Next month: April 2013
Jackpot for this draw £853,300,000
Now I'll be the first to admit I've made that additional zero typo myself a few times in the past, but it's fairly rare for this to happen on the official site. Needless to say, my scraper code used this for the UK value in sterling and messed that up on my pages (along with the currency exchange rate as you'd expect).
Luckily, both myself and Camelot were quick to spot the extra zero and we'd corrected the figure within 5 minutes or so. To prevent it happening again, I now divide any UK sterling jackpot grabbed from Camelot's site by 10 first if it is >= £200m. I've also fixed a minor Super Draw indicator issue that didn't manifest itself on the site but malformatted the next draw's line in my Euro Millions data file.
I have adjusted the code so that the home page will not announce the individual jackpot prize or the number of winners (and similarly, the individual draw page prize table will not be shown) until the Irish site reports on country split. Since the Irish site almost always announces this before the UK site announces anything at all, this shouldn't hold up the display of UK results when they finally turn up.
I'll deal with the spreadsheets shortly (will be converted to CSVs via the handy LibreOffice 4 on Linux and then processed with a script to compare them against my data files and flag up differences), but first up is the full e-mail discussion presented as a new Q And A With Camelot page.
I've taken a look at the first Excel spreadsheet Camelot sent me - the one for the Euro Millions draws. Predictably, it's a massive disappointment for several reasons. The biggest issue I have is that it's manually edited by a human who doesn't always keep every line in the same format (not good for a spreadhseet!) and is clearly not doing a cut and paste from the latest draw data to avoid mistakes.
To be frank, it's quite shocking they don't generate the spreadsheet (or preferably a CSV, which is much more processable outside of Excel) from the same database that drives the official site's draw pages.
Another annoyance is that the Millionaire Raffle numbers have been placed in a completely separate spreadsheet file when they clearly belong with the main Euro Millions data. The killer irritation though is that they've completely omitted the number of European winners in each tier from the spreadsheet, despite actually having that info on the official Web site (and in a bizarre reversal, the spreadsheet does have the number of UK winners, unlike the Web site).
After converting it to a CSV, I scripted something to produce a space-separated file that could be compared with a similar file I produced by running some awk massaging of my draw data file - though the spreadsheet only went to draw 567 and you have to adjust the number fields from draw 379 onwards (the game changed its tiers then).
Once I did that (and the draws were sorted in reverse draw number order), I did a "diff" which compared the winning numbers (ascending and drawn order), the number of UK winners and the UK prize amounts. It found 51 out of the 567 draw lines differed between my data file and the spreadsheet! Some of the diffs were less than 180 days ago, so I went to the official Web site to check them and my data matched the official site, so the spreadsheet had human typos in it, basically making it completely untrustworthy!
Typical differences included single digit typos, missing off the decimal point from a UK prize (so it ends up being 100 times bigger!) and not putting a .00 suffix on some UK prize values. The most recent example you can still see on both sites is draw #533. My site and the official site both agree that the 5-match UK prize is £43,732.70. However, the spreadsheet sent to me by Camelot says it's £43,732.00! Yes, a small difference, but even one typo is enough to destroy all confidence in the data.
One particularly nasty example of why it's bad for a human to type numbers whilst reading them manually from a computerised source is that they managed to screw up the number of UK winners data from draw #522 completely in the spreadsheet! Instead of typing the number of UK winners in each cell (0, 1, 2, 12 etc.), they typed the number of European winners instead (0, 5, 7, 45 etc.) - a seriously poor mistake that they've repeated on many of the earlier draws too (more than 180 days old, so I can't show you the proof on the official site).
Previous month: February 2013