A few weeks ago, I got into a reasonably-friendly back-and-forth on Twitter with Marshall Sutcliffe (@MarshallLRCast), host of the “Limited Resources” podcast, about the fact that Overrun was reprinted in M12 at uncommon.

The resulting chatter informed me that a great many high-level limited players dislike losing to the card, but they were also implying that it was a “bomb” and that it was too powerful to be uncommon because it showed up far too often that way.

Magic doesn’t do a lot of data-driven game design stuff on a card-by-card basis, mostly becuase–unlike games like World of Warcraft, League of Legends, and Farmville–we can’t change what our game pieces do on the fly in a live environment. In the past we’ve relied on tournament results, sales data, market research, and good old intuition to make decisions, although we’ve recently started getting better data out of Magic Online.

We don’t have access to anything that happens in-game (that stuff happens on different servers than the ones doing the data recording), but we can still see a lot of stuff. I’m not going to share a ton of data here, mostly because we still need to have a good plan for what exactly we want to share and how to best do so. I’m sensitive to handing players “the answers” about what cards to draft and play based on aggregated statistics, but there’s so much interesting material that it would be a shame to keep it all locked up.

So here’s a taste. The goal was to determine how much having any given card in your deck affected your win percentage in M10 sealed deck (I have no fear sharing these numbers as M10 sealed as a format is two years old and no one is playing it any more). The methodology was such: We looked at every games played in Round 1 of every M10 sealed event ever played on MTGO (later rounds were ignored to avoid selection bias towards good decks). Every card in the set was looked at–what percentage of games were won by decks containing at least one copy of the card in question?

Here are the top 25 finishers:

It takes a second to realize that there’s no speculation here, no guessing. This is a historical record. A series of facts.

From these numbers it’s hard to say Overrun is a bomb. It isn’t the top green card, nor the top uncommon. Heck, it isn’t the top green uncommon. That said, it still can be disheartening to lose to.

We’ve just scratched the surface; there are many other ways to analyze data like this, many more questions to ask. Which cards do better in the hands of more skilled players? How does color imbalance affect things? We’re going to be taking deeper dives into this kind of data, and my hope is that we’ll eventually share more of it on dailymtg.com or mtgonline.com, either as part of Tom LaPille’s column or via some new outlet. If you have ideas of how to better slice what we have, I’d love to hear it.

Other factoids:

• Only 70 of the 229 cards in M10 have win % over 50%.
• The top-finishing card in full Scars block limited is Wurmcoil Engine at 59.79%. What a beast.
• Mind Control is in 6th place here and in 2nd place in M11. Now that’s a power uncommon. This data may be the death knell for uncommon “Control Magic” effects, so enjoy Mind Control in M12 while you can.