- Thursday 30th November
Received my £10 cheque from Camelot, so I scanned it in and put in
on my
"what I've spent" page. Hey, a £10 cheque
from Camelot is a rarity because you can only get it if you manage a
3-match via a subscription entry.
- Wednesday 29th November
Decided that it was time to add the
last appearance page
to the InterLotto section.
- Tuesday 28th November
Back to work and time to upload the results for
Lottery #54. I now ignore spaces and
commas when people type in their positive integer answer in the virtual
lottery form. I also re-format the number in the returned page to include
comma separators for numbers >= 1,000.
- Sunday 26th November
Discovered that I'd finally won £10 after over a year of waiting -
ironic that I was out of the country and unaware of the win when it happened.
It was on my subscription of course, so I'll get a cheque at some point from
Camelot.
- Saturday 25th November
Switched to Amsterdam and blitzed the CD stores for bargains - how about the
"Woodstock 94" double CD for 20 guilders (£8.50) ?
Anyway, the best music chain to go to in Holland is undoubtedly
Free Record Shop, which
has a sale section so you don't have to pay the excessive 40+ guilder price
for an album purchase. Oh, I missed the lottery draw on TV (yes, Holland can
get BBC 1 and 2)...
- Wednesday 22nd November
I was in Delft for the next 3 days for an
HP-UX PD archive workshop meeting and,
yes, I had Internet access, but just used it to check my e-mail.
- Monday 20th November
Uploaded the results for
Lottery #53. Oops - I realised that the
answers to the virtual lottery were actually available as a text file that
was visible on (but obviously not linked to) the Connect WWW server ! This has
been fixed by coding the latest answer into the CGI binary and moving the
question/answer text file off the Connect WWW server altogether. Using the
answer from the file would have only given a marginal advantage anyway because
a) the answer is fairly easy to find via the link clue and b) the entrant still
has to match at least 4 "real" numbers as well to win.
- Saturday 18th November
Nice to see that UK teletext finally mentioned these pages after "only" a
year (probably because of the lottery anniversary), although it wasn't
the prominent ITV page 123, but actually Leslie Bunder's weekly Internet
feature on Channel 4 page 652. Meanwhile, ITV teletext page 326 had a
poll asking whether the lottery was "good or bad", but they had the wrong
number of jackpot millionaires again (they said it was 102, but it's actually
132 if you don't count
Lottery #53 since that hadn't taken place
yet).
Meanwhile, my mother won £10 for the first time with a 3-match for
Lottery #53 !
When I looked at her ticket before the draw, it noticed it had a 3-match
for the previous lottery (#52) and was sceptical that it would win two draws
in a row. How wrong can you be ?
- Friday 17th November
I decided that the "tickets that have won the most prizes" page wasn't really
very interesting (only listed one or two tickets usually), despite taking up
to 5 minutes to generate (and not scanning all the combinations either !),
so I've removed it.
To replace it, I cooked up some code to actually scan all 14 million combinations
and do a complete comparison, enabling me to draw up a performance table for each
"best match" category. Interesting to see that I use one of 5 million sets of
tickets that have never won and also that there's only 647 tickets left that
have failed to manage a 2-match. Anyway, all this info is in a new
"best performing tickets" page
(takes 15 minutes to generate on my slower home machine) and
there's an
InterLotto version too
(only 30 seconds at the moment).
- Thursday 16th November
Received a list from Camelot of
winning numbers,
individual prizes and
number of winners
in each category for all 52 draws. All my figures matched exactly, but I still
had to phone them up to confirm the jackpot prizes that would have been won on
each of the 9 rollovers so far (they'd put a figure of "0" in the list for all
unwon jackpots, whereas they can reveal how large the jackpot would have
been).
- Tuesday 14th November
Slightly adjusted the
calculations
for average number of right/wrong answers in the
virtual lottery because the first two virtual lotteries
shouldn't count (there wasn't a question set for those).
Monday's ITV teletext page 324 claimed that there's been "more than £1.53
billion paid out in
on-line prizes", but I make it a
fraction under £1.5 billion, so I phoned Camelot and asked them to
send me all 52 draws' ticket sales and total prize pool figures via snail-mail
(it would take too long on the phone and Camelot still doesn't have
external e-mail...groan !). I wanted to do this at some point anyway, just
to verify that my prize figures are correct and any differences with my figures
can be queried on the phone later of course.
Mind you, ITV teletext had ludicrously conflicting claims about jackpot winners
on its Monday pages. Firstly, we had the lottery page (123) saying that there
were 283 jackpot winners and 124 millionaires. Then we had a factfile page
(324) saying there were 633 (!!) jackpot winners and 102 millionaires.
However, I disagree with both and reckon it's 291 and 132 respectively. Now
page 123's figures are exactly 8 less than they should be for both figures,
indicating they've taken old figures from after lottery #50 (since there were
8 millionaire jackpot winners in total for lotteries #51 and #52). However,
page 324's figures are just nonsense [taking scratchcards into account as
well may explain the 633 figure, but definitely doesn't account for the 102
figure]. Now you see why I run these lottery pages - at least I make
the effort of checking my figures, unlike the lazy swines that make up the
UK media...
- Monday 13th November
Uploaded the results for
Lottery #52.
Added the number of entrants count to the
home page of the virtual lottery and also moved things
around slightly (and fixed a grammar error that had been there for weeks !).
Added a couple "Quick Pick" buttons to the
virtual lottery entry form - one for totally random numbers and another for
my RKL random numbers. These changes
convinced me that each part of the form needed to be highlighted as a
separate section, so I've duly split it into 4 parts. Created an
answers page so that users could see if
they got the question right. Updated the
lottery calendar thanks to some diary info
that cropped up ITV teletext page 322 on Sunday.
- Monday 6th November
Uploaded the results for
Lottery #51. Activated the extra question
people have to answer in the
virtual lottery. Someone
actually won virtual lottery #2 and netted themselves a cool £100
today (no, not me - I'm barred from entering of course !).
With 390 entrants and a 1,000 to 1 chance of winning, I wasn't expecting a
winner...
- Saturday 4th November
Ho hum, we'd moved mail-handing from one fileserver to another this week and
I'd adjusted the code to filter incoming teletext (or InterLotto) e-mail on
the new server, but then duly forgot to install it ! As a result, neither the
UK lottery nor the Interlotto pages were updated. The new code has now
been installed and it works too. Needless to say, I got a couple of e-mails
from people wondering where the updates were (the Net never sleeps at
weekends).
- Thursday 2nd November
Moved the "Next month" link near to the top of each of these What's New pages,
which is a much more logical place since the text is in reverse date order.
Removed virtual lottery C code from InterLotto CGI and page generator programs,
since there's no virtual lottery run by InterLotto of course (it's a real
one !). The code was never actually called naturally enough, but it had still
been wastefully included in the InterLotto binaries until now.
The
virtual lottery was always intended as a "magnet" to
bring people to this WWW server, but I suspect many people just save it as a
bookmark, visit it once a week and don't go anywhere else on the server.
That's a real shame because there's a huge amount of info here (OK, so it's
almost all about Merseyside, but that's our remit !). I did have an idea
for a "treasure hunt" several months ago, where a set of questions pointed to
various parts of this server where the answers could be found and entered into
a form. However, with the virtual lottery proving fairly popular, it makes
more sense to integrate it into that and effectively add a bit of "skill"
by insisting that a question relating to information on this server is
answered to retain the chance of a winning a prize.
It may put people off having to do a bit of surfing to find the answer (no,
any link "clues" in the question won't link directly to the page with the
answer on !), but I think it justifies the existence of the virtual lottery
as part of the Connect server - it's effectively a promotional tool for a
differing part of our server each week and should suck in people who normally
wouldn't ever browse the rest of our server. Anyway, all of this will start
next Monday with virtual lottery #3 and it'll be interesting to see if there's
a drop-off in the number of entries (if there's a rollover to bring the prize
to £150, I doubt it !). I have adjusted the virtual lottery
home page,
rules and
FAQ to mention this impending change to the
playing rules.
- Wednesday 1st November
To workaround what looks suspiciously like a bug in the Netscape Commerce
Server, I've now set an alarm signal for one hour, causing an exit of the
CGI programs ("say", "lottery" and "InterLotto") if the alarm goes off.
It appears that
the Netscape Commerce Server doesn't properly signal a "slow" CGI program
(i.e. one that has a lot of output or takes a significant time to run) to quit
if the user closes the connection (e.g. via "Stop" on their browser or simply
exiting their browser). This lead to several hung processes (not zombies -
just procs waiting to output to a non-existent parent), but we didn't get
this problem with the NCSA server. Hopefully, this alarm signal kludge should
fix this (it did...).