Rik said:
Because although you can argue for all your life about stats, they are useless
Yes, I rather agree.
unless you take the circumstances into account. Smith bats at number 3, Collingwood at 4-5. Collingwood bats in the CC 2nd Division, Smith bats in the CC 1st division. I wouldn't be disagreeing if it didn't make sense.
Seriously, though, I quite agree that if one must deploy statistics, let them at least be sensible ones with decent context.
I don't think it's reasonable to take into account Collingwood's performances after recovering from injury at the dog-end of last season, the only point of which were, as far as I could see, to prove to the selectors that he was actually fit enough to go on tour.
In the previous two seasons, Collingwood averaged 50 batting at 4/5 in div 2 while Smith averaged 40 batting at 3 in div 1. (In 2002, Collingwood's record is uncomplicated since he had no not outs and still averaged about a dozen more than Smith.)
On the raw numbers, Collingwood wins hands down. BUT, but, but HOW do you assess whether a div 1 40 is better, worse or the same as a div 2 50? Is the difference between the divisions really so large that it accounts for being 25% better? I'd happily concede 10%, but I wonder whether 25% isn't pushing the bounds of credulity a little far - given that both men's seasons included matches against several of the same teams over the two years owing to promotion and relegation.
I've just had a little dig through Smith and Collingwood's big scores in 2001 and 2002. I'm slightly interested to note that all of Collingwood's three centuries were against teams which had at least one current or very recent Test bowler (Vasbert Drakes, Andy Bichel and Vaas/Zoysa/Chuckera), whereas Smith's four were against one side with no international bowler, one against Glamorgan boasting Steve Watkin who hadn't played for England for yonks (and Collingwood managed a 99 off them himself), one against Leics involving Ormond and Devon Malcolm and 103* coming in at #7 (not #3) against a Yorkshire attack with 2-3 Test bowlers in it.
As samples go, that's not very big, so it would be daft to draw any hard-and-fast conclusions, but there is at least a hint that Collingwood is a little more successful than Smith against international class bowling.
And of course, just looking at the scorecard doesn't tell you anything about whether the bowlers were actually bowling well (although wickets seem to have fallen more rapidly around Collingwood than Smith, for what little that may be worth) or what the conditions were like. For all I know, one of Smith's 40s could have been a far more accomplished innings than his ton against Leics.
I can't find any way in which you can sensibly slice the statistics in Smith's favour - except in that Collingwood was a later developer and has more early failures on his statistical conscience which probably ought not to be considered all that relevant today. In order to get the figures to work in Smnith's favour, it seems to me, you have to add qualifications and adjustments, which means you then also have to find out whether the assumption (Div 1 > Div 2, #3 vis-a-vis #4) is what's really at work - and when you do that, lo and behold, there is prima facie evidence that the assumptions are not in fact valid in this particular case however reasonable they might seem as generalisations.
I'm very grateful to you for this debate, as it has pushed me into doign a lot more research about Collingwood and Smith than I'd ever considered doing before, and I'm fascinated to find that each new angle of argument seems to reflect more and more favourably on Collingwood. It's been very educational for me indeed.
Cheers,
Mike