year, and in 2000 at the grand old age of 37 managed to collect 66 wickets at 18, including 34 at 12 in 5 tests against England. Truly outstanding, and you wouldn't put him higher than 20?

 

[Top_Cat]

Agreed but it's not like they just based it on the opinion of a few people who work there. It seems well thought out and generally the criteria are very suitable.

Well we differ then. I personally believe any attempt at a list of this sort with inevitably fail due to the myriad of factors involved in what makes someone a great player but even in this case, the amount of anomalies exceeds that which would be expected.

I checked out Ponsford and his record looks pretty impressive, but he was very inconsistent. He made some big scores, then struggled for a few years and was in and out of the side, hence only playing 29 tests in 10 years.

1) Playing 29 Tests in that time would hardly be considered 'in and out' of the side considering Australia played only 40 Tests in that same period. They just didn't play many Tests back then. So your judgement that he only played 29 Tests in 10 years ergo wasn't a regular member of the side is grossly unfair. In one of those series, he was injured in the second Test, preventing him for partaking in the rest of the series too.

He also made some big scores against England when Bradman was no doubt showing how it should be done, and these might not have counted for much in high scoring games.

You're kidding. You could say the same thing about Viv in the 80's whilst Greenidge and Haynes were belting bowlers to all parts. So their scores don't mean as much because someone else got a big score in the same innings? You're inferring then that somehow a person's big score in a relatively weak team means more which is the basis of the biggest problems with this list.

b]For example on debut he scored 110 against England. Good effort. But in this timeless match there were 6 centuries and an average of more than 40 runs per wicket, so that puts it into perspective a bit. In his next match he got 128, but that was out of 600, and again over 40 runs per wicket on average, plus another 4 centuries, 3 bigger than his. He scored 110 against England a few years later (out of 695, Bradman 232), and Sutcliffe got 161 for England.[/b]

Sure pitches were better back then but so what? Considering the strength of England's bowling attack of the time, it's extremely harsh (and remarkably easy in hindsight, I might add) to criticise him on that basis.

His career best 266 came in an innings of 701 (Bradman 244). His next best 181 came in an innings of 584 (Bradman 304).

Yeah that's right, against the mere pop-gun attack of Hedley Verity, Bill Bowes, Wally Hammond, Maurice Leyland and 'Gubby' Allen. Damn, how easy would THAT have been...............:rolleyes:

I think you get what I'm trying to say - averaging 48 is good but scores were higher in those days and all his big innings came against weak opponents or were overshadowed by better knocks, hence don't count for as much.

Rubbish. I think you'll notice that not only did most of his REALLY big scores come when Bradman entered the side but most of them were in PARTNERSHIP with Bradman. More striking is that in most of those big scores, he and Bradman were the only guys with significant scores in the innings. This suggests that he was one of the few players good enough to stay with Bradman long enough for him to dominate the opposition.

I mean, let's look at the scores you mention there:

In that score of 266 where Bradman scored 244, the next highest score was 49.

http://statserver.cricket.org/link_.../1934/AUS_IN_ENG/AUS_ENG_T5_18-22AUG1934.html

In that 181 where Bradman scored 304, the next highest score was 27.

http://statserver.cricket.org/link_.../1934/AUS_IN_ENG/AUS_ENG_T4_20-24JUL1934.html

Seeing a pattern here yet?

And again, the scores other than against the WI, were hardly against weak bowling attacks. And even the West Indian side had bowlers like Leary Constantine and Herman Griffith in the side.

All in all, having had a look at his stats in more detail, I've gone from fairly sure to uterly convinced that he should have been in the top 50. Of all the names in Bradman's time who stand out as great players, Ponsford is one of them and not only have his contemporaries prasied him as such in the past (Bradman, McCabe, Woodfull etc.), but it'sborne out of his stats too. You're judging him FAR too harshly.

Looking at his career break down, he averaged under 30 nearly every year, and in 2000 at the grand old age of 37 managed to collect 66 wickets at 18, including 34 at 12 in 5 tests against England. Truly outstanding, and you wouldn't put him higher than 20?

That's right. I only put him as high on the list as the number of times he's been dropped from the WI side. What was it, 19? You want to talk about being in-and-out of the side...........

To put Courtney above Garner, Holding and Roberts is disgraceful, in my opinion. It also shows up one of the fundamental flaws of the criteria used and shows it wasn't comprehensive enough. It's biased and as a professional statatician I can tell you, it's flawed. I can do a detailed analysis of WHY its flawed if you like...............

 

[Bazza]

Originally posted by Top_Cat
Well we differ then. I personally believe any attempt at a list of this sort with inevitably fail due to the myriad of factors involved in what makes someone a great player but even in this case, the amount of anomalies exceeds that which would be expected.

Well obviously any list will cause controversy because everyone has different opinions. The main reason for this would have been to strike up some conversations, arguments even and get people talking about cricket. That is a good thing for which I praise Wisden, and IMO they've done a good job, though I admit there are some eyebrow-raising results.

1) Playing 29 Tests in that time would hardly be considered 'in and out' of the side considering Australia played only 40 Tests in that same period. They just didn't play many Tests back then. So your judgement that he only played 29 Tests in 10 years ergo wasn't a regular member of the side is grossly unfair. In one of those series, he was injured in the second Test, preventing him for partaking in the rest of the series too.

OK I take your point here - you obviously know more about cricket in the 1920s and 1930s that myself.

You're inferring then that somehow a person's big score in a relatively weak team means more which is the basis of the biggest problems with this list.

Of course it is. Some of the best innings in recent years were Brian Lara's against Australia - in 1998/99 WI lost the first test by 312 runs after being bowled out for 51. In the next match Lara came back with 213 against the same bowling attack (McGrath, Gillespie, Warne and MacGill - no mugs!) and they narrowly failed to secure an innings victory. The side only containd four others of any note - Adams, Jacobs, Ambrose and Walsh. In the 3rd test he came up with a very unsupported 153* as WI successfully chased 308 for a 1 wicket win. In the fourth test he got another ton (off just 84 balls), but it was to no avail as the Aussies wrapped up a comfortable 176 run victory to tie the series.

Cut to Australia 2 years later, and McGrath grabs 10-27 from 33 overs as AUS take an easy inns+126 win in the first test, and they follow it up with another inns triumph in test 2 (by this stage Aus have Lee and WI have lost Ambrose, so are even weaker). In the 3rd test Lara makes a stunning 182, but nothing can stop the Aus juggernaut on it's way to a 5-0 series win. The other games were won by 352 runs and 6 wkts.

These innings were much better than many others you are likely to see as they were made in the face of adversity. That's what made Laxman's innings so incredible. The best this summer was Dravid's rearguard action to save a match allowing India to come back and draw the series, it wasn't either of Vaughan's 190-odd's.

That is why Gooch's 154* at Headingley against WI is rated the best of all time, followed by the above mentioned 213 from Lara, and not his 375 against England on a pancake pitch.

Sure pitches were better back then but so what? Considering the strength of England's bowling attack of the time, it's extremely harsh (and remarkably easy in hindsight, I might add) to criticise him on that basis.

I don't think England's attack was particularly outstanding, and they seem to have made two changes between every match (a bit like current proceedings then!)

Yeah that's right, against the mere pop-gun attack of Hedley Verity, Bill Bowes, Wally Hammond, Maurice Leyland and 'Gubby' Allen. Damn, how easy would THAT have been...............:rolleyes:

Hammond was a batsman who turned his arm over, and Leyland only took 6 test match wickets, though granted the others ranged from fairly handy to pretty good!

I think you'll notice that not only did most of his REALLY big scores come when Bradman entered the side but most of them were in PARTNERSHIP with Bradman. More striking is that in most of those big scores, he and Bradman were the only guys with significant scores in the innings. This suggests that he was one of the few players good enough to stay with Bradman long enough for him to dominate the opposition.

True but you can look at it both ways - I think anyone would have more success
if they had played with Bradman - I would imagine he took a fair bit of pressure away from the others!!

And again, the scores other than against the WI, were hardly against weak bowling attacks. And even the West Indian side had bowlers like Leary Constantine and Herman Griffith in the side.

Constantine and Griffith don't stand out that much, and look at some of the others - Scott 22@42, Martin 8@77. Headley is the only name who stands out for me, hence Aus winning the first match by 10 wkts, and the next 3 by an innings and plenty.

judging him FAR too harshly.

You're right, maybe I am.

That's right. I only put him (Walsh now) as high on the list as the number of times he's been dropped from the WI side. What was it, 19? You want to talk about being in-and-out of the side...........

To put Courtney above Garner, Holding and Roberts is disgraceful, in my opinion. It also shows up one of the fundamental flaws of the criteria used and shows it wasn't comprehensive enough. It's biased and as a professional statatician I can tell you, it's flawed. I can do a detailed analysis of WHY its flawed if you like...............

I'd be interested to see that yes. I am very interested in stats as you may realise! I think they tell alot especially in cricket anyway.

As for Walsh, well I can't form any opinion as to whether he was better than Holding, Marshall etc though the general consensus from people who can is that he wasn't. For me though he is the best bowler I have ever seen, along with Ambrose.

I suppose the bottom line is that these lists will always be difficult, impossible even, and certainly never definitive, but they are very interesting to say the least!

 

[Mr Mxyzptlk]

Originally posted by Top_Cat
the number of times he's been dropped from the WI side. What was it, 19

Walsh was dropped 19 times? I find this hard to believe. This could not have been in the last decade because Walsh and Ambrose were unseparable and devastating. You must mean injured or rested. Surely you are mistaken.

 

[Top_Cat]

I'd be interested to see that yes. I am very interested in stats as you may realise! I think they tell alot especially in cricket anyway.

Hmmmmmmm...............okay well a bit of background first.

Do the terms chi-squared test, t-test, least-squares regression and multi-variate analysis mean anything to you? If so, I'll proceed. If not, I'll do my best to explain it. Maybe I'll post an article with a statistical analysis of the list, eh?

Walsh was dropped 19 times? I find this hard to believe. This could not have been in the last decade because Walsh and Ambrose were unseparable and devastating. You must mean injured or rested. Surely you are mistaken.

Ah, you're right. Not including injuries, he's been dropped for lack of form 13 times.

http://www.cricket.org/link_to_database/ARCHIVE/CRICKET_NEWS/2000/MAR/027749_CI_29MAR2000.html

I suppose the bottom line is that these lists will always be difficult, impossible even, and certainly never definitive, but they are very interesting to say the least!

This is true. Unfortunately some people choose to take these lists as definative when they never can be because what makes the perception of a great player is more than pure ability. Opinions vary as to how much ability a player has too.

To me, lists like this should be banned because they can easily be taken apart as a load of crapola. If I do get around to doing these stats tests, maybe I'll send it to Wisden.

 

[Bazza]

Originally posted by Top_Cat
Hmmmmmmm...............okay well a bit of background first.

Do the terms chi-squared test, t-test, least-squares regression and multi-variate analysis mean anything to you? If so, I'll proceed. If not, I'll do my best to explain it. Maybe I'll post an article with a statistical analysis of the list, eh?

Erm... no! :duh:

But I am interested...

 

[Top_Cat]

Okay, just quickly:

Chi-Square

Chi-square test is based on the chi square distribution. It is used for the goodness-of-fit test because it asks whether there is a good fit between the data (O = Observed frequency) and the theory (E = Expected frequency). Chi-square determines whether the differences between the observed and expected scores can be attributed to some actual difference in behavior or if this difference between the scores is caused by chance using the formula chi square = sum{ (O-E)^2) / E}.

(gawd that formula looks horrendous in that form!)

T-test

Effectively a test of inference. You try to completely disprove something and do this to indirectly prove something else. In short, in a certain situation where you might want to prove how a cohort of people died, you use the t-test to prove what DIDN'T kill them.

Here's an explanation:

http://www-stat.stanford.edu/~owen/courses/262/ttest.pdf


Linear Least-Squares Regression

Here's where you effectively plot certain points on a curve (a straight line in stats is still considered a 'curve') and use that to prove that a phenomena will continue with that behaviour and thereby prove something else.

For example; if you want to prove that cigarette smoke causes cancer, you might plot something like 'amount of cigarette smoke the subject is exposed to' against the probability of developing a tumour. A straight line behaviour assumes that there is a direct correlation between amount of cigarette smoke exposure and the probability of getting cancer ergo, if you smoke, you have a certain probability of developing cancer. If you smoke more, the probability increases etc. It seems almost idiotic to assume otherwise these days but these very correlations I'm talking of which showed as convincingly as possible the effects of cigarette smoke on people were ignored in the 70's and 80's by most western governments.

Another example is how the quantity 'Absolute Zero' was found.

http://www.riverdeep.net/science/middle_school_gateways/msg_handouts/chem/gsl/mgsld4i.pdf

Notice the plots of volume vs temperature plots were 'extrapolated' back past zero degrees celcius to find the point at which energy was zero and that it ended up at -273 degree celcius. This was then assume to be 'absolute zero' where there is NO energy in the system, hence it's the lowest possible temperature. Of course, this then resulted in a new scale of temperature measurement being created called 'Kelvin' where 0 Kelvin IS absolute zero.

Technically this was a statistical correlation until people actually cooled down stuff to near absolute zero (according to the one of the Laws of Thermodynamics, matter cannot be cooled to absolute zero) so this is where a statistical correlation has effectively proven something which was later confirmed by experimentation.

Here's a bit of theory on least-squares regression:

http://www.itl.nist.gov/div898/handbook/pmd/section1/pmd141.htm

Multivariate analysis

http://www.stat.psu.edu/~rathbun/stat505/slides/Intro.pdf

Pretty much says it all.

Either way, these are a few of the things I could use to prove a few things about the Wisden ratings. If I get around to it sometime soon, I'll post it.



[Edited on 19/12/2002 by Top_Cat]

 

[marc71178]

That's all very good TC, but how exactly does it help here?

 

[Top_Cat]

Just a quick summation of some of the techniques I would use to do a stats analysis of the Wisden lists is all. Bazza asked, after all.

 

[hourn]

Top_Cat,


***goes straight over my head***


the best way to analyse players etc,. is to just not over complicate if you know what i mean. Look at players records, look at them in different countries and against different players, but dont go overboard and analyse every little thing about them, because then you get very flawed lists like Wisdens.

 

[Bazza]

Yes hourn, but if you just use something simple there are many more flaws. If you base the list purely on average there are so many arguments which can be put forward, so you try to remove them by balancing out pitch factors, the strength of opposing bowlers, etc. Then it just all snowballs from there!!!

PS: Absolute zero is -273.15 degrees celcius, to be precise... :P

 

[marc71178]

Originally posted by Top_Cat
Just a quick summation of some of the techniques I would use to do a stats analysis of the Wisden lists is all. Bazza asked, after all.

But how would you use those methods?