Best batsman in cricket (ranking by median)

[silentstriker]

I was interested in using another method besides average to look at a player's score, and the obvious one is median. Medians can tell you a lot that averages miss, and considering it tells you the halfway point of all the innings (eg. half the innings were above this mark, and half below), it gives a different perspective of what someone is 'likely' to score.

Someone with a high median and a low average says a lot about their batting. This is obviously not the only criteria you should use to rank someone, but I found the results extremely illuminating, and you can tell a lot about some players by this stat. One thing it doesn't do is take into account not outs.

If you want to know a player's median, let me know. In case of even innings, the two medians are averaged.

Ranked from the highest median to the lowest:

1. Bradman: 56.5
2. Hobbs: 40
3. Dravid: 35
4. Kallis: 34.5
5. Lara: 33.5
6. Hayden: 33
7. Ponting: 32
8. Hutton: 32
9. Tendulkar: 31.5
10. Gavaskar: 29

 

[Ikki]

I was interested in using another method besides average to look at a player's score, and the obvious one is median. Medians can tell you a lot that averages miss, and considering it tells you the halfway point of all the innings (eg. half the innings were above this mark, and half below), it gives a different perspective of what someone is 'likely' to score.

Someone with a high median and a low average says a lot about their batting. This is obviously not the only criteria you should use to rank someone, but I found the results extremely illuminating, and you can tell a lot about some players by this stat. One thing it doesn't do is take into account not outs.

If you want to know a player's median, let me know. In case of even innings, the two medians are averaged.

Ranked from the highest median to the lowest:

1. Bradman: 56.5
2. Hobbs: 40
3. Dravid: 35
4. Kallis: 34.5
5. Lara: 33.5
6. Hayden: 33
7. Ponting: 32
8. Hutton: 32
9. Tendulkar: 31.5
10. Gavaskar: 29

SS, if it isn't too much trouble could you please look at: Viv Richards, Greg Chappell, Alan Border and Steve Waugh. Thanks.

 

[silentstriker]

1. Bradman 56.5
2. Hobbs 40
3. Dravid 35
4. Kallis 34.5
5. Lara 33.5
6. Hayden 33
7. Viv Richards 32.5
8. Ponting 32
9. Hutton 32
10. Tendulkar 31.5
11. Chappell 31
12. Border 31
13. Gavaskar 29
14. S Waugh 25.5

 

[open365]

Mathew Hayden?

 

[Fusion]

How about a few Pakistani's: Javed, Inzi, Yousaf.

 

[silentstriker]

Mathew Hayden?

He is there.

How about a few Pakistani's: Javed, Inzi, Yousaf.

1. Bradman 56.5
2. Hobbs 40
3. Dravid 35
4. Kallis 34.5
5. Lara 33.5
6. Yousuf 33.5
7. Hayden 33
8. Viv Richards 32.5
9. Ponting 32
10. Hutton 32
11. Tendulkar 31.5
12. Chappell 31
13. Border 31
14. Miandad 30
15. Inzamam 30
16. Gavaskar 29
17. S Waugh 25.5

Very interesting stuff. Surprised that every time Waugh came to bat, half the time he'd score less than 26. The rest of the players are pretty close together, with Hobbs above the next guy as much as Waugh is below. Obviously, Bradman is a freak.

 

[cpr]

Good way of looking at it IMO. Obviously someone like Lara is skewed by the 400 and 375, count nothing more than any other hundred in this (doesnt make this way perfect mind, been a few times Lara's scored 100-200 and REALLY turned a game, that doesnt get reflected). Likewise, those who could bash a weak attack, but regularly got out cheaply to decent teams are going to get penalised.


Be interesting to see the figures for some solid middle order batters imo, someone who rarely wavered, hardly shined. Cant think who to say off the top of my head. Be interesting to see a few more Englishmen in there aswell, particularly Boycott, Gooch and Atherton.

Comparing Botham, Khan and Dev this way would add some fuel to a few fires too

 

[Perm]

How exactly are you working this out ss? I'd be interested to see and work out a couple for myself if possible.

 

[cpr]

If its what i'm thinking, then its listing all the innings scores of a players test career in numerical order, and finding the mid point.

So if someones played 100 innings, they might have 4 ducks, 5 centuries, so it'd line up 0,0,0,0,......then upwards. The mid point would be between the 50 and 51st number in the secquence, so if they were 35 and 36 for example, the figure would be 35.5

Basically if someone has a habit of getting out before they reach 30, but has hit some MAJOR scores in their career, those big scores wont affect the result as much, but the numerous poor performances would push the median figure down. Likewise, someone who might get the nervous 90's every time, but rarely gets out for less than 25 will see their median figure be higher

 

[Flem274*]

This sounds interesting. A few Kiwis perhaps?

Maybe Fleming, Astle, Suttcliffe, Cairns, Crowe?

 

[Perm]

This sounds interesting. A few Kiwis perhaps?

Maybe Fleming, Astle, Suttcliffe, Cairns, Crowe?

I already had Fleming and Astle done, comparing them to some of the other batsman to play for us in recent times (Styris, Oram, McCullum, Marshall). Will do the others now.

Fleming - 27
Astle - 20.5

 

[Athlai]

Hussey? For the Kallis-Hussey debate.

 

[Perm]

Bert Sutcliffe - 23
Martin Crowe - 39
Chris Cairns - 20
Michael Hussey - 58

 

[Athlai]

58? Better than Bradmans

 

[Perm]

58? Better than Bradmans

Yup. 5 centuries in 26 innings is pretty damn impressive, and he's only had three single figure scores.

 

[Speersy]

Kevin Pietersen - 37
Adam Gilchrist - 27

Did these two then got bored. Did you make a spreadsheet ss?

EDIT: Gilchrist has had 46 innings below 10, including not outs. He has never been out or not out on 8!

 

[silentstriker]

If its what i'm thinking, then its listing all the innings scores of a players test career in numerical order, and finding the mid point.

So if someones played 100 innings, they might have 4 ducks, 5 centuries, so it'd line up 0,0,0,0,......then upwards. The mid point would be between the 50 and 51st number in the secquence, so if they were 35 and 36 for example, the figure would be 35.5

Basically if someone has a habit of getting out before they reach 30, but has hit some MAJOR scores in their career, those big scores wont affect the result as much, but the numerous poor performances would push the median figure down. Likewise, someone who might get the nervous 90's every time, but rarely gets out for less than 25 will see their median figure be higher

Yup.

 

[Swervy]

this is interesting stuff....the issue with not outs needs to be addressed though, but yeah nice one SS

 

[jeevan]

Wouldnt the median be somewhat insenstive to not outs (unlike average). No division involved.

Definitely seems like looking at average & median both tell a story (as in many other forms of statistics) and makes you wonder why no one else thought of it before! Kudos to SilentStriker.

Perhaps a histogram of scores per batsman 0-9,10-19,etc would be instructive. Since we principally use the web, can see this as being possible to maintain easily.

 

[silentstriker]

Wouldnt the median be somewhat insenstive to not outs (unlike average). No division involved.

It is. Though I can argue that the current method of calculating an average is overly sensitive to not outs. You have some tail enders whose average is sometimes higher than their highest score. But you are right, a not out is an important part of batting and this is ignored by median. Of course, a lot of batsmen (esp. openers) and higher middle order have a very low amount of not outs, so its not a huge problem until you start getting to #7 or so.

Definitely seems like looking at average & median both tell a story (as in many other forms of statistics)

Yup, they are both important IMO.

Perhaps a histogram of scores per batsman 0-9,10-19,etc would be instructive. Since we principally use the web, can see this as being possible to maintain easily.

Not a bad idea, I'll work on it.