Kallis vs Ponting as test batsmen

[Top_Cat]

If you're going to make the distribution argument, though, arbitrarily setting a number of innings' cut-off won't work either, no batters' distributions would look the same especially taking into account wildly different freqs when it comes to how many times they've played against different opposition/matches per country/etc. Even raw runs scored by a batter in an innings, as I've said many times on this site, is a very insensitive measure. I liken it to climbing both Everest and Kilimanjaro and deciding Everest was tougher purely by counting the number of steps.

A serious analysis of Anyone vs Anyone would require the identification of strata and systematic samples taken otherwise you're arguing noise vs noise with no idea of what the signal is. Even then, only taking raw runs into account, you're missing a great deal of the picture of each inning so there'll be strong caveats associated with any conclusions drawn which may even render the whole exercise worthless.

 

[Uppercut]

If you're going to make the distribution argument, though, arbitrarily setting a number of innings' cut-off won't work either, no batters' distributions would look the same especially taking into account wildly different freqs when it comes to how many times they've played against different opposition/matches per country/etc. Even raw runs scored by a batter in an innings, as I've said many times on this site, is a very insensitive measure. I liken it to climbing both Everest and Kilimanjaro and deciding Everest was tougher purely by counting the number of steps.

A serious analysis of Anyone vs Anyone would require the identification of strata and systematic samples taken otherwise you're arguing noise vs noise with no idea of what the signal is. Even then, only taking raw runs into account, you're missing a great deal of the picture of each inning so there'll be strong caveats associated with any conclusions drawn which may even render the whole exercise worthless.

This is all true, but if you really want to know who the better batsman is, there really isn't any good way of judging. Doing so by watching them bat is equally hopeless for all kinds of reasons- if your prolonged education in statistics tells you how worthless this option is then I'm sure your education in Psychology gives an equally damning assessment of the alternative .

To restate my personal view, if they've scored a similar number of runs over a similar period of time at a similar average, they're surely of similar quality- or at least, that's the only conclusion anyone can come to on the subject without using some extremely flimsy evidence.

 

[Furball]

I believe the word you are looking for at the end of the second paragraph is "bias."

 

[Top_Cat]

This is all true, but if you really want to know who the better batsman is, there really isn't any good way of judging. Doing so by watching them bat is equally hopeless for all kinds of reasons- if your prolonged education in statistics tells you how worthless this option is then I'm sure your education in Psychology gives an equally damning assessment of the alternative .

Yeah, there's no way of getting around the insensitivity of runs scored as a measure to be used for comparison. No amount of statistical manipulations will compensate for the GIGO effect.

As for qualitative descriptions, I don't think it's worthless but it'd be very difficult. Usually in these kinds of things, you need someone or several someones considered expert in the field and somehow draw together what they say. Even then, what's a cricketing expert? Someone with raw technical expertise (Langer) or someone who can't articulate why they scored a ton of runs but was a high performer (Greg Chappell). Is Glenn McGrath a better expert than Troy Cooley or David Saker because he did it at the top level? Etc., etc.

Running an experiment for a few years on cricketers where you watched enough games and took measurements of variables of interest other than just how many runs they scored would be fun, though.

 

[silentstriker]

To restate my personal view, if they've scored a similar number of runs over a similar period of time at a similar average, they're surely of similar quality- or at least, that's the only conclusion anyone can come to on the subject without using some extremely flimsy evidence.

Assuming they do it at similar S/R, and against similar opposition, you'd be correct.

 

[akilana]

and similar conditions.

 

[Top_Cat]

The (statistical) horror.

 

[Uppercut]

As for qualitative descriptions, I don't think it's worthless but it'd be very difficult. Usually in these kinds of things, you need someone or several someones considered expert in the field and somehow draw together what they say. Even then, what's a cricketing expert? Someone with raw technical expertise (Langer) or someone who can't articulate why they scored a ton of runs but was a high performer (Greg Chappell). Is Glenn McGrath a better expert than Troy Cooley or David Saker because he did it at the top level? Etc., etc.

Running an experiment for a few years on cricketers where you watched enough games and took measurements of variables of interest other than just how many runs they scored would be fun, though.

I doubt a test batsman would ever play enough matches to get a meaningful enough sample size to make an assessment like that worth something. Amongst many other things, the fact that every innings ends as a result of one particular delivery is the source of an absolutely insane amount of variance.

Could work for bowling though.

 

[Uppercut]

Assuming they do it at similar S/R, and against similar opposition, you'd be correct.

Maybe I should rephrase- I haven't got the slightest idea who's better- none of the data available, either statistically or qualitatively, is remotely useful on the subject. That applies to assessing the value of strike rate too.

 

[Top_Cat]

I doubt a test batsman would ever play enough matches to get a meaningful enough sample size to make an assessment like that worth something. Amongst many other things, the fact that every innings ends as a result of one particular delivery is the source of an absolutely insane amount of variance.

It'd be an observational experiment, you don't have to wait for them to play enough. Would just need to take a stratified random sample of the matches a batter has played and, well, watch them. Just have to decide on your input variables (lots of them, obv) and analyse as normal, probably a fair amount of factor analysis needed and I'd guess there'd be plenty you could group. Theoretically, it's not that big a deal.

The practicalities of it are where it would fall down. Would take an immense time and monetary commitment, deciding on the strata would be hard enough and one wonders what the point of it would be. To finally end all doubt on CW about whether Sachin is better than Lara?!

Hmmm..... I've decided it's worth it.

 

[honestbharani]

It'd be an observational experiment, you don't have to wait for them to play enough. Would just need to take a stratified random sample of the matches a batter has played and, well, watch them. Just have to decide on your input variables (lots of them, obv) and analyse as normal, probably a fair amount of factor analysis needed and I'd guess there'd be plenty you could group. Theoretically, it's not that big a deal.

The practicalities of it are where it would fall down. Would take an immense time and monetary commitment, deciding on the strata would be hard enough and one wonders what the point of it would be. To finally end all doubt on CW about whether Sachin is better than Lara?!

Hmmm..... I've decided it's worth it.

Don't tell that to the guys with IQ of 170...

 

[JBH001]

*deleted/edited*

It'd be an observational experiment, you don't have to wait for them to play enough. Would just need to take a stratified random sample of the matches a batter has played and, well, watch them. Just have to decide on your input variables (lots of them, obv) and analyse as normal, probably a fair amount of factor analysis needed and I'd guess there'd be plenty you could group. Theoretically, it's not that big a deal.

The practicalities of it are where it would fall down. Would take an immense time and monetary commitment, deciding on the strata would be hard enough and one wonders what the point of it would be. To finally end all doubt on CW about whether Sachin is better than Lara?!

Hmmm..... I've decided it's worth it.

Does this mean what I think it means?

 

[ankitj]

If you're going to make the distribution argument, though, arbitrarily setting a number of innings' cut-off won't work either, no batters' distributions would look the same especially taking into account wildly different freqs when it comes to how many times they've played against different opposition/matches per country/etc. Even raw runs scored by a batter in an innings, as I've said many times on this site, is a very insensitive measure. I liken it to climbing both Everest and Kilimanjaro and deciding Everest was tougher purely by counting the number of steps.

A serious analysis of Anyone vs Anyone would require the identification of strata and systematic samples taken otherwise you're arguing noise vs noise with no idea of what the signal is. Even then, only taking raw runs into account, you're missing a great deal of the picture of each inning so there'll be strong caveats associated with any conclusions drawn which may even render the whole exercise worthless.

I agree with all of that. And that is the reason I generally believe that number of runs or batting averages don't reveal enough, not quite as much number of wickets and bowling averages do. Always believed that statistical analysis of bowling makes more sense than that of batting.

 

[Top_Cat]

*deleted/edited*



Does this mean what I think it means?

What, that I'm going to do it?! No way. Got better things to do with my time than watch Tests and take measurements. Like just watching Tests.

 

[Avada Kedavra]

actually, no. He's not rubbish against good attacks. Just that he's not as good as say a sachin or a lara or a steve waugh in the past 2 decades.

I'd say he,kallis,dravid,yousuf,sanga,mahela etc have cashed in quite a bit on the weaker attacks of the decade and and on the flatter pitches . Doesn't mean they're not very good. Just not as good against quality bowling as sachin, lara,steve IMO

Yep he cashed in vs weak wi and nz teams and ind at home.His away avg is less than 50 and his avg in weak team is 40....shows he cant carry a weak team like true greats.gud player but not atg imo.

 

[Burgey]

How did Lara, Tendulkar or Waugh carry weak teams? When their teams were weak, the results of the team were poor. When the teams got better, so did their results.

 

[centurymaker]

How did Lara, Tendulkar or Waugh carry weak teams? When their teams were weak, the results of the team were poor. When the teams got better, so did their results.

They performed above their avgs.

 

[Burgey]

I see what you mean. I thought it was being said thy they had somehow caused the teams to win more when they were no good.

It's interesting though, because all players have their peaks and troughs. Lara had a year or two where his performances weren't that great - may have been the late-ish 90s, I can't recall exactly.. His average dropped below 50 for a time. Likewise Tendulkar had his period where his elbow injury affected him. Ponting is going rubbish now too.

I think you need to be careful in saying these things. I mean, for virtually all Lara's career from about 97 afterwards, WI were pretty poor for the most - certainly post Ambrose's retirement. When Tendulkar had his first really great incarnation in the 90s, India weren't great either. I dont know how it's a determinative factor when rating a batsman to give them points for batting during their primes when their team is not good. Likewise, I don't know that it's a determinative factor if their peak came during their side's dominant period (like Ponting). Who can say if their being in a good side makes it easy for so individual a thing as batting, or whether their peak contributes to the team being great? Who can say? And it may not matter anyways - they are all great players.

 

[Hit Wicket]

Ambrose and Walsh played till 2000-01. And both were still very good when they retired.

It's pretty basic cricket understanding that a player is more likely to perform and thrive in a stronger team and setup than a weak one. One can give numerous examples to make the point. If your bowlers have bowled out the opposition for 200 odd, the opposing team is going to start putting defensive fields much more early in the game. The openers will thrive in the knowledge that they have a solid middle order to follow, and the middle order will thrive on the consistently good starts given by the openers. The bowlers in their turn will benefit from having big totals to bowl to. Everyone is a beneficiary in a strong setup.

It's a possibility that Ponting has genuinely encountered a bad run of form, which has nothing to do with the strength of the Australian team after the retirements of Warne, McGrath, Hayden, Gilchrist, and Langer. However, the timing of it does not seem to lend much credence to it. In all likelihood, it has been both factors which have fed on each other - a weaker team leading him to a fall in form, and the fall in form contributing to an even weaker team. Much the same way his golden run was augmented by being part of an all time great team.

 

[Burgey]

Ambrose and Walsh played till 2000-01. And both were still very good when they retired.

It's pretty basic cricket understanding that a player is more likely to perform and thrive in a stronger team and setup than a weak one. One can give numerous examples to make the point. If your bowlers have bowled out the opposition for 200 odd, the opposing team is going to start putting defensive fields much more early in the game. The openers will thrive in the knowledge that they have a solid middle order to follow, and the middle order will thrive on the consistently good starts given by the openers. The bowlers in their turn will benefit from having big totals to bowl to. Everyone is a beneficiary in a strong setup.

It's a possibility that Ponting has genuinely encountered a bad run of form, which has nothing to do with the strength of the Australian team after the retirements of Warne, McGrath, Hayden, Gilchrist, and Langer. However, the timing of it does not seem to lend much credence to it. In all likelihood, it has been both factors which have fed on each other - a weaker team leading him to a fall in form, and the fall in form contributing to an even weaker team. Much the same way his golden run was augmented by being part of an all time great team.

I see what you're saying, though there have been a lot of discussions on here about whether it's better to bowl in a good attack or to stand out in a lesser one.

I tend to think Ponting's form problems relate more to age than to much else. Combine that with the captaincy of a team that isn't performing well - which he must inevitably devote his time to as well - and it kind of feeds on itself.

I mean, i personally think it's difficult to say Player X is > Player Y because when he was in his pomp he played in a weak team and when Player Y's team went South so did his form in his mid-late 30s

As always, many variables at work.