Batting Average by position

[complan]

Check out this article on batting average by position.

Cricket Records, Information and Everything Cricket: Progression of Test Batting Averages of Various Positions

No real surprises here! But a few interesting points -

1. Difference between the top and the bottom batting positions seem to be increasing over the years. Specialization?

2. West Indies and Australia seem to have had historically the best No. 3 batsmen. A quick search for WI reveals players like Headley, Lara, Weekes, Richards etc. Even Sobers made his 365* while batting at 3. For Australia, of course Don Bradman himself may have single-handedly brought up his country's average, not to mention others like Chappell, Ponting etc.

3. This is most surprising to me. Batting Averages from 1925-1950 hold their own, even against the modern generation. Weak bowling, LBW laws?

Thoughts?

 

[Red]

The averages by position (countrywise) is interesting in terms of which is highest for each nation.



Australia's number 3 is very high, but Australia has had some great, consistent #3s through history. Clem Hill, Bradman, Harvey, IChappell, Boon, Ponting.

South Africa's number 4 is very high, which makes sense considering Kallis has been there a lot, and others such as Pollock etc.

WIs number 3s as mentioned- Headley, Weekes, Lara, Richards, Richardson.

SL's 3 and 4, basically Sanga and Jayawardene.

England's are quite low in comparison. Probably hurt by the period between the 60s and now. Would have though Hobbs, Sutcliffe and Hutton might have pushed the opener's avgs up, but I guess there was a lot of crap there as well. Probably worthwhile for this exercise to combine #1 and #2 though, because opening is opening really.

 

[Agent Nationaux]

Not bad by Pakistan. We did have Inzy, Younis and Yousuf playing together at one point.

 

[the big bambino]

Take out Bradman's contribution and the 30s ave/wkt falls to the 2nd lowest in the decades 20s to the noughties. Only 50s is lower iirc.

 

[Migara]

Take out Bradman's contribution and the 30s ave/wkt falls to the 2nd lowest in the decades 20s to the noughties. Only 50s is lower iirc.

Take out SRT, Lara, Kallis, Sanga and Dravid in 90s and 00s averages and it will drop alarmingly as well.

 

[L Trumper]

Take out SRT, Lara, Kallis, Sanga and Dravid in 90s and 00s averages and it will drop alarmingly as well.

Eh!? Take out bradman, and add another hammond/headley it is still low. Take out SRT add another Lara/Kallis there wouldn't be any difference.

 

[the big bambino]

Take out SRT, Lara, Kallis, Sanga and Dravid in 90s and 00s averages and it will drop alarmingly as well.

Do you realise you've just mentioned 5 players to counter balance Bradman? Surely even the dim witted amongst us would then realise what an outlier Bradman is? Don't you agree? Players who ave 50 or close to it are rare but every generation has had them. Including the 30s. Players like Hammond, Hutton, Dempster, Sutcliffe and Headley. If I took their runs out your comment would be relevant.

But I haven't so your comment isn't.

 

[BulkBogan]

One interesting question is how do you compare the worth of averages based on batting position...eg. is an average of 45 as an opener worth the same as 50 from a guy batting five?

It is unquestionaly more difficult to bat against the new ball than to come in at 5 so what weighting is given to runs scored at each batting position?

 

[weldone]

It is unquestionaly more difficult to bat against the new ball than to come in at 5

No.

 

[the big bambino]

Eh!? Take out bradman, and add another hammond/headley it is still low. Take out SRT add another Lara/Kallis there wouldn't be any difference.

That's a good point. I've just now added back a fictional player of Hammond's ave in the 30s (59.70) to replace Bradman's stats. While the ave for the 30s increases (as you'd expect) it makes no difference to the 30s ranking as the 2nd lowest decade for batting averages. However Bradman alone elevated that decade's batting ave by 4 places up the ladder.

In reality you'd add back a fringe player. Someone like Chipperfield who averaged around 32. It would make little difference to the decade's overall average.

 

[complan]

One interesting question is how do you compare the worth of averages based on batting position...eg. is an average of 45 as an opener worth the same as 50 from a guy batting five?

It is unquestionaly more difficult to bat against the new ball than to come in at 5 so what weighting is given to runs scored at each batting position?

A good point. However, if we start giving weightage to batting position, the slope starts getting slippery, because we then have to factor in the quality of the new-ball bowlers, the weather conditions when the match/day starts etc. Plus fast bowlers can wreak havoc with the old ball too sometimes (reverse-swing). How about the freshness of the bowlers when we move onto the next day?

Anyway, I'm not sure how we can resolve the question of batting position. Do the ICC rankings take that into consideration?

 

[Migara]

Eh!? Take out bradman, and add another hammond/headley it is still low. Take out SRT add another Lara/Kallis there wouldn't be any difference.

The number of players in 90s and 00s (or rather number of innings) is so large compared to that of 30s and 40s. Hypothetical batsman with and av erage of 60 would make more effect on 30s and 40s group compared to 90s and 00s group. And the number is roughly about five times when player pool is compared.

 

[Migara]

Do you realise you've just mentioned 5 players to counter balance Bradman? Surely even the dim witted amongst us would then realise what an outlier Bradman is? Don't you agree? Players who ave 50 or close to it are rare but every generation has had them. Including the 30s. Players like Hammond, Hutton, Dempster, Sutcliffe and Headley. If I took their runs out your comment would be relevant.

But I haven't so your comment isn't.

Bah, still you don't get the point. Probably you would understand that you have to compare same proportion of innings across eras, because current era has lot more test innings played. If you take 5 players from 30s and 40s, you'd take 25-30 from current era.

 

[Migara]

For information of TBB:

Bradman played 80 out of 4376 innings. i.e. 1.83%

SRT played 323 out of 33576 innings i.e. 0.96%

Bradman basically played twice as much innings as SRT. SRT averages 54.3. If you replace it with a Sanga or a Kallis with higher average but still lesser innings that will equate to three or four batsmen

 

[Hurricane]

No.

?

 

[the big bambino]

Bah, still you don't get the point. Probably you would understand that you have to compare same proportion of innings across eras, because current era has lot more test innings played. If you take 5 players from 30s and 40s, you'd take 25-30 from current era.

Its a decadal comparison not a career v career comparison. If I give DGB the same % of innings as SRT in a given decade the 30s is still the 2nd lowest batting ave. That is no change.

You wanna make something else up while you're at it?

 

[L Trumper]

The number of players in 90s and 00s (or rather number of innings) is so large compared to that of 30s and 40s. Hypothetical batsman with and av erage of 60 would make more effect on 30s and 40s group compared to 90s and 00s group. And the number is roughly about five times when player pool is compared.

If we replace Sachin with another Lara there would be no discernible change. Simply because all great batsman have similar averages. But Bradman is so far above the rest, and if you are replacing Don with second best at that time you will lose roughly 45% of his output. So number of innings would be immaterial in that case.

 

[Migara]

Its a decadal comparison not a career v career comparison. If I give DGB the same % of innings as SRT in a given decade the 30s is still the 2nd lowest batting ave. That is no change.

You wanna make something else up while you're at it?

Why SRT's batting average? It should be the global batting average. The whole "Bradman effect" is blown out of proportion. Every good batsman has dented bowlers stats. Bradman had done it more than others. That doesn't make bowlers who bowled to him Marshalls, Hadlees, Warnes or Muralis.

 

[Migara]

If we replace Sachin with another Lara there would be no discernible change. Simply because all great batsman have similar averages. But Bradman is so far above the rest, and if you are replacing Don with second best at that time you will lose roughly 45% of his output. So number of innings would be immaterial in that case.

Second best batsman argument is a a joke. It should be the average batsman who averages 28-32 range. Bradman will have more effect. But Ponting + Kallis (which brings the same innings %, or little lesser) would have done it in substantial amounts too.

 

[Migara]

The whole argument boild down to what would the change of average if there was an average batsman instead of a great one.

When the batsman averages 100 (and the global average is 30), the difference is 70 x 0.018 = 1.26 runs.
At average of 58, it is 28 x 0.018 = 0.504. The net difference Bradman brings is 1.26 - 0.54, that is 0.7 runs, so called the "Bradman Effect".

0.018 since Bradman played 1.8% of test innings during his period.