Garry Sobers v Imran Khan,Test Cricket:Poll

[CaptainGrumpy]

But it does average out.

Do you want a Murali and Akram comparison done? I suspect Murali would edge out being further ahead of par than Akram, but certainly not by a factor of 1.5 times. Maybe I guesstimate that it would be Murali at like 50 and Akram at 40 ahead of par, but that is a guestimate. A top quality spinner who can take first innings wickets is a captain's dream because they can bowl more than their 'equal' share of the overs as compared to a typical fast seam bowler. That is the incredible value of Warne and Murali. That does not make Murali 1.25 times Akram. Makes him about quarter a wicket, or 10 runs ahead per match in contribution average than Akram - of which they are both positive.

Still way behind Bradman at 100 better than par.

 

[longranger]

But it does average out.

If you are still saying that, it means you did not understand my point. Saying it averages out 100% of the time is a very optimistic assumption at the best, and its being very closed minded at the worst.

 

[watson]

But it does average out.

Do you want a Murali and Akram comparison done? I suspect Murali would edge out being further ahead of par than Akram, but certainly not by a factor 1.5. I guesstimate that it would be Murali at like 26 and Akram at 20 ahead of par, but that is a guestimate. A top quality spinner who can take first innings wickets is a captain's dream because they can bowl more than their 'equal' share of the overs as compared to a typical fast seam bowler. That is the incredible value of Warne and Murali.

I think that you are trying to unweave a rainbow here.

Because there are so many variables at play, separating elite players and making a definitive claim that one in categorically better than another is a simple impossibility.

Even Bradman is not immune from comparison if we provide a certain context. For example, on a rank sticky-wicket it is very likely that there wouldn't be much difference between Bradman, Hobbs, Headley, and Hutton because the latter three batsman had supreme techniques on dispicable wickets and Bradman didn't. In that specific scenario you can more-or-less throw the usual stats and figures out the window.

 

[CaptainGrumpy]

I think that you are trying to unweave a rainbow here.

Because there are so many variables at play, separating elite players and making a definitive claim that one in categorically better than another is a simple impossibility.

Even Bradman is not immune from comparison if we provide a certain context. For example, on a rank sticky-wicket it is very likely that there wouldn't be much difference between Bradman, Hobbs, Headley, and Hutton because the latter three batsman had supreme techniques on dispicable wickets and Bradman didn't. In that specific scenario you can more-or-less throw the usual stats and figures out the window.

Bradman would reverse the batting order. Players these days do not play on uncovered pitches.

This method freely ignores "peak potential" and "troughs" and uses their career average of what they did play and perform. Match contexts average out. They don't perfectly. That is a limitation, but its very much an objective one.

 

[CaptainGrumpy]

If you are still saying that, it means you did not understand my point. Saying it averages out 100% of the time is a very optimistic assumption at the best, and its being very closed minded at the worst.

With all due respect, I understand your point, it is one that I have freely acknowledge, but I think ignoring averages comparison because they are not perfect in averaging player's match contexts over a long career is a worse case of being close minded. It is not that I don't understand your point, its that I think you are overstating it.

 

[OverratedSanity]

With all due respect, I think ignoring averages comparison because they are not perfect in averaging player's match contexts over a long career is a worse case of being close minded.

He didn't say they should be ignored.

This is a painful thread.

 

[CaptainGrumpy]

I repeat, statistics are very important. They give context to opinions. But they're just one of the many tools available to judge quality. Don't underestimate the human brain which can judge after taking in inputs from various sources.

I'm not. I am agreeing with an objective means that enables compare batting performance with bowling performance overall that is objective on actual score card performance and not using any subjectivity, which means that all people can use the same formula. Its not perfect. But is the human brain or everyone's human brain?

I really don't want to debate on the value or dangers of statistics. I clearly accept sports statistics have limitations but are more useful than not.

 

[RossTaylorsBox]

Please clarify exactly what you're trying to say about a "flaw" here. Because ten wickets is an innings. If ten batsmen lose their wicket, its innings over. Someone will be always be left "not out" in every innings. A batsman who bats lower will statistically be left "not out" more often and yet still have a lower batting average than the batsmen ahead of him. Its a skill to not lose your wicket cheaply. Its a measure of how many runs the batsman contributed for his own wicket. Individuals contribute the runs to the partnerships in differing proportion of contribution. Surely you accept and acknowledge this fact.

Not outs aren't the issue. From what I gather you're comparing the average of an individual against the benchmark of the average runs scored in a partnership between two batsmen. Any reason for not simply comparing individual averages in the eras of the batsmen?

Medians won't work. The only way batsman are penalised are those that bat a longer time to collect more extras. So Trevor Franklin's value will be possibly slightly underestimated because he only scored 23 runs per innings and gave more time to collect extras. This is offset by batsman who score runs quickly, like David Warner, but actually do less for the team in collecting in extras. But the extras are collected by all batsmen who bat. Some would say that fast scoring batsmen do more good for the team in breaking the other team's spirit or setting the team up for victory, so it averages out to some extent. Warner and Franklin are deliberately chosen as being at either end of the quick scoring opener scale.

Strike rates have literally nothing to do with it. You're using a mean average for statistics which have a really skewed distribution.

No, its really easy on a wickets per runs average. Its just a lot of work to collect and process all the information.

The issue I see is whether use to the actual innings per match batted, or the average innings batted per match, or both and average the result out.

Yeah but my point is that your method of finding an "average" player is not objective.

 

[longranger]

I'm not. I am agreeing with an objective means that enables compare batting performance with bowling performance overall that is objective on actual score card performance and not using any subjectivity, which means that all people can use the same formula. Its not perfect. But is the human brain or everyone's human brain?

Alright then, improve the measure. A simple 'runs to wickets' measure is inadequate. How about the weightage? Top order wickets are more relevant than tail-end wickets.

 

[CaptainGrumpy]

Not outs aren't the issue. From what I gather you're comparing the average of an individual against the benchmark of the average runs scored in a partnership between two batsmen. Any reason for not simply comparing individual averages in the eras of the batsmen?

Yes. The individual averages are compared after they have been standardised against the era average because of difference in eras. In the times of WG Grave a wicket was worth far less than the 36 runs it was worth in the 1980's. But a wicket is still worth a wicket. That is the baseline. The contribution to being ahead of par can be listed in a ratio as a value of a wicket. Or as runs. The player's individual averages are used.

Strike rates have literally nothing to do with it. You're using a mean average for statistics which have a really skewed distribution.

No, Bradman at 99.94 is a mere outlier.

Yeah but my point is that your method of finding an "average" player is not objective.

Not finding an "average" player. Its a above and below par metric based on results of matches.

 

[RossTaylorsBox]

Isn't there a thread here specifically designed for player ratings? If I recall correctly, the basic conclusion was that you might as well just try different ratings and see which one fits best in your own mind.

 

[CaptainGrumpy]

Alright then, improve the measure. A simple 'runs to wickets' measure is inadequate. How about the weightage? Top order wickets are more relevant than tail-end wickets.

That is too subjective. Then have to do the same with hard runs and easy runs.

You say inadequate, I say its objective. The information is what actually happened on the score cards. Some teams bat well to 7, some only to 5. But a wicket is a wicket.This information is not tainted by any human distortion. That can be done afterwards if needs be by the individual - but that is going to be fraught with subjectivity because suddenly some runs are worth more than others, and some wickets are worth more than others, and everyone will draw the line in the sand differently. Then you might as well compare who the better entertainer was.

No statistic will reveal how much I prefer watching Kane Williamson bat as opposed to Steve Smith. But they are both gun players so far. But they have many matches still to play. But the statistics do reveal how interesting say AB De Villiers is to bat, especially in ODI's. Now the stats will never demonstrate those magnificent shots that he plays, but they do reveal that he was a highly effective player and runs, lot of them, were being scored quickly when he batted.

 

[RossTaylorsBox]

Yes. The individual averages are compared after they have been standardised against the era average because of difference in eras. In the times of WG Grave a wicket was worth far less than the 36 runs it was worth in the 1980's. But a wicket is still worth a wicket. That is the baseline. The contribution to being ahead of par can be listed in a ratio as a value of a wicket. Or as runs. The player's individual averages are used.

Ah sorry, I meant the average of individual averages (actually this may turn out to be the same as partnership averages).

No, Bradman at 99.94 is a mere outlier.

I don't mean outliers, I mean the distribution of partnership/individual scores is not a normal distribution. Something like 7% of all partnerships are exactly zero runs, probably something similar for innings as well. That's a skew that will greatly affect the results especially if your assumption is that everything will average out.

Not finding an "average" player. Its a above and below par metric based on results of matches.

Whatever the case, you're using a rudimentary way of calculating the average.

 

[CaptainGrumpy]

I don't mean outliers, I mean the distribution of partnership/individual scores is not a normal distribution. Something like 7% of all partnerships are exactly zero runs, probably something similar for innings as well. That's a skew that will greatly affect the results especially if your assumption is that everything will average out.

I don't understand your point. A partnership ends when a batsman gets out. The information collected includes every player being dismissed or every partnership ending. It then focuses on the individual player's dismissals for runs scored. You keep talking about this partnership thing. Every player's individual runs scored is noted. Every dismissal. Every partnership that occured in the player's era where a single runs was scored or a batsman was not dismissed is noted. You sure you're not confusing your with a red herring? I am not understanding where you are going with this.

Whatever the case, you're using a rudimentary way of calculating the average.

In my opinion, that is the simple beauty of it. To use anything else, is to start fudging what is par. Batsman need to be ahead of par because the bowlers are not as good batsman. Specialist batsman averaging mid 30's are not often celebrated because they're not scoring enough runs. They're typically not scoring enough runs because bowlers don't bat as well as batsmen. Bowlers need to be "carried" by the top order batsman, but they do almost all the bowling. Their inferior batting is countered by their better bowling. But if that bowler averages mid 30's with the bat, and mid 20's with the ball and takes wickets, that is one heck of a cricketer.

I really didn't think this would be a case of people saying stats are not perfect. I thought sports fans unbiquitously knew that.

 

[RossTaylorsBox]

Er no, stats are never perfect. Data is perfect, stats aren't.

 

[CaptainGrumpy]

Er no, stats are never perfect. Data is perfect, stats aren't.

As I said: "I really didn't think this would be a case of people saying stats are not perfect. I thought sports fans unbiquitously knew that."

Alternatively, I thought every sport fans knows stats are not a perfect measure.

 

[OverratedSanity]

You really need to stop editing your posts after people respond to you. We can see when you've edited them.

 

[CaptainGrumpy]

You really need to stop editing your posts after people respond to you. We can see when you've edited them.

I edit them so that you read the edit. People can always reply with quote of the pre-edit.

 

[marc71178]

If you want to value wickets and runs etc, why not look like at the icc rankings which place values on such things over a career.

We all agree that Sobers' batting was his strong point and Imran's bowling was his. Looking at peak ratings, Sobers had 938 with the bat, whereas Imran had 922 with the ball. Pretty close to be fair and in line with what you'd expect for the 2 of them. Obviously both peaked at number one in the world.

If we then look at their weaker suit we see Sobers has 715 with the ball and Imran 650 with the bat, so quite a big gap there. In terms of rankings, Sobers reached number 4 whereas Imran peaked at 12. Obviously we have to take into account that there's more batsmen then bowlers so it's harder to get a higher place in the batting. However, when Sobers was number 4 with the ball, he was also number 1 with the bat. When Imran was number 12 with the bat, he was number 6 with the ball, although that is a slightly false position when you consider that he wasn't actually bowling at the time, so it doesn't look as impressive in that regard.

Finally, if you look at all-rounder peaks, Sobers gets 669 and Imran 518, which starts to show the gulf between them...

 

[Flem274*]

Do you want a Murali and Akram comparison done?

no

Er no, stats are never perfect. Data is perfect, stats aren't.

*unless collected by students aka compulsive liars and maker-uperers