About Rankings

PVP IV shows Pokemon rankings for Great League, Ultra League, Master League, and any current cups, e.g. Premier Cup. I recommend viewing these rankings in the app when possible, since you can click in to see details and run simulations for individual matchups. However, they are also made available at pvpiv.com.

Keep in mind that like any other ranking system, these rankings are somewhat arbitrary and subject to change based on changes to the algorithm described here, changes to the meta, and changes introduced in Pokemon Go.

You can use these rankings to get a general idea of which Pokemon fall into which tier, but I would avoid assuming that one Pokemon is always better than another because it’s currently ranked higher. Even gaps of 10 or more points can be closed by relatively minor changes, especially if the gap is between two Pokemon with different types/roles.

Scoring

Each Pokemon gets a score; the Pokemon with the highest score gets the highest rank and vice versa. A Pokemon’s score is based on simulated battles between it and every other Pokemon in the rankings. The simulator’s algorithm is described here.

The score a Pokemon gets for a matchup is the sum of its score for each even shielding scenario. Each even shielding scenario is given a score:

• 1: The Pokemon always wins.
• 0: The Pokemon sometimes wins and sometimes doesn’t, or always ties.
• -1: The Pokemon always loses.

The above score is multiplied by the lag turns for the shielding scenario (up to 10) then divided by 10 to get the final score for that shielding scenario. Lag turns are described in the simulator’s algorithm linked above.

Finally, to make the scores easier to visualize, they’re normalized so that the top ranked Pokemon scores 100 and the lowest ranked Pokemon scores around 60. Note that this only impacts the Pokemon’s score displayed in the app, and not its ranking relative to other Pokemon.

Note that this system is geared toward simplicity. Each Pokemon is weighted equally; winning against a highly-ranked Pokemon isn’t worth more points than winning against a lower-ranked one. Similarly, each shielding scenario is weighted equally; winning the 1 shield is worth the same as winning the 2 shield.

Here is my reasoning for these scores and weights:

• Uneven shielding scenarios are by definition unfair, and thus are excluded to simplify the ranking process and produce output more quickly.
• Even shielding scenarios are all weighted equally.
  • In the past, weights were uneven, but this weighting felt too arbitrary. The best Pokemon can perform well in any shielding scenario, and adapt their shield usage to set up weaker teammates with the shielding scenarios they perform best in.
  • While being able to win with shields but not without them often leaves you with a shield deficit, the inverse is often just as bad; being able to win without shields but not with them means you have no ability to maintain switch advantage.
• Bait-dependent matchups are essentially random, and should not work in favor of either Pokemon’s ranking.
  • This rewards consistent Pokemon that can win without baiting.
• Taking away points for losses – especially those with a large margin of lag turns – rewards safe Pokemon with fewer bad matchups.
• Lag turns are a good indicator of how likely an outcome is to be reproducible in a real battle with different IVs, initial damage/energy differences, etc. They’re often also a good indicator of how much a Pokemon can overfarm before it wins, which helps to account for the value of stored energy.
• Weighing more powerful Pokemon more heavily would be very arbitrary (who decides these weights?), and giving each Pokemon the same weight makes differences in overall win ratio more obvious than if each is weighted differently.

Note that there are some downsides to this approach:

• Pokemon that perform better in the larger context of a battle than they do in a 1 vs 1 matchup may be slightly under-ranked if that performance boost isn’t indicated by lag turns, since the rankings only consider 1 vs 1 matchups.
  • For example, if Haunter beats Registeel one shield to zero, it KOs with fast moves and leaves the matchup with enough energy to use Shadow Ball on the next Pokemon, which will usually get the shield back. Additionally, if a team is built around Haunter, it’s usually possible to commit multiple shields on the Haunter to get more value from it. This isn’t accounted for in the rankings.
  • Similarly, Meganium beats Azumarill with enough HP that it can’t just be farmed down by whatever comes in next, and this isn’t accounted for since it wins by very few lag turns.
  • As a final example, Sableye can swing a lot of close matchups with an energy advantage, and a good player will make sure Sableye is given opportunities to battle with an energy advantage. So, its ranking when used correctly may be higher than its ranking in standard 1 vs 1 battles since it would get positive points for close wins that it instead loses points for as close losses.
• Some traits – in particular bulk and favorable typing – are scarce, which makes Pokemon that have them more valuable. This isn’t accounted for in the rankings.
  • Many successful teams have 3 bulky Pokemon, but very few successful teams have no bulky Pokemon. Most Pokemon in the rankings aren’t that bulky, so the ones that do have bulk may be under-ranked.
• The rankings don’t factor in how likely each shielding scenario is to play out, which may mis-represent some matchups.
  • For example, if a Pokemon uses Power-Up Punch twice and the opponent shields both of them right away, that Pokemon will be able to swing some very unfavorable matchups in the 2 shield. However, real opponents are unlikely to shield right away against a Power-Up Punch user.
  • Similarly, good players can use fast charge moves to control which shielding scenarios play out. For example, a Zekrom player will hardly ever play out the 0-shield against Dialga, instead forcing the more favorable 0-1 or 1-0 scenario, neither of which is weighted.

IVs

Each Pokemon in the rankings is ranked based off of its highest stat product IVs. You can find these in the PVP IV app, or learn more about them in this article. A minimum IV (IV floor) of 0 is used unless the Pokemon was never available in the wild, in which case a minimum IV of 5 is used (10 for mythicals that can’t be traded).

In practice, your Pokemon probably won’t have these exact IVs, and neither will your opponents’ Pokemon. This may cause slight differences in how some IV-dependent matchups used in the rankings play out in real battles, and thus create some margin of error in the rankings. However, the use of lag turns should help to offset this, since matchups with a small margin of error should be decided by very few lag turns.

A minimum IV of 5 is more realistic for legendaries, etc. not available in the wild because their true minimum IV of 1 is only attainable through expensive and rare Good Friend Trades, and for each trade the odds of getting the best rank are less than 1/3000. However, note that the 1 vs 1 battle simulator uses the true IV floor of 1 by default.

Note that some specific Pokemon – like Galarian Stunfisk – have specific IVs that are considered “better” than Rank 1 IVs because they swing certain matchups. In my testing, lower-ranked IVs always ranked lower in the battle rankings as well, even for this type of Pokemon. While there may be exceptions I didn’t find, the clear trend is that Rank 1 IVs are the strongest, and thus are the best IVs to represent the Pokemon in the rankings.

Pokemon Selection

The Pokemon and movesets included in the rankings are hand-picked by me based on my personal experience as a Rank 10 player and avid battler. I chose this approach rather than a more objective approach for the following reasons:

• Battling every Pokemon/moveset against every other one would take a long time.
• Including weak Pokemon that will never be used would skew the rankings in favor of Pokemon that beat them.
  • For example, most fire types are weak and only a few of them are included in my rankings. Including Quilava, Emboar, Infernape, etc. in the rankings would cause the Water, Rock, and Ground types that beat them to be over-ranked.
  • This could be offset by weighting them less heavily than higher-ranked Pokemon, but picking this weight is just as arbitrary as picking which Pokemon are included in the first place.
• It’s a unique and (in my opinion) nice feature to be able to see relevant Pokemon and movesets without having to filter through a lot of irrelevant ones.
  • For example, you can see the difference between Ice Beam Azumarill and Play Rough Azumarill, without having to see the obvious difference between Bubble Azumarill and Rock Smash Azumarill.
• Ranking a Pokemon based on the power of all its movesets combined (instead of hand-picking movesets) would unfairly punish Pokemon with large movepools.
• Weighing all Pokemon evenly makes the overall power difference between two Pokemon more apparent, even if in practice this power difference can be offset by the current meta/anti-meta.
  • It’s important to remember that virtually every Pokemon included in the rankings – even those with negative scores – can be useful in some scenarios, depending on what Pokemon the opponent has.

However, there are some drawbacks as well.

• If you use a Pokemon that’s not included in the rankings, you have to do a 1 vs All battle with it to see its score, and it doesn’t impact the score of other Pokemon.
• Weighing lower-ranked Pokemon as much as higher-ranked Pokemon means that changing a small number of barely-relevant Pokemon can have large implications for the rankings of the top Pokemon.
  • Each Pokemon can cause up to 6 points difference between the scores of two other Pokemon.
• Typings with many relevant Pokemon – and Pokemon with many relevant movesets – may be weighted too heavily in the rankings.
  • For example, Registeel is very popular but only has one relevant moveset and no shadow form, so it may be under-weighted. On the other hand, Hypno has at least 3 relevant movesets (each of which is, to be fair, very different) and a shadow form, so it may be over-weighted.
• The accuracy and completeness of the rankings depends to some extent on my personal understanding of the meta.

Hopefully this article has clarified how rankings work. If you have any questions, you can always feel free to contact me.