Spreadsheet can be found here, if you’re a math nerd and want to check my work: https://docs.google.com/spreadsheets/d/1LYBGYQMHkUFJVamPEeEHbbfmY7NCjP0aIuXU3MTquC8/edit?usp=drivesdk
Polarization
Polarization is how far, on averages, matchups are from even. 0% polarization is a 50-50 matchup and 100% polarization is a 100-0 matchup. For the purposes of comparing decks against each other, two types of polarization are measured: raw polarization, which is from 50-50, and adjusted polarization, which is from the deck’s overall winrate (so for a deck with 55% overall winrate, the distance from 55-45).
Diamond 4-1 polarization
Archetype | Class | Adj. Pol. | Raw Pol. |
---|---|---|---|
Reno | Druid | 19.02% | 18.92% |
Handbuff | Paladin | 17.40% | 17.71% |
Dragon | Druid | 16.61% | 16.70% |
Zarimi | Priest | 16.48% | 18.00% |
Spell | Mage | 16.29% | 21.03% |
Shopper | DH | 15.83% | 15.16% |
overall | 15.79% | ||
Reno | Shaman | 15.34% | 14.68% |
Reno | Warrior | 15.31% | 15.13% |
Pain | Warlock | 14.90% | 14.76% |
Rainbow | Mage | 13.68% | 19.46% |
Aggro | Paladin | 13.53% | 14.09% |
Rainbow | DK | 13.03% | 14.61% |
Cutlass | Rogue | 12.29% | 14.75% |
Excavate | Rogue | 12.25% | 13.90% |
Plague | DK | 11.60% | 12.44% |
Pirate | Rogue | 11.57% | 20.29% |
Token | Hunter | 11.44% | 21.10% |
Sludge | Warlock | 11.21% | 12.62% |
Snake | Warlock | 10.51% | 10.68% |
Reno | Priest | 10.20% | 20.28% |
Wheel | Warlock | 9.02% | 12.75% |
Hybrid | Druid | 7.58% | 7.59% |
Legend polarization
Archetype | Class | Adj. Pol. | Raw Pol. |
---|---|---|---|
Dragon | Druid | 16.44% | 16.53% |
Reno | Druid | 14.84% | 14.94% |
Rainbow | Mage | 13.65% | 18.12% |
Aggro | Paladin | 13.55% | 12.49% |
Pain | Warlock | 13.18% | 13.18% |
Spell | Mage | 13.10% | 18.74% |
Reno | Shaman | 12.61% | 12.86% |
Handbuff | Paladin | 12.54% | 12.85% |
Reno | Warrior | 11.85% | 10.76% |
Rainbow | DK | 11.57% | 13.43% |
overall | 11.26% | ||
Token | Hunter | 11.20% | 16.82% |
Shopper | DH | 11.00% | 9.94% |
Zarimi | Priest | 10.96% | 14.16% |
Cutlass | Rogue | 10.46% | 10.46% |
Pirate | Rogue | 10.09% | 15.50% |
Sludge | Warlock | 9.95% | 10.30% |
Reno | Priest | 9.79% | 16.18% |
Snake | Warlock | 9.39% | 9.30% |
Excavate | Rogue | 9.13% | 10.28% |
Hybrid | Druid | 8.21% | 7.76% |
Plague | DK | 7.84% | 9.89% |
Wheel | Warlock | 7.47% | 13.80% |
The average game in D4-1 currently (and by “currently” I mean the 25th through 30th of last month) is 57.9% to 42.1%, which is quite polarizing. Historically, Legend is normally more polarizing than Diamond, but not this time; the average game there is 55.63% to 44.37%.
Skill Factor
Skill factor is based on the concept that overall winrate is the sum of the product of your matchup winrate against a deck and the popularity of that deck, for all decks in that meta. But what if we used the matchup winrate table for one meta, say Diamond 4-1, while using the popularity from a different meta, say Top 1000 Legend? In this way we can isolate how much winrate change is due to differences in deck popularity, and how much is due to differences in matchup winrates. Since the only thing that is changing as matchup winrate changes is the rank, and therefore the average skill, of the players, this allows for a measurement of how well a deck rewards increases in the skill of both players (so a deck that gets worse the better its opponent is, faster than it gets better from having a more skilled pilot, can have a negative value).
As far as a practical application for this measurement, the table includes a column that is the sum of D4-1 winrate and skill factor (which again, can be negative). That sum yields my best advice regarding what you should play if you are a highly skilled Top Legend player by end of month, but you are currently in Diamond or in lower Legend trying to climb.
Skill Factor
Archetype | Class | D4-1 WR+SF | Skill Factor |
---|---|---|---|
Token | Hunter | 54.30% | -5.23% |
Zarimi | Priest | 54.02% | 0.63% |
Reno | Warrior | 53.30% | -0.40% |
Aggro | Paladin | 53.15% | -2.79% |
Handbuff | Paladin | 52.32% | 0.48% |
Sludge | Warlock | 52.11% | -0.22% |
Shopper | DH | 51.71% | -0.04% |
Hybrid | Druid | 51.10% | 1.08% |
Pain | Warlock | 49.10% | -0.09% |
Excavate | Rogue | 48.89% | 3.14% |
Reno | Shaman | 48.17% | 0.43% |
Snake | Warlock | 48.05% | -2.32% |
Dragon | Druid | 47.94% | -0.38% |
Reno | Druid | 47.10% | -1.91% |
Cutlass | Rogue | 45.29% | -0.04% |
Rainbow | DK | 44.85% | -0.22% |
Plague | DK | 43.80% | -2.98% |
Wheel | Warlock | 43.78% | -0.43% |
Rainbow | Mage | 41.71% | 0.15% |
Spell | Mage | 40.67% | 0.24% |
Reno | Priest | 40.13% | -0.10% |
Pirate | Rogue | 39.68% | -0.53% |
To summarize the table, if you want to just climb as quickly as possible, run Token Hunter, even though it’s the lowest skill factor deck in the meta. If you want to climb almost as quickly, but feel somewhat clever as you’re doing so, run Zarimi Priest. And if you want to feel extraordinary clever while barely climbing at all, run Excavate Rogue.