It’s been a while since I ran a simulation — nearly eight months! — and with the recent announcement of the May Melee mini-tournament, I thought that now would be the perfect time to dust off the old spreadsheets and re-learn how I set them up to do this kind of madness. The only issues were:

1. With such a small portion of the season having been completed, everyone’s stats are likely to be rather imprecise.
2. Hero pools are probably making things even more imprecise.
3. The internal turmoil of certain teams (notably Boston and especially Vancouver) will make this potentially even more imprecise.

So as you can see, these predictions should be taken with a healthy grain of salt. Heck, go ahead and empty the entire shaker.

If you’re new, the way my simulations work is that I use SAMP to predict the outcomes, including map scores, of each match. I’ve run that simulation for May Melee 1,000 times and the results are presented below. SAMP only knows how teams have performed in the past and doesn’t make any guesses about how teams will get better or worse in the future, especially with regards to such things as hero pools or roster changes. So there, go ahead and heap on a few more piles of salt.

Still, I think these predictions will be broadly accurate — they think Shanghai and Philadelphia will do well, for instance — and will be fun to look back upon in three weeks, whether they were mostly correct or mostly incorrect and you’ll be free to praise or ridicule me, respectively.

(One other note regarding these simulations: Beyond differential, I just used a random number generator to break ties. Anything more complex would have been essentially impossible for me to simulate.)

First, let me present the raw data. This is the number of times each team fell at each particular seeding in its region:

[table “112” not found /]
[table “111” not found /]

And here are the average seeds and their ranking by those seeds:

[table “114” not found /]
[table “113” not found /]

Unsurprisingly, Shanghai looks nigh unbeatable. The Dragons’ SAMP is a ridiculous .967, which means that, against an average team, Shanghai would be expected to win maps 96.7% of the time. While the Dragons are having an amazing year, they probably won’t maintain a 6:1 map win:loss ratio all season long.

On the other end of the spectrum in the Asian region, we have London, which is expected to struggle, despite currently having more wins than losses. All three of their victories have come by 3-2 margins, though, and their schedule has been relatively soft, giving them a low SAMP (.430) and a difficult road against the top-end teams in Asia.

In North America, we know Philadelphia is good, and most people are high on Atlanta, but don’t sleep on Florida. Even though their SAMP (.494) ranks them as an average team, the Mayhem probably have the easiest schedule over the next few weeks, with games against Boston, Washington, and what remains of Vancouver, so 2-1 or even 3-0 is hardly out of the question. Technically, everyone is going to make these “playoffs,” so I don’t know if we should really count it as Florida’s first playoff trip, but even if the criteria were a little tougher, they’d be a near shoo-in to qualify.

Speaking of Vancouver, they’ve got the weirdest spread of seeds, probably owing to the fact that they’ll be playing four games while every other team in their region plays three. Such variance is properly befitting a team that’s going to look wildly different from what fans saw in their first few games.

And never give up hope, Chengdu, Dallas, and Houston fans — your team came out on top once each in 1,000 simulations. Imagine how you’d do with Florida’s schedule.