Number of players per class

[Atenhara.2947]

I am curious how accurate the gw2 efficiency site is, https://gw2efficiency.com/account/player-statistics .

It shows the guardian as the most played class with the engineer last. Granted the numbers are only 3% apart but it just seems odd since I see squads with mostly engineers in them fairly often. What's really telling tho is the number of hours per class.57 million hours played as engineer versus 135 million played by guardians. Is it possible to infer from these two numbers that engineers are the popular flavor of the month but that they don't get played for very long at a time?

[Erich.1783]

A few things, there are less than 400k registered accounts, so whole it is a a reasonable sample size, but it is also weighted to a certain type of player.  Also, those numbers may be influenced by things that more profit focused players do, like the key farm, which use to be done on a human guardian, multiple times a day before that change.  That is just one example I am sure there are other things that through the numbers off.  I guess my point is that in looking at any numbers from efficiency you need to think about the causes behind those numbers.

[Harfang.1507]

For engineers, there are two periods: before mechanists and after mechanists. The numbers gw2efficiency works off probably are for the total number of hours played on any given profession for all registered accounts it has access to. (That is my personal hypothesis, no more, no less.) Since that includes about nine years of data before mechanists came out, that is no surprise. If data were limited to the last six months, you can bet it would be an entirely different story.

[Serephen.3420]

GW2 efficiency isnot used by as many people as it used to be I feel. It would also hold old data of people who don't ay anymore and not necessarily any of more recent players. 

I'd say a site like wingman is more reliable than gw2efficiency for that kind of data. 

[Gehenna.3625]

3 hours ago, Serephen.3420 said:

I'd say a site like wingman is more reliable than gw2efficiency for that kind of data. 

I'd say the opposite. Wingman analyses data from parsers. Considering that only a very small segment of the population uses these and geared towards challenging content, I don't see it as more reliable except for a very specific group of people. Much more specific than those who use gw2efficiency imo.

[Astralporing.1957]

6 hours ago, Atenhara.2947 said:

I am curious how accurate the gw2 efficiency site is, https://gw2efficiency.com/account/player-statistics .

It shows the guardian as the most played class with the engineer last. Granted the numbers are only 3% apart but it just seems odd since I see squads with mostly engineers in them fairly often. What's really telling tho is the number of hours per class.57 million hours played as engineer versus 135 million played by guardians. Is it possible to infer from these two numbers that engineers are the popular flavor of the month but that they don't get played for very long at a time?

No. Remember, that efficiency data is taken over a veeeeery long time. Those stats you mention are impacted by that.

For example, among all my characters, according to playtime my engi is tenth. And yet currently i play mostly on it. The second character i use (and one i played amost exclusively before i switched to mech) is my virtuoso - and on that list it's eleventh. On top of my list is one of my guardians - but not the one i do occasionaly play content with, but the one i use to daily harvest my home instance (and haven't been doing any content on since HoT). Second is my bearbow, which was my main once, but nowadays i only do story content with and is otherwise unplayed. Third is my ele, and it's another character i haven't actively played on since HoT and is only being used for crafting now.

(notice, btw, how the two top characters, due to how i currently use them, accrue a lot of AFK time).

So, in short - stats can be mightily confusing if you're not very careful about interpreting them.

[draxynnic.3719]

Another 'lies and statistics'-style conversation? 😝

 

I'd probably be inclined to say that wingman is probably closer to being useful in reference to the current state. Partially because it gives records associated with particular timeframes rather than accrued over the game's life, and partially because it's more reflective of what people are likely to bring to more difficult group content.

It still isn't completely accurate, but it is more informative on what you'd expect to see in squads in the present balance.

[Gibson.4036]

Except wingman is even more niche than GW2Efficiency in the types of players who connect with it.

Short answer is we don’t have a good way to tell how much any given class is being played recently.

Even the impression that “mechs are everywhere” is likely skewed by novelty bias and the size and color of mechs.

[Fueki.4753]

If you are trying to get a view of the entire playerbase, gw2efficiency is not accurate in the slightest.

Aside from hypothetical official data (which would need to be provided by Arenanet themselves), there's no chance of accurate data at all.

[ugrakarma.9416]

10 hours ago, Atenhara.2947 said:

I am curious how accurate the gw2 efficiency site is, https://gw2efficiency.com/account/player-statistics .

It shows the guardian as the most played class with the engineer last. Granted the numbers are only 3% apart but it just seems odd since I see squads with mostly engineers in them fairly often. What's really telling tho is the number of hours per class.57 million hours played as engineer versus 135 million played by guardians. Is it possible to infer from these two numbers that engineers are the popular flavor of the month but that they don't get played for very long at a time?

The gw2data make sense, is just a bit outdated.

Engineer wasnt so popular before mechanist.. the engineer problems is really a very very long history and complicated discussion, to sum up, most engineer patches are 8 or 80... specially after scrapper, some patches make it excessively strong, them ppl ask for nerf, that made it near useless again, then repeat.

Edited October 24, 2022 by ugrakarma.9416

[Khisanth.2948]

12 hours ago, Erich.1783 said:

I guess my point is that in looking at any numbers from efficiency you need to think about the causes behind those numbers.

True for any statistic.

1 hour ago, Fueki.4753 said:

Aside from hypothetical official data (which would need to be provided by Arenanet themselves), there's no chance of accurate data at all.

That wouldn't be much better because there isn't really a good way to determine what the characters are actually being used for. Thief would be one of my most played classes but a lot of that is for daily crafting, home instance gathering and guild hall gathering. Sometimes other materials too since she has my glyph of bounty.

Necro, ranger, engi and formerly rev stats are also potentially heavily polluted by bots.

[Danikat.8537]

Another thing to consider: how many engineers did you typically see before EoD was released? How many non-mechanist engineers do you see? How many of the players you see using outfits are guardians?

Are engineers really more common or are they just more noticable since the mechanist was released because it's currently a very popular spec and the jade bot is impossible to miss.

I don't know the answers to those questions either, but it wouldn't surprise me if a big part of the change is that engineers have become more visible, rather than or as well as being more popular.

[Ashgar.3024]

Prior to EoD not a lot of people played Engineer.

 

Scrapper was a weird mess that only really worked in SPvP for the longest time and while Holo used to be very good it was harder to play than a lot of builds that fit in its niche (namely Dragon Hunter, Reaper and Daredevil) so not as many people bothered.

 

Even nowadays Scrapper is still completely eclipsed by Firebrand pretty much accross game modes, except maybe in WvW and Holo is left to rot in the dump as it wasn't really looked at by the devs despite the fact some recent mechanical changes to the game wrecked it, namely the heavy exposed nerf and also the spotter rework.

 

Meanwhile Guardian was always a top tier choice in every game modes since the inception of the game and doubt that will ever change both due to the nature of the class and the nature of the game.

 

But yeah, despite the "Mecanist effect", you can't exactly shift nearly a decade of data in 8 months. Would be interesting to be able to parses the data within the timeline though.

 

 

Edited October 24, 2022 by Ashgar.3024

[draxynnic.3719]

On 10/25/2022 at 1:40 AM, Khisanth.2948 said:

True for any statistic.

That wouldn't be much better because there isn't really a good way to determine what the characters are actually being used for. Thief would be one of my most played classes but a lot of that is for daily crafting, home instance gathering and guild hall gathering. Sometimes other materials too since she has my glyph of bounty.

Necro, ranger, engi and formerly rev stats are also potentially heavily polluted by bots.

Unless they actually do have statistics on what people are playing in endgame content.

If that quote of 20% of people being on Mechanist was across all modes and not just endgame PvE content... well, that's a pretty insane oversaturation.

[Astralporing.1957]

14 hours ago, draxynnic.3719 said:

Unless they actually do have statistics on what people are playing in endgame content.

If that quote of 20% of people being on Mechanist was across all modes and not just endgame PvE content... well, that's a pretty insane oversaturation.

If. We have no way of knowing. Which brings up again my point that statistical data is extremely easy to misinterpret if you don't have the whole context of it (and sometimes even then).

Is it across all modes? Across only endgame modes? Is it the same across all endgame modes, or are there variations (i.e. between 10-man and 5-man, between raids and strikes, between IBS and EoD strikes, between normal and CM)? Is it across all attempts? If so, how it differs from only succesful ones? From only unsuccesful ones?

Notice, how each of those scenarios has a different impact and presents a slightly (or sometimes even more than that) different picture. Without that kind of context, we can only play the guessing game, and likely end up with heavily biased interpretations.

[draxynnic.3719]

Well, yeah. I put that 'if' in for that reason.

With that said, I do think that if it's across the entire game, that's indicative of a much broader oversaturation than seeing a similar percentage in competitive modes or endgame PvE. Partially because the proportion of players in that content is probably still relatively low. More importantly, though, they're also the players that are more likely to gravitate to the most effective builds. So if that 20% figure is coming out of the casual open world and story players where build efficiency generally isn't a major concern (I tend to roll my eyes at 'everything works in open world' claims, but the more casual players probably tend to avoid activities where that isn't true), that's probably indicative of a much bigger problem than if we were talking about, say, 20% of endgame PvE players. Mind you, the solution in that case might be less 'nerf mechanist' and more 'introduce easier-to-play builds for other professions', but my previous post, while not going into the detail I have here, was expressing that a changing assumption regarding what the 20% figure was referring to can greatly alter the interpretation.

To be even more specific, 20% in endgame content can mostly be explained by being the preferred alacrity supplier while also having DPS builds. An appropriate response here might be to nerf alacmech or buff other alacrity suppliers so they can better compete. 20% in ALL content...unless it's just that people really, really like pet robots, that would probably be indicative of a much deeper problem.

[zeyeti.8347]

Elementalist is Op for fishing thats why those numbers .

[Nico.9361]

Couple of things:

- it shows players who simply own an engineer and guardian. I know many ppl who own both but don't really play both (me)

- the playtime is overall where we've had core guard being very popular and FB being an option couple years longer than mech

Taking statistics out of context again, huh?

[IAmNotMatthew.1058]

On 10/27/2022 at 7:35 PM, draxynnic.3719 said:

If that quote of 20% of people being on Mechanist was across all modes and not just endgame PvE content... well, that's a pretty insane oversaturation.

The strongest is always going to be played a lot until something better comes along. Look at Chrono and FB.

[Ausar.9542]

the data from there includes pre EoD as well, and Engineer was definitely not popular outside of niche in certain areas such as WvW squads.  Especially when you considered pre HoT  GW2, The Engineer was very unpopular since it had the fewest Weapon options in a Fashion based endgame. 

[draxynnic.3719]

10 hours ago, IAmNotMatthew.1058 said:

The strongest is always going to be played a lot until something better comes along. Look at Chrono and FB.

Chrono having a monopoly in endgame boon support didn't translate into an overpresence in other parts of the game since that build didn't really function for anything beyond perma quickness and alacrity. Firebrand is a bit fuzzier because quickbrand can be translated into other parts of the game, but I never had the impression that one out of five players in the open world was playing it.

Now, as I acknowledged earlier, we don't know if that's precisely what the 20% cited by ArenaNet is measuring. The point I was making, though, is that if one out of 27 elite specialisations is getting 20% representation even in a relatively build-agnostic area of the game, that represents a deeper problem than in endgame content where something can become dominant even if it's just a few percentage points more effective than the competition.

[Blood Red Arachnid.2493]

There is something that I haven't quite figured out yet.  There's a reddit thread from seven months ago:  

 

 

Unfortunately, this person didn't actually list how they tracked this information.  Was there an API or something simple that I just have not thought of?  Best I can think of is to take a snapshot, then take another snapshot a week later and do the math on the difference in hours played.  

[draxynnic.3719]

3 minutes ago, Blood Red Arachnid.2493 said:

There is something that I haven't quite figured out yet.  There's a reddit thread from seven months ago:  

 

 

Unfortunately, this person didn't actually list how they tracked this information.  Was there an API or something simple that I just have not thought of?  Best I can think of is to take a snapshot, then take another snapshot a week later and do the math on the difference in hours played.  

Considering that they say 'difference in hours played', I think that's exactly what they did. Take the second day's number, subtract the first, and get how much the profession has been played between those two measurements.

I'd personally say that to get an unbiased sample you'd probably want to do it over a week during a period without a festival, but it is a decent way to get a snapshot.

[Drakz.7051]

I assume this is from since the game started so I can see it being right for that time scale.