Updated post-ELD MTGO picture. This is only Challenges, Premiers, PTQs, and other higher level, non-curated events with T32 data available. N=368 decks, which isn't a terrible sample to get a metagame picture, but is also more limited than even a single GP. At least this has the benefit of spanning more time. These 19 decks represent about 80% of the metagame and in the old Modern Nexus updates would have encompassed both Tier 1 and Tier 2 decks.
- Eldrazi Tron: 12.8% (n=47)
- Grixis Death's Shadow: 7.9% (n=29)
- Amulet Titan: 6.3% (n=23)
- Burn: 6.3% (n=23)
- Dredge: 5.2% (n=19)
- Sultai Urza: 4.6% (n=17)
- Sultai Death's Shadow: 4.3% (n=16)
- Infect: 4.3% (n=16)
- Jund: 3.8% (n=14)
- Paradoxical Urza: 3.3% (n=12)
- Humans: 3% (n=11)
- Mono R Prowess: 3% (n=11)
- Mono G Tron: 2.4% (n=9)
- CrabVine: 2.4% (n=9)
- Bant Snow Control: 2.4% (n=9)
- Simic Eldrazi: 2.2% (n=8)
- 4C Whirza: 2.2% (n=8)
- Azorius Control: 2.2% (n=8)
- Gifts Storm: 2.2% (n=8)
We can group this Tier 1 and Tier 2 list into a few rough categories. All percentages are of the 80% of top decks, not of the whole:
Tron and big mana: 21.5% (E Tron, G Tron, Amulet)
Urza variants: 10.1% (Sultai, Paradoxical, 4C Whir)
Death's Shadow variants: 12.2% (Sultai, Grixis)
Less/non-interactive aggro: 24.2% (Burn, Dredge, Infect, Humans, Mono R Prowess, Crabvine)
Midrange: 6% (Jund, Simic Eldrazi)
Control: 4.6% (Bant Snow, Azorius)
Combo: 2.2% (Storm)
This doesn't fully capture some of the deck differences. For instance, Humans is SIGNIFICANTLY more interactive than Dredge, and yet is in that aggro category. By a related token, Simic Eldrazi is also significantly less interactive than Jund, but it's more midrangey than anything else. Urza and DS decks cross common categories and can pivot between roles, so they get their own categories. "Big mana" is also somewhat misleading, as its 3 grouped decks have pretty different gameplans, but I'm comfortable overlooking those differences for the purposes of a macro-meta analysis on a forum post.
Overall, this picture shows a relatively "diverse" Modern in terms of different types of decks and even different pillars. Some of those pillars are even pretty fun, skill-intensive, widely accepted, and uniquely Modern, namely the DS decks. Others are more polarizing. That said, it is alarming that midrange and control have so few viable options compared to the more proactive strategies. This is a hard issue to fix. If I worked in R&D and had to force this fix without waiting for changes, I'd probably do some hyper-targeted "nerf bans" at some targets that are not obviously egregious by known ban precedents, but nonetheless tilt the metagame. Top examples include:
Veil of Summer, Mycosynth Lattice, Oko, Nature's Claim, Once Upon a Time
None of these cards represent the essential identities of decks that use them. Two of them are sideboard cards that disproportionately help decks beat regulatory mechanisms (Veil and Claim), one of them is an "I win" element of an otherwise robust and interesting toolbox package (Karn), one is added and unnecessary consistency (OUaT), and one invalidates a wide variety of counterplay (Oko). Notably, all but one of these cards is green, and all but one of these cards is either a 2019 card or a direct result of a 2019 card (Claim is older, Lattice is older but problematic because of 2019 Karn). No one loses their deck as a result of these bans, as opposed to more heavy-handed targets like Urza, Opal, a Tron piece, etc. These decks just become more focused around a more fragile, less decisive gameplan, which opens up the format to other competition.