## Weighted bowling averages

Martin said that he'd like to see a bowling average in which each wicket is weighted by the batting average of the batsman dismissed. I have now calculated these averages for all bowlers who have taken 50 or more Test wickets. The figures are current to Test 1847, the second Test between South Africa and New Zealand.

It is interesting that this one weighting factor appears to even out bowling averages across eras remarkably well. The top of the list of lowest bowling averages is dominated by players from the 1800's. But weighting by the average of the batsman dismissed largely cancels out the effect of low scoring, because the batsmen have low batting averages.

Note that I've used the batsmen's career averages, rather than averages at the time of dismissal. This, while making my calculations easier, has the unfortunate side-effect that a bowler's adjusted average can change even when he's not playing.

Here is a table showing the top twenty bowlers, according to this adjusted average. The columns of the table are matches, runs conceded, wickets taken, usual average, adjusted wickets, adjusted average. If I am generous to me and Martin, I could impose a 100-wicket threshold, and note that SF Barnes comes out number one, as he should. I would treat these figures as interesting and not, of course, take them as the be-all and end-all.
`                  m   r     w    avg     adj w    adj avgIronmonger,Bert   14  1330  74   17,97   72,01    18,47Clark,Stuart      11  1093  54   20,24   58,32    18,74Ferris,J.J.       9   775   61   12,70   39,1     19,82Asif,Mohammad     11  1180  51   23,14   56,83    20,76Barnes,Sydney     27  3106  189  16,43   149,02   20,84Lohmann,George    18  1205  112  10,76   56,19    21,45Davidson,Alan     44  3819  186  20,53   176,09   21,69Marshall,Malcolm  81  7876  376  20,95   361,17   21,81Ambrose,Curtly    98  8501  405  20,99   386,46   22,00Bowes,Bill        15  1519  68   22,34   68,25    22,26Turner,Charlie    17  1670  101  16,53   74,86    22,31McGrath,Glenn     123 12144 560  21,69   542,99   22,37Higgs,Ken         15  1473  71   20,75   65,8     22,39Laker,Jim         46  4101  193  21,25   181,42   22,61Bond,Shane        17  1769  79   22,39   77,87    22,72O'Reilly,Bill     27  3254  144  22,60   143,21   22,72Croft,Colin       27  2913  125  23,30   125,29   23,25Miller,Keith      55  3906  170  22,98   166,98   23,39Adcock,Neil       26  2195  104  21,11   93,74    23,42Trueman,Fred      67  6625  307  21,58   279,67   23,69`

You probably want to know about Murali and Warne. Murali comes in at #30, with an adjusted average of 24,53. Warne is at #74, with an adjusted average of 28,05.

There are a number of ways this analysis could be taken further, but I will save myself the trouble and simply link to Charles Davis (see the table at the bottom and discussion just above).

Very interesting... how can you put things in laymans terms?

For eg, does Warne's position in the list mean that (on avg) he dismissed more batsmen with a lower average compared to someone like Asif? I know my statement is too genaral and it doesnt factor in the bowling avg... but Im just trying to understand :)

Hi Obaid: What you've said about Warne and Asif is pretty much right, though concluding stuff about what sort of batsmen a bowler dismisses involves comparing the usual and adjusted averages.

Warne's usual average is about 25, and his adjusted average here is about 28, which means that his wickets were, roughly speaking, more likely to be of worse batsmen. Asif, on the other hand, had his average adjusted downwards (from 23 to 21), so he's been dismissing better batsmen.

There's one factor here that's working against Warne, and that is that Australia's opening bowlers often dismissed most of the opposition's top order, leaving only tailenders for Warne. (Warne has a very high proportion of his wickets as tailenders.) Asif, as a paceman, usually gets a crack at all of the best opposition batsmen.