What and when and how often, revisited

Five years and one day ago I put up a table on the stats page to show how often and when, on average, items are found in the advent calendars that have been used for the competition. It was rudimentary and I always planned to upgrade it, and I’ve started a few times but always come unstuck in forming the database query that would drive it.

Until today, when I started from scratch, ran into the same problem again, googled it, found that I was on paper doing it right, looked up the syntax for the bit of the query that was clearly not working and tried something didn’t look like it would work but somehow did. So much, if not all, of this website was coded by trial and error!

Anyway, now the table shows not just how often each item has occurred, but how often on each individual day. On top I’ve layered mean and mode because I know how to calculate those. The table is not in the final form I envisaged – for that I’ll either use a scalable font size or a colour gradient on the numbers so higher ones stand out more, but that is for another day.

I also need to scan through the database and rationalise the names I’ve used for items, as, for example, Christmas Pudding appears twice as it must have an errant space or something similar to create a variety of spellings for the same item. I’ll try to get to that tomorrow!

Mild embarrassment

It’s fun to note that for a year I decreed that I want to try and have a bumper field, we are looking at having our lowest field since 2007 – admittedly I haven’t gone as gung-ho on promotion as I could have, because I’ve never really felt comfortable with the idea of it. Quality, not quantity, anyway!

Related to this I’ve fixed an element of the website that wasn’t working exactly as designed – the query that pulled the data from the database for the “number of new/returning players” bit of the stats page only actually worked if there was a new player registered for a given year – there are no new players this year so the table/graph didn’t recognise the existence of 2022. I have fixed this and all is now displaying as desired!

Tomorrow, if I have time, I might see if I can do my intended redesign of the “Item Stats” data on the same page – if not it will hopefully happen soon. The table as it stands is pretty naff!

Good luck to our 31 players, and it’s not too late to join yet if you haven’t!

Almost there!

I’ve had a bit of a nightmare sourcing a calendar this year but we have got there, albeit with a calendar that I might have rejected if I had managed to get a few good ones in front of me. We’ll see what happens! The list is not updated on the website, but as a reward for reading this post, here’s what’s in the calendar this year:

  • 1 week to go
  • 2 weeks to go
  • 3 weeks to go
  • bike
  • camper van
  • candy cane
  • children
  • Christmas tree
  • cow
  • donkey
  • fairy lights
  • gift
  • holly
  • mouse
  • penguin
  • pheasant
  • postbox
  • rabbit
  • reindeer
  • robin
  • Santa
  • Santa hat
  • sleigh
  • snowglobe
  • snowman
  • tinsel
  • toy car
  • wellies
  • Xmas pudding

I hope to get the website up and running and login details sent out tomorrow. Good luck everyone!

The 20th Challenge

It’s nuts to think that this year, if I bring it back, will be the 20th running of this weird little pre-Christmas competition.

I’m definitely bringing it back, by the way!

For some it’s a thing that they enter just out of a strange sense of loyalty, for others it’s full on part of their festive tradition, and there’s no doubt the full spectrum in-between. We’ve had 125 different people play it over the years. Only two players have played in every edition, and there’s a whole load of people for whom once was enough. It’d be nice if we could mark the milestone by having as many players return as possible, but, well. It’s tricky.

The easiest way to try and get a crazy huge field of people would be to email everyone who has ever played and invite them back, explain my thinking, etc etc, but I am very keen to not do this. Worried about bombarding people with emails about a thing they most likely don’t care about, it has always been my policy to email player details only to those people who actively played (that is, registered a guess) the previous year. There’s no mailing list, as such, but if you can’t be bothered with it just ignore the email one year and you won’t hear from the challenge again, unless you pro-actively email in yourself.

I cannot bring myself to change that, so this blog post is going to have to be the way. I’ll lazily tweet and FB this post a few times in the next four weeks up to starting day. Pass it on, if you know someone who has previously played but DIDN’T LAST YEAR give them a gentle poke so it’s not me who’s bothering them.

In summary:

This year is the 20th running of the Advent Calendar Challenge, and to mark this crazy occasion I’m looking to welcome as many returning players (and indeed new ones) as possible. If you have played in the past but didn’t play last year, get in touch and I’ll get you entered for this year. Thanks all, and I look forward to hearing from you!

If you haven’t played before, and I know you, also get in touch. New players traditionally do very well for some reason, so this could be your best shot at winning! 🙂

Let’s see how unsuccessful this campaign is…

2021 review and some rambling

Huge congratulations to Douglas Cameron who took the victory in this year’s challenge! He made a rather beneficial guess change with a couple of days to go and it was obvious that he would probably win from there. He has already been in touch to say he is looking forward to trying to defend his title next year, and I am looking forward to welcoming you all back then!

In a tired, middle of the night awakening (I had a horrible cough, but don’t worry – I’ve been tested and it’s not that) my mind wandered into the scoring. I still think it’s a bit of a nonsense that the bigger point reward comes when it’s easier to get the answer right, but I’ve not been brave enough to make the change. I debated with myself whether or not reverting it a bit more slyly to how it used to be (i.e. 24 points on day 1, 23 on 2 etc) would be a bit more palatable to the players and intuitive to follow. The way I came up with would see everyone begin on 300 points and be deducted the number of points equivalent to that day’s date if they failed to get the answer correct. By the end it would all shake out to reward earlier correct guesses more than later ones. In all honesty I think I’m very unlikely to bring this in but I was a fun 3am thought!

I also debated aloud with Mrs S how it was cool that on the final day there were three objects that you could have scored with this year, and maybe to throw an extra level of intrigue (and maybe a cat amongst the pigeons) I should’ve decreed that scoring would be weighted in favour of the less popular guess.

34 people scored on day 24. Of those, 0 had gift so we’ll discard that from our thinking. The split amongst the other two answers was 21 Christmas Trees and 13 Santas – so I could’ve said 13/34 x 24 = 9.1 (we’ll call it 10) points for guessing Christmas Tree and 21/34 x 24 = 14.8 (we’ll call that 15) points for going with Santa. This would’ve actually seen Ed Haslam win as the ten point swing would’ve been good enough, but we’d also have to look back at the igloo/penguin day which I think Ed scored on so nothing is certain.

Or is reducing the value of the day wrong? I guess it is. So maybe we say the least guessed scores the full 24 and the more popular item gets a reduction equivalent to the gap, so (34-8)/34 * 24 = 18.3 (we’ll say 19). That’s now only a 5 point swing, and DC stays champion. This is the kind of shithousery that F1 would be proud of, and I’m happy to say it was never a serious option!

I wish everyone a very happy and healthy Christmas, a 2022 full of everything you want, and I’ll see you next year for more silly advent calendar fun!

2021 updates

This year I made it so that when you guess, it no longer reloads the same page but a specific page that simply says whether or not your update was successful and then offers a link back to the homepage so that a player can then see their guess in the table (as I assumed they’d want to), but I’m still getting the behaviour where a player is saying yes to reposting their form data and as a result the same guess is being entered multiple times in quick succession. A good design will stop this happening and I thought the bodge job that I’d done would be enough – after all, why would you want to refresh the page that only says your guess change was successful!? Anyway another rethink is required; I’ll have to do something clever like throw out an error if you are trying to change your guess to what it already is…

I haven’t had the time to dedicate to tweaking the website as much as I’d want but I’m well on the way to eliminating the ‘table-headers-at-the-bottom’ lunacy that was a direct result of imitating the design of the F1 digital graphics that I have mentioned previously. Hopefully that process is one I will complete in the next few days and then I’ll be able to move on to a better version of the which-items-have-appeared-on-what-days table that I’ve been planning for a while. I started this last week and found that it’s not as simple as I thought it would be so I’m scratching my head a bit, but with a bit of time I should be able to figure it out!

A player has entered this year with not their real name and this is as good a moment as any to remind players that if they want something similar then they can get in touch and their display name will be amended accordingly. I’m reluctant to give free choice (I know enough people that will choose either something rude or something akin to Boaty McBoatface to rule that option out) but I’m sure between us we can find something agreeable. It’s actually got me thinking about whether the default of full name is actually needed – if, for example, the all time top three were listed as Simon R, Mark C and Antony B I don’t think the competition would lose anything. I’m going to ponder this one and report back, and thoughts are welcome as ever!

A couple of players have pointed out on social media that there are twenty-five items to choose from and they are correct! This means that at least one day has more than one item behind its window, and I don’t think I’m giving too much away when I state this.

I got the Facebook and Twitter links working again and reverted them back to the “post everything” mode that they existed in when they were first created. In the interim the rate limiting of the third party services that I used meant that I had they had been restricted to only updates about window opening and points scoring, and after a couple of days it became clear to me that that had been a good thing so I have deliberately limited it myself now. Please continue to subscribe on Twitter and Facebook to get the important updates placed directly into your feeds (where their algorithms think best)!

Not long left to get your first guesses in on time, and remember a late start carries a penalty so don’t be late! Good luck everyone 🙂

2021 Competition – coming soon!

We are a couple of weeks away from the competition getting underway for its 19th incarnation so I’ve logged on and started to make the preparations as necessary. This starts with the mundane (e.g. renaming the Facebook and Twitter pages from 2020 to 2021) and progresses to mild code changes with plenty of stuff inbetween.

Logging on here to write this post caused lots of problems with my browser trying to redirect to a secure (https) connection that didn’t exist and whilst this might (or might not) be my local settings being a bit aggressive, if it’s happening to me, it’ll probably happen to others too. I didn’t want this so have acquired a certificate and it’s associated padlock and so the browser errors won’t now happen – but I’ll have to test the website still works fully before the launch so add that to the list of jobs to do in the next 10 or so days!

That’s just one of a few fiddly jobs that will culminate with me taking a couple of hours to manually email out everyone’s login details for this year, so whilst I’m getting on with these tasks over the coming days here is a peak at the items that are found in this year’s calendar, in alphabetical order of course. Get strategising now!

angel
bauble
candy cane
cracker
deer
doll
elf
gift
holly
igloo
lantern
mince pie
moon
penguin
pudding
santa
skis
snowflake
snowman
spinning top
star
stocking
teddy
tree
wreath

The race to 1000

Couple of housekeeping measures before I begin – late last night I made a change that has no bearing over the actual results or standings of the competition but that I wanted to make for a while – where everything else is equal (score, number of correct guesses, highest scoring best guess etc) the current standings table now ranks players in the order they joined the competition this year. Previously it used a random and mildly nerdy method that is far to silly to explain. I’m happier now. I’ll reiterate that if all the other things are equal then the players are fully and unbreakably tied, and this is shown as such in the table!

If you head over to the stats page then you’ll see the all-time scores table in all it’s massive glory. Atop it sits Simon Rutter, who has been at the top of that table for fifteen straight years now. He is on 880, with Mark Coughlan 92 points behind in second (where he has sat for fourteen years himself), Antony Brown a further 47 back in third (no mean feat for a player who started four years after the others at the top) and Ross Turnbull is fourth, 18 points adrift of the top three. I highlight these players because all four could, in theory, hit a thousand if they play well this year!

Admittedly for Mark, Antony and Ross they’d require suspiciously high scores (212, 259 and 277 out of the maximum 300 respectively) but Simon is definitely within striking distance – 120 away is doable, last year’s victor scored 135 and Simon himself has a PB of 88 which isn’t too far away. I’m excited – this milestone is one that was never contemplated when this all started. I suppose, realistically, I shouldn’t get too excited. Simon averages 55 points per appearance so it’s statistically unlikely this year, or even next. But with a good performance, it could be this year. I’d love it to be. I might even make a first player to 1000 certificate (remember when I used to make certificates? I should start doing that again) to commemorate the occasion.

Please Simon, do well. It’d be epic!

Tomorrow

It all kicks off tomorrow, so any players who still haven’t submitted their first guess should get it in quickly!

On that… Another feature of the old 24-23-…-2-1 scoring system was that there was an incentive to play from the start. Flipping it had the unwanted side effect of allowing someone to drop in with only five days to go and win it. This seemed unfair on those who had been in from the start and for me detracted from the feel of the competition a bit. The list based guess format helped a bit but I haven’t been happy with the situation since it began. For one season only players starting late had their points scoring offset by their delay – so a player that started on day 10 and then scored on day 16 would receive only six points rather than sixteen, for example. This ended up being quite confusing (as I remember it on one day three people with three different starts scored) and so was scrapped, possibly in the middle of the same season.

For this year, should it be needed, and again on a trial basis, players starting late are deducted the relevant number of points from their total – so the player who joins after window 10 has been opened will start on -10. I like, and want players to start at the start, and hopefully this change should encourage that!

The rules and website have been updated accordingly. Good luck everyone, and get your guesses in!

Teams Scoring Changes

Before I get into what has changed about the scoring in the team’s championship, let me confess that there was very very nearly a change in the individual scoring too. I have long been of the opinion that scoring 24 points on the final day when there is only one item left to choose from is just silly. Scoring actually used to be the other way around, the (to me) sensible way around of 24 points on Day 1 all the way down to 1 point on Christmas Eve. It was changed after significant peer pressure because in the few years it ran that way, the title was wrapped up early and the rest of the competition had the “dead rubber” feel you sometimes get in proper sports when this happens.

Thing is, it has never sat well with me. I’m of the opinion that the harder to achieve achievement (that is guessing correctly on a 1 in 24 chance rather than a 1 in 1) should be rewarded more. I will do a further post about all this another time, but don’t be surprised if I flip it back to how I feel it should be one of these years. Your thoughts, players, are welcome as always!

So onto the team’s championship, where this year (2020) Leo are looking to defend their title and make it six wins in eight years. Credit to the Leonians for dominating so much, they have some heavyweights on their team and have clearly earned their results. It’s a sad side effect of the changes I have made then that they’d have actually finished fourth last year under the new scoring system. My beef with the old system is that we had uneven teams. Leo’s success is accreditable to that – they have had the most players at least half the time in their dominant spell, or one short of most the other times. 2013 was the last time they were significantly short – they won that year despite being three players lighter than their closest rival.

The dilemma of sorting this seems easily fixable by saying only the top x players for each side score, but I didn’t like that so much as that’s not really a team’s championship towards which every player contributes. Instead I’ve come up with a slightly contrived system and it works as follows. For each team, their best player contributes their whole score, their second best player contributes half their score, their third best contributes a third of their score, and so on. The teams with the best players still get the big points that they’ve earned, but for teams with lots of additional players these additional players contribute less and less – but crucially they still contribute. Fractions are always rounded up so every player who scores always contributes at least one point.

There are flaws with this, I’m aware, but I like how it creates a competition that still rewards teams with big scorers, I like how everyone is still in play for their team no matter how low their score (unless, you know, it’s zero), and I like how as teams get larger their capacity to run away with it is limited.

I will, given time, go through the team’s championship for every year and see how it would have changed under this system but for now here is what last year would have looked like:

Leo’s 2 point victory over Taurus becomes a 1 point victory for Aries. Aries had only four players compared to Leo’s seven and Taurus’s six, but with three of those in the top eight this feels like Aries have been rewarded properly for having strong players rather than many players. Over half of Leo players finished in the bottom half but the combined points haul of these “extra” players were what won it for them. Alas, what the new system cannot do is give a leg up to teams with only one or two players – Capricorn, Sagittarius and Scorpio had only single players last year and they remain down the bottom – but then if you only bring one player to a team championship I can deal with the resultant struggle. I think the new system strikes a decent, if imperfect balance. It’s worth pointing out that had Hakan, our champion last year, been representing a team all on his own that team would have finished in the top half!

As usual with such a fundamental change, we can consider this one “on trial” for this year. I’ll see how it pans out and try a different tweak next year, or even just revert it, if that seems best. But for this year, the new system is in. Good luck everyone, especially Gemini!

Recent Posts

Categories

GiottoPress by Enrique Chavez