58K Games of Magic Origins Limited Analyzed

Today we get a look at a huge source of objective information in the form of Magic Origins draft statistics provided by our good friend Rolle. You can see the full statistics page for Magic Origins here and the individual card win percentages here. Today we'll go over some of the more interesting data points from the data.

Format Speed

There are on average, 9.1 turns and 6.2 lands played per player at the end of a game of Magic Origins limited. This pegs this environment as being a bit faster than the previous draft format of Dragons of Tarkir x 2, Fate Reforged. If you look at win% on the play, it's as fast as Theros. This speed is highly influenced by the presence of Renown creatures. Aggressive starts are made more punishing by these "free" +1/+1 counters, and that means playing a creature on turn two is very important whether you're attacking or blocking. This is similar to the Dash creatures of Dragons of Tarkir that sped up the entire draft format.

Top Colors

Asking what color is "best" in a format is a loaded question. Are you trying to determine which color has the best cards, or which color has the highest win percentage? These are actually two distinct questions.

The color that has the highest win percentage is Red. Decks with Red in them won 51.4% of the games they played. But that's not the full picture. Are Red decks winning because Red is underdrafted or because Red cards are strong in Origins? Let's look at the play percentage and win percentage of the five colors in Magic Origins.

We see that Red is played more than Blue, Black, and Green. So despite being drafted more (what some may call "overdrafted") relative to these three colors, it still ends up with a higher win rate. This indicates that Red is a deep color in the format; its cards are strong enough to support more drafters while still maintaining a healthy win rate. While Red is currently overdrafted, it could probably stand to be a bit more overdrafted.

Viewing the win percentages in comparison to play percentages, Red and White are the clear front runners, Green is somewhere in the middle, while Blue and Black bring up the rear.

Top Archetypes

Looking at the ten 2-color archetypes, we see that Black/Red, Blue/Green, and White/Red have the highest win percentages. Black/Red and White/Red managed this while being played more on average, while Blue/Green is played less than any other 2-color archetype. This indicates UG is being underdrafted compared to how many wins it puts out; it is also an indication of the color pair being shallow — Blue and Green's top commons drop off in value sharply after the first few in each color.

The fact that many players view Blue and/or Green as the worst colors in Magic Origins, combined with the fact that the Blue/Green archetype doesn't have a clear direction like "Artifacts" or "Elves", has apparently led people to avoid the color combination despite its success.

At the opposite end of the spectrum, look at Black/Green. It is the second most popular 2-color combination in Origins by play percentage, but the third worst in terms of actual win percentage. This indicates that the color is probably being overdrafted by a fair margin.

The colors and archetypes in this set are quite close in power level. It's more important to find out what's open for the position that you're drafting and to read the pod correctly than it is to force a specific color combination.

Individual Cards

Here are some cards that stood out to me:

• Gideon's Phalanx does a better job of closing out a game than Mage-Ring Responder does at 74.4% win rate versus 66.9%.
• ​Akoran Jailer looks to be bad, but not laugh out loud terrible.
• Willbreaker has a higher win percentage than Soulblade Djinn. The confidence intervals are a bit wide since the number of observations for each are in the low thousands, but it's still impressive that Willbreaker is in the same arena as the Djinn.
• Might of the Masses is probably even better than you think.
• Send to Sleep is played far more often than it probably should be.
• It’s interesting to see the difference between the win% of swift reckoning vs Reave Soul. This seems to come up in every set where the removal that lets them hit you first has a much lower win% than things that let you reach out and kill something. This may just be because it means you are almost surely losing (or a least racing) by definition when you cast Swift Reckoning where you are just as likely to be ahead as behind when casting reave soul.
• The top Common creatures for each converted mana cost are Topan Freeblade, Ghirapur Gearcrafter, Ampryn Tactician (which suffers slightly from the Overrun problem), Rhox Maulers, and Vastwood Gorger.
• The top Uncommon creatures for each converted mana cost are Undercity Troll, Bounding Krasis, Whirler Rogue by a mile, Conclave Naturalists, and Sentinel of the Eternal Watch.
• The top Rare/Mythic creatures for each converted mana cost are Knight of the White Orchid, Managorger Hydra, Pia and Kiran Nalaar, Gilt-Leaf Winnower, and Woodland Bellower.

General Observations

• There are only two color “raw” power categories. Red and White (with Red slightly ahead) and everything else. Green is not the worst, it is equal to blue and black.
• Splashing is generally very poor.
• The most important factor of ORI draft is identifying the open colours for your seat. Every colour pair is viable, and the gold uncommons really are very strong.
• Draft a curve. Win % for playing a 2-drop on turn 2 (excluding 1-drops) is 56% (vs 44% for no 2-drops). Play vs Draw this is even worse.
• Mill may actually be viable. At least the pieces of it don't have straight up appalling win %'s
• This set is the opposite of Fate Reforged. I have never seen a set with such low impact rares and particularly mythics. This is the set in which the good uncommons really do outshine the mythics!

Conclusion

Explore the data on the statistics page and see if you find any interesting anomalies for yourself. If you're interested in similar analysis but for Constructed instead of Limited, see these excellent articles:

• Statistical analysis of Standard metagame and matchups
• Statistical analysis of Modern metagame and matchups

Reach out to me in the comments below or on Twitter @JakeStilesMTG if you have any questions or comments about the data or the collection methodology, or if you discover any insights to share.