Pokémon: Mathematically proving the need for a Ground-type Eeveelution

Reader, I am a self-professed idiot. A couple of weeks ago, I clocked out of work as per usual and decided to conduct a quick investigation into something that piqued my interest: The possibility of a new Eeveelution being introduced to Pokémon. After spending several hours of my own free time developing a rigorous methodology for this study while mindlessly running basic arithmetic, it suddenly dawned on me that I may, in fact, be wasting precious hours of my life on God’s green Earth.

Who am I kidding? This is science, sheeple. By the end of this article, you’ll be advocating for my Nobel Prize nomination. 

Let’s start at the beginning. There are, at the time of writing, eight Eeveelutions: Vaporeon, Jolteon, Flareon, Umbreon, Espeon, Leafeon, Glaceon, and Sylveon. It should be obvious, then, that there are also eight types of Eeveelution — Water, Electric, Fire, Dark, Psychic, Grass, Ice, and Fairy, respectively. It’s important to remember that Eevee itself technically accounts for Normal, although due to the fact it is literally Eevee and inherently cannot be an evolution of itself, we’re not going to write that off as an unused type. You can’t have an Eeveelution without the lution. That one’s going in the yearbook.

What the above means is that there are still ten types that have yet to receive an Eeveelution: Ground, Rock, Steel, Bug, Poison, Flying, Fighting, Dragon, Ghost, and the aforementioned Normal. Our mission was to prove which of these types would hypothetically have the largest impact on a new game, which is of particular interest now that Pokémon Scarlet & Violet have been revealed. Let’s get cracking.

The first point of order was to analyse existing stat spreads in order to determine what kind of builds have yet to be covered by Eeveelutions. The table below outlines the highest, second-highest, and third-highest stats of all eight known Eeveelutions, which are unanimously 130, 110, and 95 regardless of designation.

We then calculated the relative value of each individual stat by using a three-point system in which stats were assigned either one, two, or three points depending on how often they appeared as the third-highest, second-highest, or highest total. Phew.

Special Attack and Special Defense are by far the most commonly high stats across all Eeveelutions, clocking in with a whopping 12 points. Meanwhile, Defense earned seven, Speed earned six, and both HP and Attack earned a comparatively meager five. This means that the most balanced stat spread to introduce in order to maintain consistency as per the inherent function of Eeveelutions would be either HP/Attack/Speed or Attack/HP/Speed. Step one of the investigation: complete.

Step two of our research involves examining type matchups among the Eeveelutions to design a sort of inter-Eeveelution type-effectiveness chart. We won’t bore you with laborious details — you can just view the chart for yourself below.

What this demonstrates is an aggregate value for how many strengths and weaknesses each individual Eeveelution has in relation to the seven remaining ones. From this data, we can see that Jolteon, Flareon, and Sylveon are the most overpowered Eeveelutions, while Vaporeon, Espeon, and Glaceon are the weakest. This is very obviously a myopic exercise when viewed independently — for example, we all know that Vaporeon is the best Eeveelution. As part of a wider study though, it’s a useful measure to include. 

This is where things start to get interesting. In order to make better use of the data above, we ran a more specific version of the type matchup exercise to determine how many of the ten unused types were strong or weak against overpowered or underpowered existing Eeveelutions. Once again: phew.

As you can see from the table above, a Ground-type Eeveelution would have the biggest impact on inter-Eeveelution battle dynamics. Second-place goes to Poison, while the only other type to land on the positive side of our aggregates is Rock. We are now left with just three viable options for an Eeveelution that could be meaningfully integrated into the fold — so let’s run the exercise again, except this time we’re specifically comparing them to the until-now unmentioned neutral aggregate Eeveelutions, Umbreon and Leafeon. In this version of the exercise, points only count for half.

The only type to benefit from its amended total is Poison, although Ground still manages to come out on top. Considering the results more broadly, a Ground-type Eeveelution would have the most significant effect on how Eeveelutions respond to one another. If a second Eeveelution were to be added, Poison has the second-highest amended total and would also serve as a check to Leafeon’s newfound super-effectiveness towards and resistance against Ground. Rock errs on the side of irrelevance and would arguably make Ground too strong, so it’s out for now.

The final part of the investigation applies this new data to our first exercise: Assessing logical stat spreads based on the ground (eh?) that has already been covered. This means that a Ground-type or Poison-type Eeveelution would ideally pool points in either Attack/HP/Speed or HP/Attack/Speed.

Fortunately, both of these options are viable for both types. Given Ground’s +4 aggregate total prior to amendment, it would make a particularly good candidate for Attack/HP/Speed, allowing it to lean into its offensive potential. By that logic, Poison could satisfy the last remaining hole with a HP/Attack/Speed build. You’re welcome, Game Freak.

It’s also interesting to think about this hypothesis in terms of the wider meta. Ground and Poison-type Eeveelutions could check Steel-types and Fairy-types, respectively, both of which are arguably the strongest single types in the game right now. It would be nice to see one of Eevee’s various forms carving itself out a spot in the modern metagame.

To conclude: The next Eeveelution should be Ground-type, followed closely by a Poison-type. Popular fan names for these types include Terreon and Toxeon. We reckon they sound alright.

Written by Cian Maher on behalf of GLHF.