Weekend entries for the SNITCH

Author   starstruck_auPosted on Categories Announcement

Hi all, a quick note to say that I’m now recording weekend entries on both the main (15×15) SNITCH and the Quick SNITCH.

For the Quick, the Saturday puzzles appear on the day as normal. They track the solving times during the day and update the SNITCH value as for the weekday puzzles. Although the Saturday data is not used for caculation of month averages, any during-the-day results that make it into the top 100 are recorded, as are any times for club solvers that are declared in the blog.

A reminder for everyone that I’m very happy to register new Blog Solvers, i.e. people who regularly report their times in the TfTT blog and would like them tracked on the SNITCH site. But this is an opt-in arrangement, so please let me know you’d like to be included. (You can do that by responding to this blog post :-)).

For the main SNITCH, I now record the difficulty of the Saturday and Sunday puzzles when the results are posted one week after the crossword. I only use the entries that make it into the top 100 and, given the delay, there are a lot more neutrinos to ignore in the data. So the reference solver list is smaller than normal. However, I’m hoping that it will provide some guide to solvers and bloggers on the difficulty of these puzzles.

As for the Quick SNITCH, none of the weekend results are used to calculate monthly averages for either the solvers or the crosswords in general.

Thanks for your ongoing interest. Constructive and curious comments and ideas are always welcome, even if I might take a while to respond or act on them.

27 comments on “Weekend entries for the SNITCH”

  1. Much appreciated Starstruck!
    Looking at the SNITCH is the cherry on the blog. More coffee coming your way 🙂

  2. I’ve been hoping it would be possible to produce a regular weekend Snitch rating, so this is a very welcome development. I’d like to be considered to be included for the snitch ratings although I’m not sure if I qualify. I have successfully completed about two thirds of this year’s Quick Cryptics but only 11 of the the year’s 15 x 15s (plus one Sunday Times Cryptic). I realise that I’ll have to post on TfTT on a much more regular basis than I currently do if I wish to be included.

    1. Well done on your solving. All of us started somewhere and getting a steady stream of completed crosswords is a great achievement.

      I’ll make you a Tracked Solver for the Quick SNITCH; you haven’t completed enough crosswords for me to pick you up automatically yet, but I can calculate an average and start tracking. If you report your times on the blog each time you complete the Quick, the SNITCH will then pick these times up even if you don’t make it into the top 100. It will then track your average time and how you went against the SNITCH for the puzzle. (It might take me a day or two, as I’ll need to do it on my PC.)

      A word of warning – you’ll need to be logged into the TfTT site with your blog name RPE10538 (no space) for me to match up your blog name with your Crossword Club identity. (Some people have commented without logging in and used variations of their blog name – spaces or not, capitals or not – and the SNITCH is not smart enough to track all these.)

      Happy solving and I look forward to seeing your progress on the 15×15.

      1. Thanks, I didn’t realise I had accidentally inserted a space, this has now been edited out.

  3. I only have a very peripheral interest in all this since I don’t like to solve online and I don’t worry too much about completion times.
    Having said that, still I am in awe of this wonderful website and all the data it encompasses. Starstruck is amazing; and it surely can only be a matter of time before he invents faster than light travel or time travel … watch this space?

    1. Thanks, Jerry, for the kind comments even though it’s only of peripheral interest. Having just done the long haul flight from Australia to London, some sort of easy time travel or faster-than-plane travel would definitely be of value!

  4. I’d be happy to be a tracked solver. I solve on the app, so have to manually report my times here, and don’t usually report times for the QC or prize puzzles.

    1. Thanks, Amoeba. I’ve added you as a Blog Solver.

      BTW, do you log in to the Times for The Times site? Or do you just add your name each time?

      1. Hi, starstruck. With reference to the incident yesterday when it was reported a.m. that the snitch hadn’t picked up the QC puzzle, in my role as TfTT admin I noticed that the puzzle category had been set by the blogger to Daily Cryptic instead of Quick Cryptic. After I had amended that, the QC snitch appeared. Could you please confirm whether the puzzle category is a critical factor, or was this just a coincidence? If it is part of the process, does changing the category immediately bring everything up to date as if it had been set correctly from the start, or does some early data get omitted? Regards.

        1. Hi Jack, thanks for the question. Yes, picking up the TfTT blogs depends on getting the category right. For the Quick SNITCH I look for either Quick Cryptic or Uncategorised categories. When I find it, I scan the whole blog and comments from scratch, so there should be nothing missing.

      2. Sorry, I missed this previously! On the off-chance you’re still checking this thread…

        I just add my name each time! But I shall investigate the mysterious possibility of ‘logging in’.

        UPDATE – right, now I have a login set up, thanks.

        If you do see this, a couple of Wednesdays ago the Snitch says I took dead on 30 minutes. This was one of the Championship puzzles, so I think the Snitch has picked me up as saying the three together took me 30 minutes on the day, and allocated that time to that one puzzle.

        I’m (honestly!) not precious about it, but if it’s an easy fix, could you just remove that one?

        I am also now in the Crossword Club as Amoeba89.

    1. Sorry that it’s not linked in the detailed results (or elsewhere). Yours is here: https://quick.xwdsnitch.link/solvers/1103

      I’ll try to provide a link with the Blog Solver TfTT name in future. You don’t have a SNITCH alignment yet, but this will come in due course and the end-of-month algorithm picks up your results.

    2. OK, I think I’ve fixed this now (and inserted the link from the Blog Solver name). Would you mind checking that it’s working okay?

      Thanks for your patience.

      1. Hello – thank-you. The link is working and more of my best times are showing up and my avg has adjusted accordingly.

        There was one discrepancy last Tuesday*. You’ll see in the comments, I posted early – time didn’t appear on Quitch. Later in the day I deleted my post and reposted in the hope it might get picked up. There’s a bit of discussion as to what happened – it was posted in wrong category etc – so maybe that impacted matters.

        Appreciate your efforts on this and I will get you another coffee when November rolls around 👍

        * This blog here … https://timesforthetimes.co.uk/qc-2813-by-pipsqueak

  7. Many thanks. I’m doing the puzzle on The Australian online, so slow to catch up. I’m unsure which Saturday to check. The puzzle for the 19 October is blogged on 26 October. Is the relevant Snitch the one dated 19 or 26? Does the lower number of reference solvers tend to make the Snitch lower? It doesn’t seem as if many of the usually slower solvers are in it. Don’t people do the Saturday puzzle?

    1. Hi anoja, thanks for the questions. The SNITCH for the Saturday puzzles are listed on the dates that the puzzles come out, not when the solutions are released. So, you will see that they are a week behind. The date and the crossword numbers are published in the header for each crossword (e.g. https://times.xwdsnitch.link/crosswords/3532), so hopefully it’s clear which crossword they refer to. [I don’t track how the Australian numbers link to The Times originals, but I’m sure that’s available somewhere.]

      With regard to the number of solvers, the delay of one week before the Club results are released means that the top 100 solvers (on which the SNITCH calculations are based) has many more neutrinos in the list. This crowds out many of the slower real solvers, who would normally appear at some time in the leaderboard’s top 100. I suspect that many slower solvers do the Saturday crossword, it’s just that they’re results aren’t visible.

      I’m not sure if this makes the SNITCH results lower. I can’t think of any prima face reason why it would, but sometimes these effects are subtle.

      1. Thanks for your reply. I don’t think there’s any longer a link between the two sets of numbers. I just have to remember what week the puzzles were originally published. The days are the same. If I get stuck I can search on an unusual word in the puzzle. Unfortunately the online Australian repeated a Sunday puzzle one week and never fixed it, so it slipped a week behind the one printed in the Weekend Oz, or two weeks behind the Times.
        My thoughts about the Saturday Snitch being lower is that it’s based on times mostly quicker than the usual range in the weekday puzzles.

        1. Thanks, anoja. I love your interest and welcome the questions 🙂

          On the links to the puzzles in The Austrlian, can I please check that you’ve seen this link: http://crosswordclub.org/tftt/showlinks.php referenced from the right sidebar above under “Syndicated Times Puzzles”. I’m sure the folk in the TfTT team would be keen to know if this process does not work. So, if it’s not delivering the right answer please let me know and I’ll pass the message on.

          With regard to the weekend times, the SNITCH rating is based on comparing the solving times to reference solvers’ average times. So, even if the sample of reference solvers is smaller, there’s no fundamental reason why it would tend to make the SNITCH rating lower. For example, Verlaine’s time on Saturday is compared to his average time in the same way as would happen on Friday or any other day. These individual scores are then averaged to give an overall rating which should, by and large, be delivering the same sort of hardness indicator for Saturday.

          [There are some second-order effects (I think) that are hard to predict. For the less speedy reference solvers (like myself), we only get included in the Monday-Friday SNITCH calculation if we’ve remained in the top 100 at the end of the day. This means that our fast times are included but our slower times are not. My intuition is that this would lower the Monday-Friday SNITCH rating rather than increase it, but I haven’t modelled the case extensively, so I’m not sure. Because of the neutrinos on Saturdays our times never make it into the SNITCH calculation for Saturday. If my intuition is correct, this would tend to make the Saturday SNITCH rating higher rather than lower.

          Probably a longer answer than you wanted, but thanks for the question anyway ;-).]

  8. Thanks again. I expect that link still works but, if I need to search, I just use the Search in the blog. The previous version of that link worked on the publication date and that’s what I was referring to as no longer available. I hadn’t realised or possibly forgot that the SNITCH isn’t just an average. It’s been a fair while since I read the background briefing. Maybe the calculation could use the top 100 after the neutrinos are eliminated? Just a suggestion and only if it would make a difference. I’m not trying to urge more work and appreciate what you have done. I’m also in the slower half. A good day for me is under 3 verlaines and a slow day is 5 verlaines. Of course I couldn’t calculate a verlaine rating without your list.

Comments are closed.