pumpkinstrokemarrow

He then went on to demonstrate an algorithm for constructing magic squares, said to be of his own invention and to work for any positive odd number N. It started by putting a 1 in the top middle cell. There was then a small number of rules about what to do next, the important one being that if the cell above and to the right was available, put the next number in that. It worked OK with N=3 and N=5. I have not checked the illustration above with N=7. But my first thought, a few Browns down, was that this was silly. Why should such an algorithm for moving around the square deliver such a result?

Back home, I could have written some Basic code to do the job in Excel - populating the square and then checking the result. Try it for a few large numbers and if that worked the algorithm would be sound. Not a mathematical proof but good enough for ordinary mortals. I, however, wanted to play the extraordinary mortal and wanted to prove it proper and I think I am starting to see how one might approach the problem. But some way off a working proof.

Then, more prosaically, we have been pondering here in Epsom about the mysteries of AV, something which I am vaguely for, thinking it likely that it will make for a modest increase in the number of Liberal MPs. Not that is something I am particularly keen on, but I do think that it would be proper. While we pondered, up pops Andrew Gilligan, who after rocking the ship of state with lurid revelations about the invention of weapons of mass destruction in Iraq, went on to work for that bastion of Middle England, the DT. Google suggests that he is rather a strange bird; perhaps you have to be to be an investigating journalist. Yesterday though he seemed entirely benign with a fairly banal half page article about AV, mainly revolving around the antics of the Australians who already use such a system. He comes to the fairly comfortable conclusion that AV is entirely practical - in the sense that one can still count the votes - but will not make a huge amount of difference. But there was no analysis, just anecdote.

Now I would think that in the space of a day or so one could download the results of recent general elections in some suitable form and run various simulations of AV on them - which might well give one a more informed idea about what AV might do - this in a land where we do have a significant middle party, unlike the Australians. How many constituencies are there out there with a 6-4-2 split of the vote, with the dropped 2's giving the 4's the victory by a short head? (Which I think would be a reasonable outcome). Is there a hack out there with access to a research assistant with basic skills? Is the standard of debate going to rise in the interval between now and the referendum?

Will the reds and the blues move off their rather simplistic vision of AV being the end of parliamentary democracy as we know it - just because they are apt to loose a few seats and so be forced into showing the Liberals a bit of respect. Will they make any attempt at argument on principle?