Wednesday, October 01, 2014

Vintage Super League

For the last month and a half Randy Buehler has been streaming weekly vintage Magic matches from an event called Vintage Super League. It's basically 10 famous Magic players (some from the Hall of Fame, some current active pros, some famous from the vintage tournament scene itself) playing in a round robin tournament. They can switch vintage decks every 3 weeks and all of the matches are played on Magic Online and streamed through Twitch. The production values of the stream have evolved over time and they've now got it set up so you can see the cards in hand of both players live. They also bring in 2 people for each game to comment on the games as they're happening, or to comment on previous games in the event, or the metagame around the event, or just to tell stories of days gone by. I'm finding it to be very entertaining to watch, and more people should check it out. Tuesday nights, or there are VODs available on Randy's YouTube page.

Randy is at 2-4 after 6 weeks and made a comment on stream this week about how he's now more worried about making sure he doesn't come last as opposed to trying to make the playoffs. (Top 4 advance, with ties broken by playing more Magic because who doesn't want to see more games of vintage?) With almost half the field advancing my gut feeling is a winning record should be good enough a reasonable chunk of the time. 2-7 is probably tied for last but I'd think 5-4 is probably tied for 4th. I wanted to verify that gut feeling, and what better way to do so than by turning to good old Sports Club Stats? (That's a website that simulates the results of a league by 'playing it out' hundreds of thousands of times. I've used it with Blood Bowl leagues and liked looking at the numbers it generates.) So I took the liberty of building a league on there...

I set the tiebreakers up to be 'playoffs' but I'm not experienced enough with the system to know what that means. If 3 people get to 6+ wins and 2 people get to 5-4 how is it going to treat the chances of making the playoffs for those people? 100%? 50%? Adding up all of their numbers gets me to exactly 400% of everyone making the playoffs, which is exactly 4 people in every scenario. So my guess would be if the simulation results in a tie it assigns that as a proportional berth in the top 4. Meaning the numbers it gives are not the odds of making it to a play-in game, but the odds of making it out of a play-in game. So the odds of making it to a play-in game are understated. I'm also not sure if the weighted odds make more sense in a league like this or the 50-50 odds. Especially given the players can change decks every 3 weeks I don't know that your chances of beating currently 1-5 Bob Maher are significantly better than your chances of beating currently 5-1 Steve Menendian. (The weighted system assumes that because someone is 5-1 they're a harder match and has them win more of their games going forward. It's often a reasonable assumption to make, but maybe not here.) It's interesting to look at both numbers, anyway, though for now I'm just going to go with 50-50.

According to the site Randy's 2-4 position will get him down to 10th place 9% of the time. Most of those are 0-3 finishes, but he can actually get to 4-5 and still finish 10th. On the other side he makes it up to 4th place (or higher) 5% of the time. Almost always from 3-0 finishes, but there are chances to get there even with a 2-1 result. What surprises me most is that the 3-0 finish is only good enough for a top 4 spot 38.5% of the time. I'm guessing the weirdness with the play-in game really takes shape here. It's hard to imagine 5-4 being a clean break to the top 4, especially with 2 people already at 5-1 and 2 more at 4-2. But I can't think of a way to make the site handle this gracefully, so we're stuck with what we have. Just know that the odds have to be better than 5% since 5% is the number after simulating the play-in games from any ties for 4th. And if you consider 3 or 4 way ties for 4th you could end up with a significantly bigger number than just 5%. (A 2 way tie for 4th cuts out half of your advances, but a 4 way tie cuts out 75% of your advances!)

The bottom line? The last 3 games are still super relevant even for someone at 2-4. Getting to 5-4 isn't guaranteed to make the play-ins (looking at Steve Menendian I can find an entry where he gets to a 6-3 record and comes 5th) but it's definitely still live. People roll a 20 on a d20 ALL THE TIME!

No comments: