Monday 17 June 2013

Welcome to Vegas

Welcome to Las Vegas

Vegas is going to be the largest magic tournament ever. And its going to be the largest by a big margin. Which makes it an interesting exercise to try and figure out the implications for how that effects records and the cut to top 8.

For those who read my facebook notes, you will have seen my previous attempt to make a guess at what records would top 4 the Player’s Championship. Something I nailed with reasonable accuracy. The methodology is fairly simple. I assume everyone is 50/50 in every matchup and then simulate it a crapload of times. Draws are not allowed.

There were 3 problems adapting this framework to Vegas:

  • How to incorporate byes.

I figured Vegas would have about 4000 people and a large number of those would have some number of byes. The original program isn’t designed to handle that, so I figured I would compensate by just setting the number of participants at 5000.

  • Time to Run

The initial version was pretty slow, which wasn’t a big deal when I had to simulate a 12 round tournament with 16 people. I had to make some pretty big adjustments to speed it up for a 15 round 5000 person tournament. I don’t think I made any errors when making these adjustments, but who knows.

  • Trials.

I am down to about 100 trials because of how long this thing takes. The information regarding 12–3 players is based on 10 trials.

Results

  1. The record needed to top 8 will be 13–2. In all 100 trials 8th place was 13–2.

  2. Between 17–20 people end the tournament with that record or better. The average was 18.72. So on average 10 people missed the cut at 13–2.

  3. An average of 87.8 people were 12–3 or better. Which means 20+ people were missing the money with a 12–3 record. The first GP I ever travelled to (GerryT winning in Denver) that record was a lock for t8. This actually makes me suspicious that I did something wrong (because the result is so ridiculous), but I haven’t found what it could be yet.

Practical Implications

  1. You can drop at X–4.

  2. Draws are much better then they would otherwise be, especially late on day 1.

    For example, assume you are 7-1 going into the last round of Day 1. A draw is going to be significantly better than a loss here assuming you care about t8 only. With either a draw or a loss you have to win out to have a shot at Top8ing. But with a draw you are 100% to top 8 assuming you win out. With a loss you are almost 100% eliminated from top 8.
    
  3. If you are 7–2 on day 1, your odds of top 8ing even if you win out are essentially zero.

Breakers are going to a matter a lot for the top 8 cut, so losing early is costly (See above). Though you can still obviously qualify for dublin.


5 comments:

  1. Should this not be a 16 round tournament.
    Paris 2009 (biggest gp at that time) had 10 rounds on saturday and then 6 on sunday.
    don't see why they would not do this here..

    ReplyDelete
    Replies
    1. Individual GPs are now capped at 15 rounds, as per the OP policy update.

      Delete
    2. Has that kicked in yet though? Providence was 16 rounds.

      Delete
    3. Providence was an exception as a Team GP with a cut to top 4. Las Vegas will be 15 rounds.

      Delete
  2. This comment has been removed by the author.

    ReplyDelete