Thursday, January 25, 2018

GDS3: Good luck on Trial 2!

... not that you'll need any luck, with all the excellent preparation work you've been doing. I can't wait to see how many of us make it to Trial 3!

We've been asked not to discuss any questions until noon PST on Saturday, so that's when the discussion thread will go up. In the meantime, of course, please refrain from posting anything even vaguely spoiler-ish.

In the meantime, get a good night's sleep. Best wishes from all of us at Goblin Artisans. Thank you for being such an awesome, thoughtful, and creative community.

32 comments:

  1. If y'all didn't see the thread of tweets from him, there will be 75 questions.

    ReplyDelete
  2. Where's the like/retweet button on this thing?

    Good luck all. We're rooting for you.

    ReplyDelete
  3. Good luck, my friends!
    I'm proud of you.

    Don't commit to answers until you've written about each one.
    Beware: When you refresh the form, the order of answers may have changed.

    ReplyDelete
  4. You may want to check Mark's twitter feed when you start working on the test. He's already noted a typo in one of the questions.

    ReplyDelete
    Replies
    1. Thank you for pointing this out. Anyone else feel like they really earned their dinner tonight?

      Delete
    2. I would, except that I didn't really get started until after dinner...

      Delete
  5. Twelve hours later... I'm exhausted. Can't wait to talk about these when we're allowed.

    ReplyDelete
    Replies
    1. I'm exhausted too!

      We'll have a post for discussing the test at exactly 3:01 PM Eastern. Can't wait to hear everyone's thoughts.

      Delete
  6. "Get a good night's sleep"... I think you jinxed me, HavelockV. My work woke me up with an emergency customer issue at 3am PST.

    It's been an interesting twenty-four hours, but I did finish and I'm reasonably happy with how it went. Looking forward to discussing!

    ReplyDelete
  7. ~65 questions took me a minute or so each.
    ~5 took about 9 hours total.
    And 2 of them were still agonizing coin tosses!

    I'll be posting things.

    ReplyDelete
    Replies
    1. Yeah, about that. Except I didn't spend anywhere near as long.

      Delete
    2. I'm in the same boat here, except I had more like 7 agonizing coin tosses.

      Delete
    3. Don't forget the trick questions... They said they weren't doing them, but I think at least one snuck in.

      Delete
  8. Yeah, this definitely seemed to be testing things that extended outside design, like MARO said.

    ReplyDelete
  9. I have been following this blog and thanks everyone for the help.

    ReplyDelete
  10. 80/20 Rule to the effin EXTREME!!!! Pretty much did the same as Jay, finished most of it in an hour then had to chew the fat on the rest.

    If my power level is hitting Jay levels heres hope I make it to the Third Trial (And promptly get BTFO since I dont have a set proposal ready)

    ReplyDelete
  11. If even Jay is thinking he got at most a 73, it puts me a little at ease.

    ReplyDelete
    Replies
    1. Adding up all my confidences, my score is 68.57.

      Delete
    2. I am feeling 68 is going to be the cutoff. I hope I get there.

      Delete
    3. Those are oddly specific numbers! I know I got at least 60 and probably 65+, but beyond that I wouldn't care to speculate.

      Delete
    4. Well will be easier speculate on Monday when we know how many out of 3000 entered.

      Delete
  12. I find it interesting they don't start with card design, seems an easy way to just get a lot of cards.

    ReplyDelete
    Replies
    1. It's more work to sift through cards from an unvetted source than to make them yourself.

      Delete
    2. It would have been interesting to just be like, give us your best card. Pretty impossible thing to ask. but 3,000 cards is more than they'll end up looking through but I just feel the multiple choice does not show off what they end up looking for.

      Delete
    3. A bunch of the questions are trivial if you do (or have done and internalized) the work of reading his articles and checking gatherer and google. They're just there to weed out the folks who aren't committed enough to do that work. And then there are the really hard questions, to ensure they have a top 100 rather than a top 500. Some of those are a very good measure, and some are not. While this method could (and does) eliminate good candidates, they can be sure their 100 folks will be worth looking at. (So they'll get false-negatives but not false-positives.)

      Delete
    4. The other thing is, it's prohibitively time consuming to grade 3000 of anything. Any sort of test on that scale really must be automated.

      Delete
    5. I'm salty about that typo, it made the question so much easier (and I actually think the typo was a good questiom given its reference to an obscure but important rules interaction)

      Delete
  13. I spent my entire time at the gym yesterday afternoon chewing over one of the questions. I'm sure I got way too many wrong, but the experience itself is very helpful.

    ReplyDelete
  14. The one piece of advice I really wanted them to follow, they didn't: To give the second-best answer a non-zero value. By keeping each question's score completely binary, they're throwing away data, making the results of their test a less accurate measurement of Magic knowledge and design ability. And by making questions needlessly ambiguous, they reward the ability to take multiple choice tests, which is not a relevant skill for the position.

    (That's public information MaRo shared yesterday.)

    ReplyDelete
    Replies
    1. Or worse, it’s actively getting them worse candidates, as designers who think “too much” and are aware of the complexities get it wrong while level-one thinkers don’t even realize there could be more to it.

      Delete
    2. Yeah this one had some MURDUEROUS questions. Several "grey area" ones imho.

      If anything this means the curve grading will be smoother.

      Delete
  15. I also feel if we get into a situation like GDS 2 where 80% of the people answer one answer that is wrong the question should be thrown out.

    ReplyDelete