New defender rankings based on *real* gym battles (not simulations). Alpha testing (V0)

Note: Please read explanations before coming to any conclusions about this.

Introduction

You may remember me for my Quantitative gym defender rankings series  where I used simulations to rank defenders while accounting for type based on survey data. I am now taking this to a new level. These are the first rankings, as far as I know, that use actual gym battles. In these rankings I sample hundreds of gym battles from players of all levels ranging from novice to hardcore and assign the defender quotient (the same quotient from my previous rankings) to defenders. The defender quotient takes into account both time wasted and damage done.
Since this uses actual gym battles it is the only ranking of its kind that accounts for the autoselector, the full range of dodging strategies, lag, the death loop and more!
Because of how ambitious this project is, I am only ranking the big 7 defenders(anything with a CP higher than Vaporeon) because it is impractical to collect data on rare gym Pokémon. As of now, although I have hundreds of gym battles I still do not have enough to rank each Pokémon by its moveset (but I will, stay tuned). I am releasing prelim results now in case Niantic changes before I have collected all the data.

Methodology

I calculate defender quotient for defenders by examining real battles uploaded to YouTube by Pokémon Go players. In order for battles to be eligible they must meet the following criteria:

1. Have occurred after the latest moveset update
2. Be 1 participant only
3. Not ‘show’ battles with deliberate selection of esoteric pokemon (like magikarp or togepi) for the video.
4. Not mass produced by a single very high level player (PokeAk was disqualified). Note expert players with near perfect dodging skills and extremely high levels are still included.
5. Not associated with hacking, spoofing, or click bait/spammy type content.
6. Attacking only, no training.

Results

Note: Tyranitar was not included as there is currently not enough data to justify it. This ranking only contains the Big 7 minus Tyranitar.

	Position	Pokemon	Defender Quotient
	1	Blissey	160
	2	Dragonite	59
	3	Snorlax	56
	4	Vaporeon	54
	5	Gyarados	53
	6	Rhydon	43

Explanation (please read before jumping to conclusions)

I was also very surprised by these results, the reason they are so different to our preconceptions is because most players do not play with anywhere near as much skill as those on silph road. About 20% of the base, consistently play very well and make good selections but this is not true of the majority. I will explain by each Pokémon.

Blissey

More than 50% of attempts time out on Blissey. She is a huge problem for most of the player base. However, despite this, most players make reasonable attacker choices with most opting to go for Dragonite, Rhydon or Flareon. A significant fraction of players also use Machamp, this is especially true of the higher skilled players.

Dragonite

Very poor choices are generally made when it comes to Dragonite. Highly skilled players reach for Lapras or Cloyster but this is, unfortunately, in the minority. There are a suprising number of players that use Machamp against Dragonite. Vaporeon is also a popular choice but many players do not optimize at all for Dragonite and just go with whatever they are given. It seems many players are still not aware of how weak Dragonite is to ice. Perhaps worse, a minority of players were content to use grass types against dragonite!

Snorlax

Snorlax has a low defender quotient because it was the single most well optimized pokemon here. Even relatively low level players knew to go for Machamp and, in fact, more players chose to use Machamp against Snorlax than even Blissey! Snorlax has a theoretical defender quotient at around 25, so about 50% of the potential is still being lost, but that is nothing compared to 75% lost for dragonite and near 80% for Gyarados! I suspect attacker tier lists showing machamp as the best Pokémon play a role here. Snorlax is tackled much better than anything else.

Vaporeon

Most players know that Vaporeon is weak to grass and electric, but only a few seem to realize that the former is almost always better than the later. Too often do people opt for Jolteon against Vaporeon rather than a high level grass type. However, overall, this was a relatively well tackled defender. Dodging, though, was very poor against Vaporeon with Hydro pump’s short charge time being a problem for many players and aqua tail tending to go off when players used their charged attacks.

Gyarados

Performance against Gyarados was extremely poor. This was, unequivocally, the worst handled defender of all. It was quite astounding how few players opted to choose Jolteon or another high level electric type. Instead, a mixture of grass types and water types were preferred. Dodging on Gyarados was also fairly poor. Hydro pump and twister proved to be challenging and many players did not attempt to dodge any quick moves. However, skilled players were extremely effective against Gyarados with most knocking it down in under 30 seconds with hardly any damage taken.

Rhydon

Rhydon was generally handled well, most players opted to choose a strong grass or water type with Vaporeon being a very popular choice. A significant minority of players also chose Machamp which also functions well against Rhydon. It is notable that most players opted to go for dodge all strategies against Rhydon with Mud slap and Rock smash dodged effectively even by low level players. Stone edge was extremely challenging but dodging was generally very good against rhydon.

Future

My aim is to release the set of rankings with movesets included, together with further reports of how players coped against the most common defenders. I’d be happy to answer any questions 🙂