Tuesday, December 18, 2012

end of the century

So the question I was asked, following the last post, was: that's all very interesting, but is 99 a statistical outlier in this regard? Are you more likely to be out for 99 than, say, 92, as the nerves set in as you approach the landmark score? And does it drop off again after you get to 100? These are all good questions and cannot be left unanswered.


So here's a graph of the number of times people have been out for each of the scores in a range of ten runs either side of 99 (i.e. 89 to 109). A couple of obvious things stand out, most obviously that 99 clearly isn't a statistical outlier in terms of the number of people who've been dismissed for that score, indeed more people have been out for 89, 90, 91, 92, 93 and 96 than have been out for 99, though you could argue that those are just nervous nineties syndrome getting a grip a bit sooner for some people than for others. However, you'll also notice that more people have been out for 100 than for 99. Perhaps we're seeing two overlapping phenomena here, with twitchiness approaching 100 immediately giving way to relief, euphoria and carelesssness after reaching it.

What you'll also notice, however, if you look at the smaller (beige) columns, is that 99 (with fifteen victimsis a statistical outlier in terms of the number of people who have been run out for that score. The nearest competitor in the range we're looking at here is 90, a score on which seven people have been run out, including the legendary West Indian Sir Everton Weekes, ten runs short of what would have been his sixth Test century in successive innings in January 1949. So while in general there seems no reason to conclude that people are any more likely to get out for 99 than any other score in the immediate vicinity, there certainly does seem to be some reason to believe their judgment of a quick single may be compromised.

Saturday, December 15, 2012

99 flakes

Another morsel of cricket-related statistical trainspottery and trivia for you: when the Indian captain Mahendra Singh Dhoni took a quick single to mid-off earlier today and got himself run out for 99, he became the 15th man in Test cricket history to perish in such a way. I know this because of the excellent Statsguru facility provided by Cricinfo, which permits all sorts of arcane queries to be answered. So if I want to know how many batsmen have been stumped for exactly 64 runs while batting at number 4, then I can easily ascertain that the answer is just one, our very own Kevin Pietersen.


Getting run out for 99 seems especially poignant, though - setting off for a run with the expectation of being able to raise your bat in acknowledgement of a century, only to watch in horror as the stumps are demolished with you either still floundering halfway down the pitch, or having to undergo (as Dhoni did) the drawn-out agony of a referral to the third umpire and a prolonged series of TV replays. Here are South Africa's Jacques Kallis and Neil McKenzie getting run out by the same fielder (Australia's Damien Martyn) within 3 months of each other in December 2001 and March 2002 respectively, and, most agonisingly of all, England's Mike Atherton falling over, getting up and then falling over again for what seems like hours before Ian Healy puts him out of his misery against Australia at Lord's in 1993. On the other hand, Atherton did make plenty of Test hundreds (sixteen in fact, although he never made one at Lord's) so perhaps the award should instead go to New Zealand's John Beck, who was run out for 99 against South Africa in 1954 and never made a Test hundred.

The run out for 199 club, by contrast, has only one member: Pakistan's Younis Khan.

Wednesday, December 05, 2012

tis the season to be pedantic

It's December, Christmas is in the air, there's a bit of frost on the ground in the cold winter mornings....or is there? Well, yes, the frost is definitely there, but what I mean is: is it winter yet?

It's funny how stuff that's been in place since before you were born seems normal and generally passes without question, even when it's fundamentally absurd when you stop to think about it. I'm sure you can see where I'm going with this - this is basically how religion works. But that's not what I'm about here.

My concern here is: when does winter start? And, more generally, what are the dates of the seasons? Now obviously I'm taking a northern-hemisphere-temperate-zone-centric view of the world here, because, well, that's where I live, and some of the terms I'm going to be bandying about are only really relevant to that region. If you live at a similar latitude south of the equator you'll find your seasons are out of phase with ours by six months, whereas if you live near the equator you'll probably find significantly less year-round variation.

My recollection, for instance, of living in Java in the late 1970s is that there was very little year-round variation in temperature, but that half the year was designated the "dry season" and the other half the "rainy season", the only difference between the two being that during the rainy season it would rain, torrentially, at 4pm every afternoon, regular as clockwork. One other interesting factoid from that link is that at Indonesia's latitude the difference between the longest and shortest days of the year (in terms of daylight, yes, I know they're 24 hours long everywhere) is a mere 48 minutes (it's about 6-8 hours in the temperate zones).

We're not really getting anywhere here; let's try and focus. When's the first day of winter? The winter solstice? Having December 21st (or thereabouts) as the first day of winter doesn't seem absurd, particularly when we in the UK are used to the coldest months typically being January and February anyway. But it seems somehow less sensible to have June 21st being the first day of summer; most of us would instinctively feel that it should be a bit earlier than that. And what of the convention of calling June 21st "Midsummer's Day"? Or of calling December 21st "Midwinter's Day", come to that? I don't think people would necessarily demand that it be smack dab in the middle of the season, but it being right at the start seems wrong.

So what to do? Well, one could adhere rigidly to the "midsummer"/"midwinter" thing and say, OK, the seasons are 365/4 = 91 days long (give or take the odd day) so we'll just go 45 days each side of the solstices and fill in the gaps from there. This gives you the seasonal dates as follows:
  • Spring: 5th February to 6th May
  • Summer: 6th May to 5th August
  • Autumn: 6th August to 4th November
  • Winter: 5th November to 4th February
Those start dates, on the other hand, seem a bit early. So we could do what the UK Met Office do and designate the first day of the month in which the solstice or equinox falls as the first day of the season - this is quite handy because it divides the year into nice easy-to-remember chunks of three whole months each, as follows:
  • Spring: March, April, May
  • Summer: June, July, August
  • Autumn: September, October, November
  • Winter: December, January, February
This is fine, and a convention I'd be perfectly happy to adopt, but the important point here is that there isn't really a definitive answer to this question. Again, I'm perfectly happy with that; life in general is, after all, a series of grey areas with fuzzy and ill-defined boundaries.

It's a useful indicator of personality type, though, to see how people respond to the notion that there is no answer - the more authoritarian types (small-c conservative, broadly speaking) get vaguely uncomfortable and annoyed that there isn't some central authority that will TELL THEM THE RUDDY ANSWER, on this topic as on many others where you just have to stick an arbitrary stake in the ground and no two people agree on where it should be: things like the age of consent, the safe weekly intake of alcohol, that sort of thing.

Bob Altemeyer's fascinating, very readable to non-academics, and freely downloadable (in PDF format) The Authoritarians is the canonical work of behavioural research on this topic. It doesn't say much about when winter starts, but, hey, nobody's perfect.

Monday, December 03, 2012

is it a bird? no. is it a plane? yes; yes it is

Here's a thing you might like to try: take a trip - cyberspatially I mean - along the M4, via the magic of Google Maps, and turn off at junction 10 onto the A329(M) (and thence the A329 and A322). Then head down through Bracknell and past the prestigious portals of The Berkshire golf club towards Bagshot and the M3. Wait a minute, you'll think to yourself, as I did, what's that in the bottom right corner of the picture there, half a mile or so north of Bagshot railway station?


Let's zoom in a bit. Crikey, it's the world's biggest plane parked in a field. Or possibly an up-to-date aeroplane-shaped version of one of those chalk hill figures typically found further west in Oxfordshire, Berkshire and Wiltshire.


Of course what it actually is is a regular aeroplane that just happened to be on a flight path that took it under the Googlecopter at the precise moment that it took its picture of that particular area. Have a look at it on Google Maps here - zoom in on the tail and you'll see it's orange, which would suggest an easyJet plane, but the blue engine nacelles and the suspicion of a logo by the front side windows suggests maybe it isn't. Since Heathrow is probably less than 20 miles away to the north-east it seems more than likely it took off from there, and I don't think easyJet operate out of Heathrow, so I don't know. Nor do I know what sort of plane it is; maybe someone can help me out.

It turns out that this sort of anomaly isn't all that uncommon - here's one in New York, one in Los Angeles, one in South Carolina, one in Chicago, one in Florida, one in Germany and one in Russell Square, London. That last one is from this list, and you'll notice there's another entry in that list which says British Airways Boeing 777 In Flight Over Bagshot, which on the one hand is slightly galling as it means I haven't got a scoop, but on the other hand is good as it identifies the plane and airline, assuming we trust the link, and there isn't very much context at that link to back it up, though this link seems to confirm the identification. The Boeing 777 ticks some basic boxes like having the right number of engines, though; the orange tail must be one of those crazy commemorative ones.

Don't bother actually going to Bagshot to have a look at it, though, as I expect it's gone by now.