WMF, WTF?

[peterbirks]

Cannny observers of this sequential drivel might notice that the web adddress has changed. Live Journal restructured all its users addresses to http/username/liverjournal.com. Unfortunately, this doesn't work for usernames that began or ended with an underscore. They offered a free name change, so I took it. The old favourite will still redirect, but if you want to save a nanosecond, use the new address.

All of which appears to part of the new year's security flaws. LiveJournal (sort of) explained the reason for the change, but I didn't really understand it and I had no real desire so to do. I'm not that geeky. But security flaws seem to be the hip issue of the day. I went to the Gutshot site yesterday morning and it was "down for maintenance". Last night I went to the forum and my machine at home had a fit. It claimed to have found half a dozen Windows Meta Files (the files through which trojans are stored in images) on the Gutshot forum site. It casually deleted them and steadfastly refused to display the forum at all. An event which, I fear caused me no great sense of loss.

+++

I see that the Euro Lottery has a nine-week rollover. I've not paid much attention to the mathematics of this, so forgive me if my numbers are wrong — the luxury of time for research is denied me just at the moment. But it seems as if you are getting 85m-to-1 on a 76m-to-1 shot. So, with such massive positive expected value, clearly we will all be putting all our wealth into the Euro lottery this week.

Well, no, we won't. Which is a good example of the marginal value of money and the significance of big numbers. Even with positive EV, the chance of the big win is minuscule, so you are hardly likely to place your entire wealth on a single roll of the dice. Bankroll management, you see.

So, what would be a sensible proportion of your wealth to put on the Euro lottery this week? Well, that depends on your marginal valuation of money, I guess. 1%? 5%? 10%? I have no easy answers here. For me the answer is 0%, simply because I can't be bothered. But it would be nice to know if there had been any rational research on this, rather than the cobblers spouted by most people on poker forums.

+++++

I went to the police station yesterday to give my witness statement and to sign in as potential evidence the mobile phone that I found in my non-stolen car (see Sunday blog). I typed out the statement beforehand but, of course, bureaucracy requires that a policeman then write out my typed statement by hand so that I can sign it. That's the way they do things, you see. Modern society? Impenetrable, mate.

How long before I can get out of here?

Comments

[geoffchall.livejournal.com]

I'm afraid I still don't understand the maths of the lottery. I thought the think behind rollovers is that the inflow of money chasing an illusory prize was so much bigger that you still didn't get positive odds, although I presume there is a limit. How about the concept that the whole thing sucks and you should have nothing to do with it.

Apart from perniciously encouraging gambling, it's a regressive tax on those on lower incomes, enables the government to apparently get poorer people to subsidise the NHS and it's a Crap Bet.

[jellymillion.livejournal.com]

If one were to make the stipulation that only one ticket would win the Euro-lottery, then one could simply buy 76m tickets (I'm assuming your numbers are right, being far too lazy to check) and take EUR9m home on Friday. Or whenever.

I think the Kelly Criterion ought to come into play here, somehow, but I can't quite think how. I probably need more coffee.

Buying 76m tickets

[peterbirks.livejournal.com]

Interestingly, this play has been attempted in the past in the US, but that was in the days of a mere 47 numbers (10.7m combinations of six, if my memory serves me correctly). As you say, there is the risk that the winning combination will also be chosen by a number of other players. On the flip side, you are also guaranteed all the consolation bonuses, which pushes up the overlay considerably.

However, the logistics for even 10.7m combinations were horrific. The entry cards were printed out on a computer by the Wednesday or thereabouts, distributed nationwide to a pre-employed staff of thousands just in time, all of whom overwhelmed their local newsagents with about $10,000-worth of entries. Clearly some of the people failed to put on their bets, which left the syndicate petrified that one of these might be the winning combination. The syndicate made money, I think, but it wasn't a massive return on equity, given the work involved.

Kelly Criterion? I'm sure I have an album of hers somewhere.

Re: Buying 76m tickets

[(Anonymous)]

Pete, I remember the TV programme about that lottery, the syndicate only managed a 80-85% coverage. A lot of lotteries subsequently banned the computer generated multiple entries. The trick in the lotteries when there is a multi-rollover is knowing which number combinations aren't chosen at all and selecting all of those!
There are combinations 1-2-3-4-5-6 for example that can have over 1000 ticktes, imagine how peeved the jackpot winners would be getting a few thousand each, lol.

JG

Re: Buying 76m tickets

[peterbirks.livejournal.com]

In an old Greatest Hits I did a statistical analysis of what combinations of numbers had the best chance of being "single winner". I'll see if I can dig it out. What I did conclude was that it was not so much a matter of the numbers as their position on the entry form. If all the numbers were in the far left and far right rows, for example, then you had a good chance of a rollover.

The introduction of the random number ticket obviously smoothed out, but did not eliminate, this anomaly.

PJ