Modern Izzet Phoenix—Building a Better Bird Based on Data

Izzet Phoenix is the hottest new addition Modern has seen in a while. Its win percentage across the two most recent GPs proved higher even than KCI’s, and its metagame share was still growing when the circuit last touched base with the format. Eli Kassis’s victory at Grand Prix Oakland could be explained away as an outlier. But one cannot argue with the four copies of the deck in Grand Prix Portland’s Top 16 nor with its 503-377 aggregate record over the course of the two tournaments. The latter even included a negative 15-21 performance against Ironworks, which shall not be repeated in the future.

Speaking of the future—expect the trend to continue. If any archetype can rise to the challenge and unseat Spirits from the top of the metagame breakdown, it is Izzet Phoenix. At the top of the food chain it arrived long ago.

When a deck enjoys its first major breakthrough, there are often various versions in circulation. Now this deck went from one major breakthrough right on to the next major breakthrough, so it isn’t exactly uncharted territory anymore. But while the deck design space may no longer look like the Wild West, Izzet definitely has enough of the steampunk going on to make everything feel at least a bit like the Wild Wild West. Factors include that the possibilities in this format are plentiful, Modern players are notoriously loath to adapt, and published deck lists were few.

Well, let’s change this, shall we? If you’re interested, take a look at a Google Doc with all the Izzet Phoenix main decks I could get my hands on that players registered for GP Port- and Oakland. The list of lists is sorted by the match points their pilots earned, including Top 8 playoff results. What I omitted are the pilots’ names. After all, we don’t want to shame anyone for going 0-3 drop, do we?

The full thing is somewhat unwieldy and hard to parse. Feel free to copy the source material and do your own analysis. Or rely on me, and read on.

Four Degrees of Separation

Comparative Arclight Ornithology is a new field of study, so I had to make up some ground rules on the, ahem, fly. I was interested to see what separated the successful Izzet Phoenix versions from their less fortunate cousins. For this, it was necessary to divide the full population into different categories.

The division is somewhat arbitrary. I tried to follow meaningful cutoff lines, but I also favored separations that led to samples large enough on their own and at least somewhat comparable in size.

29 players crashed and burned themselves rather than burn anyone else. They earned between 0 and 10 match points.

Another 36 players were in the running for some time but also failed to cross the Day 2 threshold. They finished on 12 to 15 match points.

23 players won seven to ten of their matches. Very few made it into the Top 100 at GP Oakland with a record of 10-5, and none did at the much larger event in Portland.

17 players went 10-4-1 or better.

Unequivocal Cries of the Firebird

A couple staples of the archetype proved universal. First of all, one hundred percent of all lists included four copies of Arclight Phoenix, and the same was true for Faithless Looting and Manamorphose.

Lightning Bolt and Thing in the Ice were a 4-of in 104 of 105 decks. The one exception was a spicy hybrid build with some Jeskai Ascendancy combo mixed in, where sadly not only the Lightning Bolts but all the nuts and bolts came off. The person in question eked out a mere three wins at GP Portland. Then the same person switched to a regular version for GP Oakland and went 9-6, proving that it wasn’t gross incompetence as a player that led to the earlier brew’s downfall.

The next most common spell was Serum Visions:

The differences between the tiers of success don’t reveal much of anything here. I’m just showing this chart to illustrate what you’ll find more of below. Almost everyone used four Serum Visions. One player at 12–15 points went with two and one with zero copies. Among the 17 most successful players, one went with a single copy. All in all, it seems dubious not to have a full playset.

Next in Line

The next most played spell also was the first to exhibit an interesting pattern of employment.

Top finishers used fewer Opts than the people at the bottom of the ranking. This isn’t because they managed with fewer cantrips. Rather they ran an alternative at higher rates than anyone else.

Permanent Payoff

In addition to the namesake and Thing in the Ice, Phoenix decks typically include a few other permanents that benefit from the presence of the many non-permanents. Most of the available options are further creatures.

Although the results aren’t quite conclusive, there is a visible trend here. 86.2% of the least successful decks contained Crackling Drake, compared to 58.8% of the most successful decks. It’s great that the research aerostaticasters of the Izzet League came up with two new flyers frequenting Modern, and it’s neat how this one circumvents Rest in Peace, but maybe Crackling Drake isn’t all it is cracked up to be. At the very least, the data is pretty clear on the topic of the third copy: less than half of the best decks included three, whereas more than half of all the others did.

Then again, I haven’t seen enough of the mirror yet. If toughness 4 is a big deal in the matchup, it may be time to get cracking with a vengeance. I’d prefer to bet on Flame Slash. The card showed up in three main decks, twice as a 1-of, while the one player who made room for two Slashes finished first at GP Oakland.

This picture is much less clear, which indeed is valuable insight in its own right. Without a doubt, most folks who jump on the Phoenix hype plane now will copy the GP winning list and will run no Monastery Swiftspear. But the data does not actually settle the question whether to spear swiftly or not to spear at all.

A couple of interesting notes nevertheless: 45% of players who reached Day 2 took a complete pass on the 1-drop, whereas only 32.3% of those who failed did. Both players who made it to the semifinals had none. At the same time, the most successful people also were the most likely to go all out with four copies. So, although two was the most common number of Swiftspears across the board, it may not be the best number.

Bedlam Reveler remains too much of a fringe inclusion that the data couldn’t reveal any big trends, except one: Three copies or more were so much more common at the top than anywhere else that the card deserves another look.

Baby-face Jace is another case of dubious sample size. Yet the chart shows a definite correlation between his presence and a player’s among the top performers. 12.5% of Day 2 decks contained Jace, but only 3.1% of the others.

That’s it, mostly, for payoff creatures. Young Pyromancer played a huge role out of Eli Kassis’s sideboard, but only one showed up in all of these 105 main decks. Three decks with a singleton Snapcaster Mage each failed to qualify for the second day, as did the one splashing for two Tasigur, the Golden Fang. Splashing for three Monastery Mentor, on the other hand, resulted in a 12-3 record that could have been a fluke but might very well become a trend. I’ve seen Mentor fight alongside Arclight Phoenix since. Sometimes it’s too clunky, but when it’s good, it’s scary.

Eli Kassis voiced disappointment with his pair of main deck Pyromancer Ascension. The GP champion even wanted to bid the enchantment farewell when updating his deck. Actually, though, Ascension fared pretty well already.

Scouring and Discarding

Cards such as Jace, Vryn’s Prodigy, Pyromancer Ascension, Snapcaster Mage, and Crackling Drake like to find the graveyard fully stocked, and Arclight Phoenix wants to get in there itself. Faithless Looting is the best way to achieve either, but this supply alone doesn’t meet the demand.

The chart tells a story torn in two: simple at first, then less so. Throughout the lower three tiers, more Thought Scour seems to equal more points. Only at the very top do things gain a bit of nuance. To put more Thought Scour into your deck yields better results, but to put more thought into your deck yields the best results.

Listing all the various inter-dependencies goes beyond the scope of this article. Just know that Pyromancer Ascension coincided most reliably with a maximum of Thought Scour. All Ascension players used four, except one person who ran Ascension with zero. Such folly was rewarded with elimination on Day 1. This in turn makes the enchantment’s results a little more impressive still: Two-thirds of Izzet Phoenixes with Ascension made it to Day 2, but of all reasonable Phoenix pilots with both Ascension and a playset of Thought Scour three quarters did.

People who wonder about the second-best actual discard option for an Izzet deck may jump to, oh, of course, Izzet Charm. Duh! As iconic and as charming as it may be, the card’s prevalence was lowest among the highest finishers.

In fact, Izzet Charm wasn’t even the most common discard option next to Faithless Looting. The 105 main decks contained a total of 110 Charms.

…while Lightning Axe weighed in at 168 copies. Roughly 90% of players across all levels went to work with one or two Axes. Considering how many Thing in the Ice and Crackling Drake one has to deal with in the mirror, this sounds like a smart move.

Chart a Course still played a role as well, or two roles even. While it’s super important to discard the Phoenix, the deck doesn’t have a lot of other stuff to discard for value. (The average deck contained 0.38 Fiery Temper, the 17 top finishers ran an average of 0.53.) So it’s just as well that Chart a Course sometimes acts as a source of card advantage and always maintains card parity.

Maximize Velocity is another card with a dual role. When you’re low on ways to discard, it is one, and when you have too many you can store Maximize Velocity in your graveyard for future use. It isn’t a good idea to actually maximize the numbers here, but 23.5% of the 17 top performers ran one, whereas only 16.2% of the total population did.

Shot Through the Heart

The perfect Arclight opening involves three instants and/or sorceries on turn 2. Casting three spells off of 2 mana sounds more like a tall order than it is in practice, mostly thanks to Manamorphose and…

Once again, the first three bars show a relatively neat progression, only for the final column to topple every assumption so formed. One possible explanation why the number of Shots goes down again at the very top is because the number of decks with four Monastery Swiftspear went up. As a consequence, we can find more Mutagenic Growth per player in this segment than anywhere else.



Scroll to Top