Wednesday 8 September 2010

2010 T20 Champions' League Preview Thought

One of the questions I have tinkered with off and on is trying to gauge the level of quality of different domestic cricket competitions. My limited research so far suggests that Australia's domestic level in first-class cricket is very high, and quite probably higher than some Test series between weaker sides.

On the surface, the T20 Champions' League might seem to make an easy job of it, but the number of matches played is so few that I would be very wary of using it as a barometer. However, the more of these competitions played, the better we'll be able to make judgements. What will be particularly of interest this year will be whether Indian sides again fail to rise above the mediocre level.

Sunday 29 August 2010

2010 Match-Fixing Scandal

Knee-jerk reaction is never altogether helpful in a matter like this.

However, listening to The World This Weekend on the Home Service this lunchtime, one heard exactly what the root of the problem was: How does a young Pakistani cricketer make a realistic wage in an industry where his place is subject to the whims of an unstable organization? Mihir Bose suggested that ready money from the gamblers trumped the graft of becoming a big star and gaining sponsorships.

Bose said that the problem was that international cricket 'lacked the teeth' to enforce the real change needed to end the repeated influence of gamblers on international cricket once and for all. He blamed the relaxed gentlemanly attitudes of the 'English club', as opposed to the French-style officiousness of other international sporting organizations such as FIFA. Then he suggested that the BCCI would never give up the power it wielded to an international body. So the end is that there is no solution.

I'd argue that if the problem is the distribution of income to players, then the solution rests in the players' hands, as the relaxed gentlemanly attitudes suggest. Basically, Pakistani cricketers need a strong players' organization to fight on their behalf to ensure that selection is stable and that the rewards of playing international cricket are sufficient that only those who are not sporting cricketers in the first place will be tempted by the gamblers.

In other words, a strong and powerful union, which will need to function at an international level to be effective.

Saturday 28 August 2010

Pakistan in England, 2010, 4th Test.3

Hic sunt leones, a medieval cartographer might say of the parts unknown in which we find ourselves, win expectancy-wise.

Basically, England have about a 70 per cent chance of winning this match. But we've seen some violent swings in win expectancy today, and I'd be surprised if there aren't more during the rest of this match.

Friday 27 August 2010

Pakistan n England 2010, 4th Test.2

Phew! You'd think Pakistan would be on top of the world after that first session. And you'd almost be right.

Unsurprisingly, breaking out the win expectancy calculator reveals that Pakistan's best chance of winning occurred after the fall of Morgan's wicket.

However, in between, surprisingly, the fall of Collingwood's wicket saw a big swing to England. When wickets fall fast, it's a case of what's sauce for the goose is sauce for the gander, as the pattern is set for the match. The side bowling second gets all the traditional advantages of the side bowling second, so wickets tumble even faster.

Pakistan's key wickets were those of Pietersen and Morgan. Even better for Pakistan, the current Trott-Prior partnership has seen England's win expectancy fall to its lowest level of the match, as the chance of a draw gradually rises.

At the moment, Pakistan have about a two-thirds chance of winning this match, and nearly a 9/10 chance of not losing it.

Thursday 26 August 2010

Pakistan in England 2010, 4th Test.1

The dismissal of Strauss had minimal impact on the Win Expectancy. England's chances of victory went down less that 0.1 per cent. Which is just another data point in my developing case that you don't actually need a big run total from your opening pair.

Wednesday 25 August 2010

The Bradman Class and Sabermetrics

My link in yesterday's blog, which was to a discussion on the validity of sabermetric methods for cricket, was itself generated by the publication of this paper in The Journal of Quantitative Analysis in Sports.

The authors propose a new method of evaluating Test batsmen, which tries to incorporate the consistency of a batsman's innings. Someone like Alistair Cook has a high average this year, but it's a consequence of a couple of big innings, rather than consistency. His median score is quite low. The question the paper's authors don't address, however, is whether a consistent batsman is more likely to produce Test match wins. What they do look at is the relative importance of a batsman's average to that of his team-mates, but this isn't the same thing.

This is where most of what passes for sabermetric research in cricket misses the point. Ranking batsmen is jolly good fun for a discussion down the pub. What I want to know, however, is the relationship between runs and wins.

And that's what I still regard as my key discovery: it is more important to stop your wicket from being taken than to score masses of runs. This means it is more important to have bowlers who take wickets than to have batsmen who score runs. The secret to success in Test cricket is the right balance between how quickly you take wickets, and how slowly you lose them. The first building block in this is to find bowlers, not batsmen. And that raises an interesting question about the relationship between bowlers' overuse and injuries or declines in effectiveness.

Tuesday 24 August 2010

Getting That Wand Out Again

In the 1988 Baseball Abstract, Bill James famously 'broke his wand'. He'd been increasingly annoyed, he said, by the abusive attitude of people that fame had brought him into contact with. So he gave up writing the Abstracts and turned to other projects. Time proved this episode to be a minor blip in the Long March of Sabermetrics, but for about a year people like me were left wondering from where we would get our annual fix about the 'objective understanding of baseball'. In the end, the rights to the Abstract were passed on, and Don Malcolm and Brock Hanke eventually started to produce their own baseball annuals that kept the flame alive until Usenet equipped amateur sabermetrics with a quicker means of debate and discussion. Out of this came Baseball Prospectus and all the sabermetric sites and thoughts we know today in the baseball world.

Last week, I noted that the Journal of Quantitative Analysis in Sports had published a study of batsmen. What I didn't know until yesterday was that Andy Bull, at The Guardian, had written about it. In the comments below his article, a cursory debate about the validity of sabermetrics for cricket ensued. Sadly, I missed all this.

I'm more than a trifle annoyed that, for all the large gaps in my publishing record, someone like Andy Bull has never, ever mentioned that Phil and I were the first people publicly even to connect sabermetrics to cricket. There are plenty of good and bad reasons why I don't post frequently, but all the conclusions I believe are valid have been published in some form or other, either here or on the old site. We deserve some credit for that, the same as F C Lane and George Lindsey get credit in any sound history of performance analysis of baseball statistics.

I've allowed this to turn into a sour rant, but I'm going to let it stand anyway. I actually meant to talk about the scepticism shown by many posters about the validity of sabermetric methods applied to cricket statistics. In particular, the comment by a poster named 'quebecer' that sabermetrics is not transferable to cricket 'except in the broadest possible sense' is, I think, based on a misunderstanding of sabermetrics. However, I'll leave that for another day, or answer him by posting more findings over time, to illustrate why he's wrong.

Monday 23 August 2010

Test Ladder Updated

I have updated the Test Ladder through the start of the 2010 Northern Hemisphere season.

Two changes occurred:

Australia rose to the top of the ladder owing to the win over New Zealand in New Zealand. (A drawn or lost series would have seen India pass Australia to claim second place.) England's draw against South Africa weighed too heavily in the face of Australian successes against weaker teams.

New Zealand, despite the loss at home, still managed to accumulate enough points to get past the makeshift West Indies' elevens beaten by Bangladesh (hence the unlikely high placing of the weakest of Test sides) and Australia.

England currently look likely to head into the Ashes series in top spot again, though.

Sunday 22 August 2010

Alastair Cook, Stay or Go?

Alastair Cook is 25 years old, a left-handed bat, and bats in the first, second or third positions in the batting order during his career.

I took all batsmen who meet those qualifications in what might be called the 'Bangladesh Era' of Test cricket (since 2000), plus all England batsmen who also meet those qualifications, with a minimum of ten innings of batting. I then compared their Innings Average, their Bowling Overs per Innings and the Standard Deviation of their innings' scores. For those who think a hundred against Bangladesh might only be worth, say 80 runs against a better Test side, I also calculated the Innings Average by adjusting the value of Cook's runs downwards by 10 per cent, 20 per cent, 25 per cent and 33.3 per cent. (When you get to 50 per cent, you may as well exclude Bangladesh altogether, so I did that, too.) I'll rank them by Innings Average.
Player           InnAve   BO/I     StDev   Dates
Cook (-10%) 48.02
Alastair Cook 45.44 13.70 49.87 2009/10
Cook (-20%) 40.99
Kumar Sangakkara 40.40 13.00 29.39 2002/03
Cook (-25%) 39.88
Cook (no BAN) 37.91 42.96
Cook (-33.3%) 37.87
David Gower 37.62 13.63 30.02 1982/83
Wavell Hinds 36.70 11.50 34.09 2002
Salman Butt 34.80 12.26 30.59 2009/10
Graeme Smith 32.24 8.32 26.60 2006/07
Mark Butcher 30.87 13.70 33.27 1998
Sadagoppan Ramesh27.58 10.84 22.29 2000/01

The idea of dropping Cook because of his current poor run of form overlooks the fact that he isn't actually having a bad year. In fact, in this set of batsmen he is having the best age-25 year of them all, with the best Innings Average and tied for the best BO/I numbers with Mark Butcher's 1998.

The only warning sign is the very high standard deviation in Cook's innings' scores. As you can see, the better players in this list have much narrower deviations. As in the case of Ramesh, however, it is possible to have too narrow a deviation. My studies of the significance of standard deviations are still rudimentary. It does seem that there is a sweet spot of these in which class batters tend to maintain, and players who fall either side of this are flattering to deceive if they seem any good.

So, if there's an obvious alternative to Cook, then it might be worth including him in any Ashes squad. But Cook still seems to deserve his place, as far as the statistics can show us.

Friday 20 August 2010

3rd Test, England vs Pakistan, 2010, Tea 3rd day

Well, actually, I looked at the win expectancy at the fall of Cook's wicket.

England should not take heart, although on a limited sample. Their chance of winning is poor, and a Pakistan victory seems the most likely outcome. England have a better chance of battling to a draw.

Of course, this is mercurial Pakistan, and if any team can upset the win expectancy apple cart, it is them.

Monday 19 July 2010

No Play Galle

Wet conditions prevented Sri Lanka and India from getting any play in today at Galle, in the first test of India's tour. Who is more likely to be frustrated, Sri Lanka or India?

Looking over previous Galle tests, you'd have to think it was Sri Lanka. They made an excellent start, and half of those teams whose first innings total was 250 or more went on to win the match here.

The only loser was a 2001 West Indian side, despite a 178 from Brian Lara. Loose bowling and a top-order collapse in the second innings ended their chance of a victory. Reading the match commentary, it looks as if Captain Hooper's tactics were a little faulty.

India managed to win a match here over four days in 2008 after making 329 in their first innings, so Sri Lanka still have some hope, provided the weather co-operates.

Thursday 15 July 2010

Lord's Collapsed

Way back in 2004, Phil and I devised a method of defining what exactly a collapse in cricket was.

The working definition we came up with was:

A collapse occurs when three or more wickets fall for 60 or fewer runs, with no partnership adding more than 35 runs.

Well, looking at the current test at Lord's we see a lot of collapsing going on.

In their first innings, Australia suffered a collapse when wickets 4 through 9 went for 51.

In their first innings, Pakistan never really stopped collapsing:
Collapse #1 Wickets 1-3 54/3;
Collapse #2 Wickets 3-5 29/3;
Collapse #3 Wickets 5-9 58/5;
Collapse #4 Wickets 7-10 31/4.

And then Australia collapsed again in their second innings: Wickets 2-4, 36/3.

Friday 12 February 2010

Conventional Wisdom Wisdom

In the comments to this article it is suggested that conventional wisdom says the best batsmen should be third in the order.

While baseball sabermetrics has challenged the conventional wisdom of the lineup, in cricket my haphazard studies of win expectancy suggest that in test cricket's case the conventional wisdom is absolutely spot on. The crucial wicket to defend are the second and third ones. If the first wicket goes cheaply, it's important that the second one goes for a par score. If the first two wickets go down for not-very-many, it's vital that the third wicket puts up a high score.

Once you're three down and below par, you're in real trouble.

Monday 8 February 2010

2010 Nagpur Test, India vs South Africa #2

My model seems to have been right! I haven't actually finished fiddling with it, and now I'm not sure I will bother to get it done for this test.

However, I was looking at the scorecard after close of play, and I wondered about the difference between the standard deviations of the two sides' first innings.

Including extras, the standard deviation for the South Africans was 89.73. That for the Indians was 32.51. Does this have any meaning?

It may indicate that Nagpur's wicket is treacherous, at least for this test. If you can get in, it's possible for an exceptional batsman to put up a decent score, but otherwise you're scrabbling for runs and in danger of losing a wicket at any moment. If that's the case, it makes this is a result pitch, and future tourists should take note.

Sunday 7 February 2010

2010 Nagpur Test, India vs South Africa (Win Expectancy)

I tuned into the play until lunch yesterday, and I managed to catch all the wicket-taking excitement.

I left off mostly talking about win expectancies. In this case, one needs to tinker with the model a little, since the situation is rather unique. Here's how significant the first wickets were.

After Prince was the victim of an umpiring mistake, South Africa had boosted their chances of winning by +.007. (There's a joke there, somewhere.)

Smith, however, 'popped out', as they say in baseball, to the wicketkeeper, and reduced the chance to 19 per cent, a loss of -15.

Kallis and Amlah set up 'shot' and hey-ho, took South Africa's chances of winning to 100 per cent. Except I don't believe that. The model has spit out a 'porky'. I'll come back and report after I've done a bit of figuring. I think I know how to handle this.

Meanwhile, readers with longer memories might like to recall this match.