Monday, 18 October 2010

Time your babies.

Statistics can be a pretty dreary subject and I promise not to go on and on about it, but greatly to my surprise I've just learnt something from the Sun newspaper, other than what Steven Gerrard's wife looks like in close-up in a very scanty swimsuit.

I wasn't really wanting to drool over a picture of Steven Gerrard's wife, of course not, but I was on a train last Saturday, and another man had been engrossed in it for 10 minutes or so, and when he reached his destination and left the paper behind it seemed a pity not to pick it up and see what he found so interesting. Hmmm!

Where was I? Oh yes, in the same newspaper there was an article (16.10.10, p. 21) saying that "Mum Barbara Soper beat odds of 50 million to one when her new baby was the third to be born on a date where all three numbers were the same". Baby Chloe arrived on 08.08.08, brother Cameron came on 09.09.09 and little Cearra Nicole arrived on 10.10.10. Barbara, of Rockford, Michigan is not, thank goodness, aiming to produce another child on 11.11.11.

According to the Sun journalist "Statistics experts say the odds of such an achievement are more then 50 million to one".

Well you don't need to be a statistics expert, and you don't even need to have a computer, though a pencil and a small piece of paper helps, to know that the odds are less than 50 million to one, they are in fact 48,627,125 to one.

But as Doug pointed out in his Comment with regard to my last Post in which I took issue with the Telegraph's wild over-estimate for 3 babies sharing the same birth date, the odds change greatly if you don't specify a particular couple and a particular birth date. The odds for a particular couple having 3 babies all born on the same date may be around 133,000 to one, but if you have 133,000 couples each having 3 babies, then you are very likely indeed to have one couple with 3 babies sharing a same birth date. Doug points out in his Comment that you only need 23 people in a room for there to be a more than 50% chance that two of them will share the same birth date. If you have 60 people in a room it is virtually certain that two of them will share the same birth date.

The Sun article illuminated this problem, and I now see why the Telegraph 'expert' claimed odds of 48 million to one, because he assumed they had chosen one particlar couple and one particular birth date, when the odds are indeed 48 million to one. And though the Sun figure is almost correct for a particular couple starting their family on 08.08.08, they haven't considered the possibility of a sequence of 3 babies starting on 03.03.03, 04.04.04, 05.05.05, and so on.

You can see why people talk about lies, damned lies and statistics.

I'm much more exercised by today's poser, which is to calculate the odds of babies being born at a particular time on a particular date so that they have a sequence of the same number. Supposing a child had been born at 20.02hr. on 20.02 in 2002. A woman lucky enough to have already had a baby at that exact time and date might like to try for her next to be born at 21.12hr. on 21.12.2112. Plenty of time to practice.

That should get the journalists' "expert statisticians" going. She'll make her fortune, if she can do it.


  1. Interesting tidbit to know but I have got all muddled up with the stats.

  2. Mystica- Don't worry, everybody does. It's only if you get your dates muddled that you need to worry.

  3. I *know* that this was in English, but anything to do with figures just leaves me floundering in the deep end.

    Other than the figures in swimsuits, I get them ok ;-)

    Ali x