My Hasse diagram done better

[simont]

Almost four years ago I had a silly idea, and used Graphviz to generate a Hasse diagram of the 2008 Olympic medal table, under the partial order with which any sensible ranking of medal counts (irrespective of the country producing them) must agree no matter what relative importance it assigns to gold, silver and bronze.

I had an email from the New York Times a few weeks ago, saying that they were planning to pick up my idea and do the same thing this year. All they wanted to know was how to credit me accurately, but I mentioned to several friends a sneaking suspicion – based on previous interactions with journalists interested in my maths-and-computers frivolities – that round about now I might find then coming back and asking for a little help with the coding.

Apparently I greatly underestimated them. I've just checked my web logs and found this page: http://www.nytimes.com/interactive/2012/08/07/sports/olympics/the-best-and-worst-countries-in-the-medal-count.html. Not only did they have no trouble at all repeating my analysis, but they've managed to automate it on a web page so conveniently that they're able to drop in new data sets effortlessly – and so they're able to put it up before the Olympics are over (and presumably, therefore, it keeps running track as more medals are won), and have it also show the Beijing results and have a button to adjust for population. And their version is prettier and has mouseovers. Nice work, NYT!

Comments

[simont]

and presumably, therefore, it keeps running track as more medals are won

Addendum: we can drop 'presumably', since I've now seen it do so.

[lnr]

Thanks for linking to it too, it's fab, and great to see not only a cool idea being borrowed with credit but being done well.

[simont]

It's a rare pleasure for me, actually, to see an idea of mine getting this sort of treatment. Mostly I think of myself as an engineer rather than an inventor – it's very common for me to take someone else's idea and apply solid engineering to turn it into a polished implementation (e.g. almost all my puzzles are implementations of existing game ideas, PuTTY is an implementation of existing protocols and terminal standards, my day job is mostly implementation too), but far less common for me to have an interesting original idea myself. So to see someone else picking up one of my own ideas and applying the sort of engineering I'd have been proud of is a delight. I hope I've given the same pleasure to at least some of the people whose ideas I've implemented!

[gerald_duck]

On the other hand, it is also possible to invent algorithms and patterns. Tactics rather than strategy. And I know you do that.

[simont]

Occasionally, yes. I did say mostly.

[pseudomonas]

Nice :)

My brother's brother-in-law works on the NYT web team, from what he says they seem like quite hi-tech within the industry, quite interested in developing new technologies rather than just keeping the website ticking over, so their competence with this doesn't surprise me entirely.

[sunflowerinrain]

Very cool :)

[holdthesky]

Ooh, I do like non-parametric solutions to problems. All those constants are so arbitrary. The New York Times!

It's always nice to see American journalists treat people properly. It reminds is that in Britain the Mail would probably have stolen it, got the maths wrong, say the law doesn't apply to them and then threatened any complainant with a smear campaign of dubious veracity. Continental journalists can be quite respectable too, but in a different way. American print journalists always get the details right but sometimes paint the wrong picture, continental print journalists are better at big pictures, but a bit smudgy. British print journalists (and, it seems, the CPS) seem mainly to be wearing dirty macs and sniffing around the bins for discarded polaroids.

Interestingly, in the IOC's charter it is explicitly prohibited from producing medal tables by country as the Olympics "is" a competition between individual athletes (boggle).

Chapter 1, Section 6:
The Olympic Games are competitions between athletes in individual or team events and not between countries

Chapter 5, Section 58:
The IOC and the OCOG shall not draw up any global ranking per country. A roll of honour bearing the names of medal winners and those awarded diplomas in each event shall be established by the OCOG and the names of the medal winners shall be featured prominently and be on permanent display in the main stadium.

In light of this, I think we need to also consider a rejection of the validity of the addition operation (as each individual is a "different unit"). The only reasonable way to do that, I think, is to draw up a nullary medal table where all "moves" on medal counts are equally invalid (and so valid) and so every country is incomparable (and so comparable) to every other, (and so every medal scores zero).

So I think there's also a trivial reasonable medal table in which all countries are better than all others (including ones which they are worse than) or equivalently no country is comparable to any other (I suspect some countries are not comparable to themselves, but that's another problem!).

The New York Times could have a button where everything either collapses to a dot or else flies away to infinity.

:-)

[pne]

About the producing medal tables by country as the Olympics "is" a competition between individual athletes - I had heard that but hadn't seen "chapter and verse" (so to speak). Interesting; thanks!

Still, the diagram is nifty, as is the fact that the NYT picked it up and credited Simon!

[holdthesky]

Indeed, and clearly medal tables are an important aspect of the Olympics, despite the IOC's quixotic position, and the parametric kind of scoring which you see is very unsatisfactory. I really like identities, nulls, matrix kernels, and so on, and so was (perhaps a little too) enthusiastic when this one came to mind!

[cartesiandaemon.livejournal.com]

Huh. That's amazingly ethical and competent of them! I guess it's a shame I'm surprised a major newspaper is ethical and competent, but still, kudos to them. And to you for the idea.

[simont]

and presumably, therefore, it keeps running track as more medals are won

Addendum: we can drop 'presumably', since I've now seen it do so.

[gerald_duck]

I see Matthew Bloch's name on the article. Matt Bee was at King's College a decade or so ago, may be familiar to many people reading this and is known, by me at least, to be competent. (-8

[gerald_duck]

I note in passing that the NY Times logo for London 2012 is yet another example of something nicer than the official Simpsons-fellation one.

[1ngi.livejournal.com]

Cor!

[atreic.livejournal.com]

That's very nicely done.

The adjusting for population one is very funny though - I like the strategy of 'win one gold medal, have very few people' adopted by Grenada!

[simont]

Yes, that is unfortunately the problem with per-head or per-dollar-GDP systems; they seem a fairer way in principle to judge merit irrespective of size, but when countries get down to the point where the expected rate of medal-winning is a fraction of a medal per Olympics they develop the problem that the interesting parts of the table are composed of noise rather than signal...

[pseudomonas.livejournal.com]

Though Jamaica has done rather well even not taking population into account :)