Guild Wars 2

IndigoSundown.5419:

And having an understanding of statistics tells us that Vinceman is correct. A biased sample is unlikely to be representative of the entire population. All GW2E can tell us is what percentage of people who signed up for GW2E have LI’s or MS’s.

And those who joined GW2E aren’t “random” enough?
I think people are underestimating how many are using GW2E for their daily activities, inventory management, crafting guides and not for hardcore play like Raids. There is zero evidence that the GW2E population is mostly a group of hardcore players

Ayrilana.1396:

So you’re telling me that the player population on that website isn’t a large enough sample size?

Of course, it could and would be a large enough sample size but nobody knows how biased that sample is. That is the problem we are facing here.

Ayrilana.1396:

You’re telling me that only a very specific segment of the population is attracted to that website? Any evidence to support that?

No, we aren’t telling that but we also have no evidence if that isn’t the case.
We have to be careful when interpreting data.

maddoctor.2738:

And those who joined GW2E aren’t “random” enough?

The fastest thing that comes to my mind is that the overall “casual” (hate to name it like that) is obviously not using gw2efficiency because he/she doesn’t even know that this site even exists.
Additionally, the average player will have one game account while it is most likely that veteran players with additional accounts using gw2efficiency will use it for all of their accounts.
So, even these two little examples can bring in a huge bias into the system.

The thing is, we don’t know anything.

Vinceman.4572:

Ayrilana.1396:

So you’re telling me that the player population on that website isn’t a large enough sample size?

Of course, it could and would be a large enough sample size but nobody knows how biased that sample is. That is the problem we are facing here.

Ayrilana.1396:

You’re telling me that only a very specific segment of the population is attracted to that website? Any evidence to support that?

No, we aren’t telling that but we also have no evidence if that isn’t the case.
We have to be careful when interpreting data.

You’ve given no evidence that it is biased. Just because something has the chance to be biased doesn’t mean that it is so otherwise practically all surveys would be unreliable.

Assuming that players who have not beaten an encounter cannot consume shards, you can realiably use that as a metric to determine how much effort they have put into the raids. It becomes unreliable once players have access to spend them.

LI’s will become unreliable the second legendary armor becomes available. That is unless there’s a way to create a lifetime accumulation metric.

One limitation of that website is that it’s only for a specific point in time and doesn’t appear to track historical trends.

Ayrilana.1396:

Vinceman.4572:

Ayrilana.1396:

So you’re telling me that the player population on that website isn’t a large enough sample size?

Of course, it could and would be a large enough sample size but nobody knows how biased that sample is. That is the problem we are facing here.

Ayrilana.1396:

You’re telling me that only a very specific segment of the population is attracted to that website? Any evidence to support that?

No, we aren’t telling that but we also have no evidence if that isn’t the case.
We have to be careful when interpreting data.

You’ve given no evidence that it is biased. Just because something has the chance to be biased doesn’t mean that it is so otherwise practically all surveys would be unreliable.

But they are unreliable. We had a good example last year, on surveys Trump was sure to lose elections, but the surveys didnt count if the people would actually go to vote, people just said “I will vote for X not trump” and these people never went to vote so Trump won.
Its the same with gw2effiency, you dont know who goes there, so taking it as true does more harm then good.

Ayrilana.1396:

Vinceman.4572:

Ayrilana.1396:

So you’re telling me that the player population on that website isn’t a large enough sample size?

Of course, it could and would be a large enough sample size but nobody knows how biased that sample is. That is the problem we are facing here.

Ayrilana.1396:

You’re telling me that only a very specific segment of the population is attracted to that website? Any evidence to support that?

No, we aren’t telling that but we also have no evidence if that isn’t the case.
We have to be careful when interpreting data.

You’ve given no evidence that it is biased. Just because something has the chance to be biased doesn’t mean that it is so otherwise practically all surveys would be unreliable.

Sorry, but as a natural scientist I have to be skeptical and have to question (and every serious one is and has) when using data with almost 0 background information. Everything else would be highly disputable. And the data of gw2efficiency doesn’t give us enough of this background info.

OnizukaBR.8537:

Its the same with gw2effiency, you dont know who goes there, so taking it as true does more harm then good.

Excatly.

maddoctor.2738:

OriOri.8724:

https://gw2efficiency.com/account/statistics/statistics.legendaryInsights

The key with Raids is to identify what a “raider” is first and then compare those players with those who could be running them in the first place.

This is 100% true. Without establishing what a Raider is you can’t say who big the community is

@maddoctor – I agree with you, its just meant as a baseline. It shows that, at best, less than 1/3 of the community raids. Which is what I said in my original post.

Vinceman.4572:

OriOri.8724:

representative of GW2 as a whole

I disagree here. We already had discussions here why it is not sure to be representative of GW2 as a whole.

We don’t know if gw2efficiency is problably used more often by raiders than others or if it’s the other way round. There are indications for both.

These numbers are numbers but they actually are of very little value.

Not true. This is where statistics come in. With such a large sample this is still useful data, if the sample was only a couple thousand then yes, it wouldn’t be of much use. But with over 110 thousand accounts, there are enough accounts on GW2Efficiency to be useful in a statistical sense.

And this still doesn’t change the fact of what I said. At most, less than 1/3 of the community raids. Even if you don’t believe in statistics whatsoever and think that people on GW2Efficiency are more predisposed to raiding than general accounts, what I said still holds true. Because it would still mean that less than 1/3 of the people who play this game participate in raids.

OnizukaBR.8537:

Ayrilana.1396:

Vinceman.4572:

Ayrilana.1396:

So you’re telling me that the player population on that website isn’t a large enough sample size?

Of course, it could and would be a large enough sample size but nobody knows how biased that sample is. That is the problem we are facing here.

Ayrilana.1396:

You’re telling me that only a very specific segment of the population is attracted to that website? Any evidence to support that?

No, we aren’t telling that but we also have no evidence if that isn’t the case.
We have to be careful when interpreting data.

You’ve given no evidence that it is biased. Just because something has the chance to be biased doesn’t mean that it is so otherwise practically all surveys would be unreliable.

But they are unreliable. We had a good example last year, on surveys Trump was sure to lose elections, but the surveys didnt count if the people would actually go to vote, people just said “I will vote for X not trump” and these people never went to vote so Trump won.
Its the same with gw2effiency, you dont know who goes there, so taking it as true does more harm then good.

There were other issues that were preventing people from being open about their choice but this isn’t the place to discuss that. Surveys are most often reliable if done correctly.

maddoctor.2738:

IndigoSundown.5419:

And having an understanding of statistics tells us that Vinceman is correct. A biased sample is unlikely to be representative of the entire population. All GW2E can tell us is what percentage of people who signed up for GW2E have LI’s or MS’s.

And those who joined GW2E aren’t “random” enough?
I think people are underestimating how many are using GW2E for their daily activities, inventory management, crafting guides and not for hardcore play like Raids. There is zero evidence that the GW2E population is mostly a group of hardcore players

Exactly. The most bias I see here is people trying to force their imaginary biases onto the community in GW2Efficiency. It existed before Raids were available, it has dozens of extra, highly useful features that have nothing to do with Raids. It is not designed for raiders. It does not cater to them over any other group of players.

Vinceman.4572:

Ayrilana.1396:

Vinceman.4572:

Ayrilana.1396:

So you’re telling me that the player population on that website isn’t a large enough sample size?

Of course, it could and would be a large enough sample size but nobody knows how biased that sample is. That is the problem we are facing here.

Ayrilana.1396:

You’re telling me that only a very specific segment of the population is attracted to that website? Any evidence to support that?

No, we aren’t telling that but we also have no evidence if that isn’t the case.
We have to be careful when interpreting data.

You’ve given no evidence that it is biased. Just because something has the chance to be biased doesn’t mean that it is so otherwise practically all surveys would be unreliable.

Sorry, but as a natural scientist I have to be skeptical and have to question (and every serious one is and has) when using data with almost 0 background information. Everything else would be highly disputable. And the data of gw2efficiency doesn’t give us enough of this background info.

OnizukaBR.8537:

Its the same with gw2effiency, you dont know who goes there, so taking it as true does more harm then good.

Excatly.

Not all systems and sources are perfect. Unless you have reasonable evidence to support that there would be bias or errors in the population data of that website, I don’t see any issues with it being used.

OriOri.8724:

Not true. This is where statistics come in. With such a large sample this is still useful data, if the sample was only a couple thousand then yes, it wouldn’t be of much use. But with over 110 thousand accounts, there are enough accounts on GW2Efficiency to be useful in a statistical sense.

No, you are mistaken here. 100,000 paid data points are 100,000 invalid data points when it comes to statistics. Statistics is not just about the sample size, you also need a good sample design.

OriOri.8724:

And this still doesn’t change the fact of what I said. At most, less than 1/3 of the community raids. Even if you don’t believe in statistics whatsoever and think that people on GW2Efficiency are more predisposed to raiding than general accounts, what I said still holds true. Because it would still mean that less than 1/3 of the people who play this game participate in raids.

I’ve never questioned the part “less than 1/3 of the community raids” but it would sound noncredible if the actual number is less than 1/30.

Ayrilana.1396:

Not all systems and sources are perfect. Unless you have reasonable evidence to support that there would be bias or errors in the population data of that website, I don’t see any issues with it being used.

See above, keyword sample design.

zealex.9410:

I believe it would be apropriate for anet to release some numbers about how big is the raiding community

No, it is not appropriate.

• ANet has never released numbers, even percentages about participation in the game or any portion of the game’s content.
• Releasing numbers without context is asking for trouble. Without knowing goals, targets, industry benchmarks, it’s very difficult to make any statements about whether numbers are robust or even whether they are sufficient to justify a dedicated team.

I’d love to see ANet, as well as Blizzard and other companies, release actual numbers about games owned, active players, peak and trough concurrency for the game and each of its modes, and so on. But as long as they don’t publish any of that information, releasing any single metric won’t be helpful to understanding or discussing the game.

Vinceman.4572:

OriOri.8724:

Not true. This is where statistics come in. With such a large sample this is still useful data, if the sample was only a couple thousand then yes, it wouldn’t be of much use. But with over 110 thousand accounts, there are enough accounts on GW2Efficiency to be useful in a statistical sense.

No, you are mistaken here. 100,000 paid data points are 100,000 invalid data points when it comes to statistics. Statistics is not just about the sample size, you also need a good sample design.

OriOri.8724:

And this still doesn’t change the fact of what I said. At most, less than 1/3 of the community raids. Even if you don’t believe in statistics whatsoever and think that people on GW2Efficiency are more predisposed to raiding than general accounts, what I said still holds true. Because it would still mean that less than 1/3 of the people who play this game participate in raids.

I’ve never questioned the part “less than 1/3 of the community raids” but it would sound noncredible if the actual number is less than 1/30.

Ayrilana.1396:

Not all systems and sources are perfect. Unless you have reasonable evidence to support that there would be bias or errors in the population data of that website, I don’t see any issues with it being used.

See above, keyword sample design.

You’ve given no evidence that the sample design is impaired enough that it cannot used with a reasonable degree of accuracy.

Rednik.3809:

BlaqueFyre.5678:

Can a player not create their own Raid group? Can a Player not post their own LFG? Can a Player not lead their own Group? Can a Player not look for an actual Raid training Guild? Can a Player not join a Static Raid Guild? Can a Player not group with similar minded people to them? Can a Player not take their own initiative?

Exactly that list was addressed in Bastion of Penitent topic, yesterday. No, he cannot because this list is not working.

Iirc since the posts don’t seem to bein the BoP thread anymore, he complained he had other obligations IRL that took priority, and he said he plays normally early morning and that no Guilds/few groups Raid at that time. Which again his IRL obligations have zero importance when it comes to this games designs and is not a fault of the games design.

I doubt he exhausted all avenues to make his own Guild/group or to join a static Raiding Guild/Group, since there are plenty of EU and NA based guilds on either region that Raid at most hours of every day, I should know I raided with an EU/OCX based Guild on NA servers when I used to play at 0200 MTD do to my schedule, guess what *I*made it work since I wanted to raid. Again the player you are referring to made excuses on why he couldn’t Raid not that he wasn’t allowed to, not that the raids are inaccessible, he was stopped by his own Self imposed limitations and lack of Initiative to find/create a Static group that works with his needs and schedule.

Again there are only Self imposed Accessibility issues, every single player can create and lead their own Raids, can look for an try to find Guilds that Raid and meet their needs, can adjust their schedules to in game or irl to accommodate their need to Raid, nothing outside of the Players own Self imposed limits stops this from happening.

All a player needs to Raid are

Time
Adaptability
Teamwork
Being Social
Basic understanding of class and mechanics
Putting forth Initiative
Level 80
At the extreme minimum Rares/Exotic Gear

No other requirements are needed

All other Accessibility “issues” are just excuses.

Ayrilana.1396:

Vinceman.4572:

OriOri.8724:

Not true. This is where statistics come in. With such a large sample this is still useful data, if the sample was only a couple thousand then yes, it wouldn’t be of much use. But with over 110 thousand accounts, there are enough accounts on GW2Efficiency to be useful in a statistical sense.

No, you are mistaken here. 100,000 paid data points are 100,000 invalid data points when it comes to statistics. Statistics is not just about the sample size, you also need a good sample design.

OriOri.8724:

And this still doesn’t change the fact of what I said. At most, less than 1/3 of the community raids. Even if you don’t believe in statistics whatsoever and think that people on GW2Efficiency are more predisposed to raiding than general accounts, what I said still holds true. Because it would still mean that less than 1/3 of the people who play this game participate in raids.

I’ve never questioned the part “less than 1/3 of the community raids” but it would sound noncredible if the actual number is less than 1/30.

Ayrilana.1396:

Not all systems and sources are perfect. Unless you have reasonable evidence to support that there would be bias or errors in the population data of that website, I don’t see any issues with it being used.

See above, keyword sample design.

You’ve given no evidence that the sample design is impaired enough that it cannot used with a reasonable degree of accuracy.

You are the one that need to prove that this sample design is good enough. By default its is a bad sample.

Its just like i’m saying i have an unicorn, and you have to prove me i’m wrong instead of i’m having to prove that i’m right. Here by default its a lie that i have an unicorn until i prove otherwise.

Ayrilana.1396:

Vinceman.4572:

OriOri.8724:

Not true. This is where statistics come in. With such a large sample this is still useful data, if the sample was only a couple thousand then yes, it wouldn’t be of much use. But with over 110 thousand accounts, there are enough accounts on GW2Efficiency to be useful in a statistical sense.

No, you are mistaken here. 100,000 paid data points are 100,000 invalid data points when it comes to statistics. Statistics is not just about the sample size, you also need a good sample design.

OriOri.8724:

And this still doesn’t change the fact of what I said. At most, less than 1/3 of the community raids. Even if you don’t believe in statistics whatsoever and think that people on GW2Efficiency are more predisposed to raiding than general accounts, what I said still holds true. Because it would still mean that less than 1/3 of the people who play this game participate in raids.

I’ve never questioned the part “less than 1/3 of the community raids” but it would sound noncredible if the actual number is less than 1/30.

Ayrilana.1396:

Not all systems and sources are perfect. Unless you have reasonable evidence to support that there would be bias or errors in the population data of that website, I don’t see any issues with it being used.

See above, keyword sample design.

You’ve given no evidence that the sample design is impaired enough that it cannot used with a reasonable degree of accuracy.

I’ve given two little examples that can easily have a huge and disastrous influence on the actual validity. And there are some more.
Btw. in science it’s on the people who are using the data to show that there isn’t any impairment or conflict involved. Please don’t take the bus Trump fans are driving. ^^

Since you won’t be able to I still hold my point to be carefully with numbers of gw2efficiency.

OnizukaBR.8537:

Ayrilana.1396:

Vinceman.4572:

OriOri.8724:

Not true. This is where statistics come in. With such a large sample this is still useful data, if the sample was only a couple thousand then yes, it wouldn’t be of much use. But with over 110 thousand accounts, there are enough accounts on GW2Efficiency to be useful in a statistical sense.

No, you are mistaken here. 100,000 paid data points are 100,000 invalid data points when it comes to statistics. Statistics is not just about the sample size, you also need a good sample design.

OriOri.8724:

And this still doesn’t change the fact of what I said. At most, less than 1/3 of the community raids. Even if you don’t believe in statistics whatsoever and think that people on GW2Efficiency are more predisposed to raiding than general accounts, what I said still holds true. Because it would still mean that less than 1/3 of the people who play this game participate in raids.

I’ve never questioned the part “less than 1/3 of the community raids” but it would sound noncredible if the actual number is less than 1/30.

Ayrilana.1396:

Not all systems and sources are perfect. Unless you have reasonable evidence to support that there would be bias or errors in the population data of that website, I don’t see any issues with it being used.

See above, keyword sample design.

You’ve given no evidence that the sample design is impaired enough that it cannot used with a reasonable degree of accuracy.

You are the one that need to prove that this sample design is good enough. By default its is a bad sample.

Its just like i’m saying i have an unicorn, and you have to prove me i’m wrong instead of i’m having to prove that i’m right. Here by default its a lie that i have an unicorn until i prove otherwise.

Anyone can register to it and there’s no evidence to suggest that there is any bias to those who registers as it has aspects that can appeal to everyone.

Valik Shin.9027:

The raid community is bigger than the pvp community

Possibly, but remember that pvp community is pretty much dying nowadays.

Ayrilana.1396:

So you’re telling me that the player population on that website isn’t a large enough sample size?

Yes. The size is not everything. 100 million would usually be enough sample size for statistics about world population, but if you were to prefilter those, by, say, picking only a population of a single country (US, for example), you might arrive at some hilariously wrong results.
That’s what happening here. You are picking a population of a single country, and treating it as a representative of a whole world.

Ayrilana.1396:

You’re telling me that only a very specific segment of the population is attracted to that website?

Yes. People that use gaming websites, for one. We already know that such bias can be really heavy (remember that one LOtRO thread?).

maddoctor.2738:

IndigoSundown.5419:

And having an understanding of statistics tells us that Vinceman is correct. A biased sample is unlikely to be representative of the entire population. All GW2E can tell us is what percentage of people who signed up for GW2E have LI’s or MS’s.

And those who joined GW2E aren’t “random” enough?
I think people are underestimating how many are using GW2E for their daily activities, inventory management, crafting guides and not for hardcore play like Raids. There is zero evidence that the GW2E population is mostly a group of hardcore players

The act of using a site like this for daily activities already sets them above the average crowd. Most players do not even care enough to use forums, much less more advanced sites. Quite often, because they don’t even know such sites exist.

Astralporing.1957:

The act of using a site like this for daily activities already sets them above the average crowd. Most players do not even care enough to use forums, much less more advanced sites. Quite often, because they don’t even know such sites exist.

First, the forums have much less activity than GW2efficiency.
Second, if they are in a guild they should know such a site exist. If they’ve been to many meta events chances are they’ve heard about it too.
Unless they are an anti-social person who plays completely on their own, or someone who is in a small guild of similar players, chances are someone in their surrounding has used the site (or gaming sites in general) at least once.

Also, about using “advanced sites”, how many times have you been at an event and someone asked when is the event coming and the response was “check the timer”? Or how many times someone in map chat said “use the wiki /wiki”, or “check dulfy” when someone asks something that has a guide. Every single player of GW2 must’ve heard about timers and “advanced websites” at least once while playing the game. This is a game that depends heavily on external websites to work properly because sadly in game timers do not exist and most high end events depend on timers.

Rednik.3809:

Wasabi Kitty.8247:

Raiding guilds already create multiple windows of opportunity for new players to get into raiding.

And raiding community is working hard to close them.

ofc not like wut?

Astralporing.1957:

Valik Shin.9027:

The raid community is bigger than the pvp community

Possibly, but remember that pvp community is pretty much dying nowadays.

Ayrilana.1396:

So you’re telling me that the player population on that website isn’t a large enough sample size?

Yes. The size is not everything. 100 million would usually be enough sample size for statistics about world population, but if you were to prefilter those, by, say, picking only a population of a single country (US, for example), you might arrive at some hilariously wrong results.
That’s what happening here. You are picking a population of a single country, and treating it as a representative of a whole world.

Ayrilana.1396:

You’re telling me that only a very specific segment of the population is attracted to that website?

Yes. People that use gaming websites, for one. We already know that such bias can be really heavy (remember that one LOtRO thread?).

maddoctor.2738:

IndigoSundown.5419:

And having an understanding of statistics tells us that Vinceman is correct. A biased sample is unlikely to be representative of the entire population. All GW2E can tell us is what percentage of people who signed up for GW2E have LI’s or MS’s.

And those who joined GW2E aren’t “random” enough?
I think people are underestimating how many are using GW2E for their daily activities, inventory management, crafting guides and not for hardcore play like Raids. There is zero evidence that the GW2E population is mostly a group of hardcore players

The act of using a site like this for daily activities already sets them above the average crowd. Most players do not even care enough to use forums, much less more advanced sites. Quite often, because they don’t even know such sites exist.

That website works for both NA and EU players so I have no idea at what you’re trying to get at.

maddoctor.2738:

Astralporing.1957:

The act of using a site like this for daily activities already sets them above the average crowd. Most players do not even care enough to use forums, much less more advanced sites. Quite often, because they don’t even know such sites exist.

First, the forums have much less activity than GW2efficiency.
Second, if they are in a guild they should know such a site exist. If they’ve been to many meta events chances are they’ve heard about it too

Being in a big guild creates a bias in itself.

Astralporing.1957:

Being in a big guild creates a bias in itself.

I didn’t say “big” guild, I said guild in general. How many players are in personal or family-only guilds and how many are in actual guilds out there? I think you are trying to create an imaginary “average player” here.

Never ever read “gw2efficiency” in an open world chat or squad chat during meta events tbh.

Ayrilana.1396:

That website works for both NA and EU players so I have no idea at what you’re trying to get at.

That you pick a specific subgroup of players, that is not really taken randomly. That in itself creates a bias. And you have no info about the nature of that bias, so you can’t even try to correct for it in any way.

But if you want a nongeographical example, let’s try this one: you are attempting to estimate spending habits worldwide. Basing it only on statistical data from people that have bank accounts. It just can’t end well.

Seriously, that’s the very basis of statistics. One of the most common and basic errors, that often can have a massive impact on the results. It’s one of the first things you’re being taught to be wary of…

Vinceman.4572:

Never ever read “gw2efficiency” in an open world chat or squad chat during meta events tbh.

What about other “advanced websites”? Dulfy? A timer website? The wiki?

maddoctor.2738:

Vinceman.4572:

Never ever read “gw2efficiency” in an open world chat or squad chat during meta events tbh.

What about other “advanced websites”? Dulfy? A timer website? The wiki?

Wiki, of course. Dulfy and Timer websites rarely.

Vinceman.4572:

maddoctor.2738:

Vinceman.4572:

Never ever read “gw2efficiency” in an open world chat or squad chat during meta events tbh.

What about other “advanced websites”? Dulfy? A timer website? The wiki?

Wiki, of course. Dulfy and Timer websites rarely.

in many cases gw2 ef is far better than the wiki and used more id expect

Astralporing.1957:

Ayrilana.1396:

That website works for both NA and EU players so I have no idea at what you’re trying to get at.

That you pick a specific subgroup of players, that is not really taken randomly. That in itself creates a bias. And you have no info about the nature of that bias, so you can’t even try to correct for it in any way.

But if you want a nongeographical example, let’s try this one: you are attempting to estimate spending habits worldwide. Basing it only on statistical data from people that have bank accounts. It just can’t end well.

Seriously, that’s the very basis of statistics. One of the most common and basic errors, that often can have a massive impact on the results. It’s one of the first things you’re being taught to be wary of…

Exactly what subgroup is being specifically picked? You only have access to the data in aggregate form so it’s not like you can pick and choose.

zealex.9410:

Vinceman.4572:

maddoctor.2738:

Vinceman.4572:

Never ever read “gw2efficiency” in an open world chat or squad chat during meta events tbh.

What about other “advanced websites”? Dulfy? A timer website? The wiki?

Wiki, of course. Dulfy and Timer websites rarely.

in many cases gw2 ef is far better than the wiki and used more id expect

Maybe used more often but I doubt it. Can only speak for my guild, no one uses gw2ef but timers, dulfy and the wiki almost daily. Also to mention: You can use the site without having an account.

OriOri.8724:

Not true. This is where statistics come in. With such a large sample this is still useful data, if the sample was only a couple thousand then yes, it wouldn’t be of much use. But with over 110 thousand accounts, there are enough accounts on GW2Efficiency to be useful in a statistical sense.

So I feel the need to correct you here because what you’ve stated is dangerously incorrect. The simple fact that gw2e has a large number of accounts on it does not make that information useful in a statistical sense as you claim. What matters far more than raw size of your sample is whether that sample is representative of your population. Let me illustrate this with an example.

Polling is a big part of politics today. Polls always report the number of people polled, and usually that number is a couple thousand. The key is that those couple thousand people are carefully selected so that they represent the broader population that the poll is trying to examine. Now imagine that the poll only contacted people making over 100k and living in cities. Would that still produce data representative of the population in a country as a whole? Obviously not, no matter how many extra people they polled because their sampling is flawed.

The same issue exists with trying to use gw2e as a source of useful information. We’re unable to control the sample, and this means that we’re also unable to draw useful information from it. What if it turns out that most raiders don’t use gw2e because they check gear in other ways? Now the numbers on gw2e are too low. Or if it turns out that mostly raiders use gw2e because of other reasons. Now the numbers are too high. We don’t have any way of getting that sampling information and that makes gw2e a completely worthless source for the purpose of gauging the gw2 community as a whole.

Vinceman.4572:

Maybe used more often but I doubt it. Can only speak for my guild, no one uses gw2ef but timers, dulfy and the wiki almost daily. Also to mention: You can use the site without having an account.

I used the second part of my post to counter this point:

Quite often, because they don’t even know such sites exist.

It wasn’t about gw2efficiency in particular.
I doubt the average Guild Wars 2 player doesn’t know external websites, that can help them play the game, do not even exist. In fact I guess that almost every single player of this game knows or has used at least once an external website, be it dulfy, wiki, dragon timers, gw2crafts, gw2spidy, gw2eff or any other such website that exists out there. With some rare exceptions that always exist but they are the exception, not the rule.

If everyone had an account on gw2eff then it’d have way more than 100k accounts. I know players that Raid or run T4 fractals that do not have a gw2eff account. Having an account there doesn’t require them to be hardcore players, there are even some really nasty statistics you can check that show that the gw2eff isn’t in fact “hardcore” at all.

https://gw2efficiency.com/account/statistics/statistics.completedDungeons
28.5% of “everybody” has run a dungeon at least once

https://gw2efficiency.com/account/statistics/statistics.fractalLevel
15% has reached Fractal level 100 and 27.5% has reached T4 in Fractals

I’m sorry but if only hardcore players were making accounts on that website then something is wrong here.

Fay.2357:

OriOri.8724:

Not true. This is where statistics come in. With such a large sample this is still useful data, if the sample was only a couple thousand then yes, it wouldn’t be of much use. But with over 110 thousand accounts, there are enough accounts on GW2Efficiency to be useful in a statistical sense.

So I feel the need to correct you here because what you’ve stated is dangerously incorrect. The simple fact that gw2e has a large number of accounts on it does not make that information useful in a statistical sense as you claim. What matters far more than raw size of your sample is whether that sample is representative of your population. Let me illustrate this with an example.

Polling is a big part of politics today. Polls always report the number of people polled, and usually that number is a couple thousand. The key is that those couple thousand people are carefully selected so that they represent the broader population that the poll is trying to examine. Now imagine that the poll only contacted people making over 100k and living in cities. Would that still produce data representative of the population in a country as a whole? Obviously not, no matter how many extra people they polled because their sampling is flawed.

The same issue exists with trying to use gw2e as a source of useful information. We’re unable to control the sample, and this means that we’re also unable to draw useful information from it. What if it turns out that most raiders don’t use gw2e because they check gear in other ways? Now the numbers on gw2e are too low. Or if it turns out that mostly raiders use gw2e because of other reasons. Now the numbers are too high. We don’t have any way of getting that sampling information and that makes gw2e a completely worthless source for the purpose of gauging the gw2 community as a whole.

There’s no evidence to support one type of player is more likely to sign up to that website than another. if you check the various metrics, you’ll see what appears to support this. Until such evidence surfaces, which establishes that there are biases, you can assume that all player segments are equally represented among the websites linked account population.

Ayrilana.1396:

Fay.2357:

OriOri.8724:

Not true. This is where statistics come in. With such a large sample this is still useful data, if the sample was only a couple thousand then yes, it wouldn’t be of much use. But with over 110 thousand accounts, there are enough accounts on GW2Efficiency to be useful in a statistical sense.

So I feel the need to correct you here because what you’ve stated is dangerously incorrect. The simple fact that gw2e has a large number of accounts on it does not make that information useful in a statistical sense as you claim. What matters far more than raw size of your sample is whether that sample is representative of your population. Let me illustrate this with an example.

Polling is a big part of politics today. Polls always report the number of people polled, and usually that number is a couple thousand. The key is that those couple thousand people are carefully selected so that they represent the broader population that the poll is trying to examine. Now imagine that the poll only contacted people making over 100k and living in cities. Would that still produce data representative of the population in a country as a whole? Obviously not, no matter how many extra people they polled because their sampling is flawed.

The same issue exists with trying to use gw2e as a source of useful information. We’re unable to control the sample, and this means that we’re also unable to draw useful information from it. What if it turns out that most raiders don’t use gw2e because they check gear in other ways? Now the numbers on gw2e are too low. Or if it turns out that mostly raiders use gw2e because of other reasons. Now the numbers are too high. We don’t have any way of getting that sampling information and that makes gw2e a completely worthless source for the purpose of gauging the gw2 community as a whole.

There’s no evidence to support one type of player is more likely to sign up to that website than another. if you check the various metrics, you’ll see what appears to support this. Until such evidence surfaces, which establishes that there are biases, you can assume that all player segments are equally represented among the websites linked account population.

No, you’ve got that backwards actually. Until evidence surfaces that shows equal player representation, you have to assume that an unspecified bias exists in the sample. You can’t just say that any random sample can be trusted. That’s like saying “we haven’t found any evidence that this unspecified pill won’t fix your illness, so we’ll just assume that it works until proven otherwise.” Do you see the problem with that?

maddoctor.2738:

I’m sorry but if only hardcore players were making accounts on that website then something is wrong here.

I don’t doubt that neither. Just saying that gw2ef isn’t reliable because we don’t know the sample design.

Fay.2357:

Ayrilana.1396:

Fay.2357:

OriOri.8724:

Not true. This is where statistics come in. With such a large sample this is still useful data, if the sample was only a couple thousand then yes, it wouldn’t be of much use. But with over 110 thousand accounts, there are enough accounts on GW2Efficiency to be useful in a statistical sense.

So I feel the need to correct you here because what you’ve stated is dangerously incorrect. The simple fact that gw2e has a large number of accounts on it does not make that information useful in a statistical sense as you claim. What matters far more than raw size of your sample is whether that sample is representative of your population. Let me illustrate this with an example.

Polling is a big part of politics today. Polls always report the number of people polled, and usually that number is a couple thousand. The key is that those couple thousand people are carefully selected so that they represent the broader population that the poll is trying to examine. Now imagine that the poll only contacted people making over 100k and living in cities. Would that still produce data representative of the population in a country as a whole? Obviously not, no matter how many extra people they polled because their sampling is flawed.

The same issue exists with trying to use gw2e as a source of useful information. We’re unable to control the sample, and this means that we’re also unable to draw useful information from it. What if it turns out that most raiders don’t use gw2e because they check gear in other ways? Now the numbers on gw2e are too low. Or if it turns out that mostly raiders use gw2e because of other reasons. Now the numbers are too high. We don’t have any way of getting that sampling information and that makes gw2e a completely worthless source for the purpose of gauging the gw2 community as a whole.

There’s no evidence to support one type of player is more likely to sign up to that website than another. if you check the various metrics, you’ll see what appears to support this. Until such evidence surfaces, which establishes that there are biases, you can assume that all player segments are equally represented among the websites linked account population.

No, you’ve got that backwards actually. Until evidence surfaces that shows equal player representation, you have to assume that an unspecified bias exists in the sample. You can’t just say that any random sample can be trusted. That’s like saying “we haven’t found any evidence that this unspecified pill won’t fix your illness, so we’ll just assume that it works until proven otherwise.” Do you see the problem with that?

Nope. If there’s no evidence to suggest that one type of player would use that website over another, then there really isn’t any issue to use that source until evidence comes about that proves differently.

It’s funny though as you’re pretty much saying that all sources are biased by default and must be proven otherwise. Kind of similar to saying that all sources are unbiased by default and must be proven otherwise. A little amusing how that works out, huh? You’re innocent until proven guilty. No, you’re guilty until proven innocent.

That site has been available for years and has features that can appeal to all players. There’s no evidence to suggest that there is a biased towards one segment of the population to use the website over others. Does that mean that it’s not possible for there to be? No. But then we’re not going on 100% certainties.

For WoW even though blizzard doesn’t give data you can easily go to wowprogress and see the size and scale of the raiding community and this is a good site because it doesn’t consider or track LFR.

As for gw2 what’s wrong with seeing the numbers? That data could easily be helpful for the gw2 playerbase to give feedback so that raids can be improved.

Jman.1975:

For WoW even though blizzard doesn’t give data you can easily go to wowprogress and see the size and scale of the raiding community and this is a good site because it doesn’t consider or track LFR.

As for gw2 what’s wrong with seeing the numbers? That data could easily be helpful for the gw2 playerbase to give feedback so that raids can be improved.

Perhaps because it might go against ANets running narrative regarding numbers participating and put them in a position where they would be forced to admit that they frankly cannot afford the resources to make Tiered Raids?

Fay.2357:

OriOri.8724:

Not true. This is where statistics come in. With such a large sample this is still useful data, if the sample was only a couple thousand then yes, it wouldn’t be of much use. But with over 110 thousand accounts, there are enough accounts on GW2Efficiency to be useful in a statistical sense.

So I feel the need to correct you here because what you’ve stated is dangerously incorrect. The simple fact that gw2e has a large number of accounts on it does not make that information useful in a statistical sense as you claim. What matters far more than raw size of your sample is whether that sample is representative of your population. Let me illustrate this with an example.

Polling is a big part of politics today. Polls always report the number of people polled, and usually that number is a couple thousand. The key is that those couple thousand people are carefully selected so that they represent the broader population that the poll is trying to examine. Now imagine that the poll only contacted people making over 100k and living in cities. Would that still produce data representative of the population in a country as a whole? Obviously not, no matter how many extra people they polled because their sampling is flawed.

The same issue exists with trying to use gw2e as a source of useful information. We’re unable to control the sample, and this means that we’re also unable to draw useful information from it. What if it turns out that most raiders don’t use gw2e because they check gear in other ways? Now the numbers on gw2e are too low. Or if it turns out that mostly raiders use gw2e because of other reasons. Now the numbers are too high. We don’t have any way of getting that sampling information and that makes gw2e a completely worthless source for the purpose of gauging the gw2 community as a whole.

You bring up a good point, but I feel its mostly irrelevant. What you describe is intentionally going out and finding your sample, and cherrypicking people to get a certain narrative out of your findings. That is definitely a problem, but GW2Efficiency does not follow that pattern. That site never went out and decided who would make an account. It never cherrypicked people to include in its statistics, so I don’t feel as if your analogy is the least bit relevant to GW2Efficiency.

Ayrilana.1396:

Fay.2357:

Ayrilana.1396:

Fay.2357:

OriOri.8724:

Not true. This is where statistics come in. With such a large sample this is still useful data, if the sample was only a couple thousand then yes, it wouldn’t be of much use. But with over 110 thousand accounts, there are enough accounts on GW2Efficiency to be useful in a statistical sense.

So I feel the need to correct you here because what you’ve stated is dangerously incorrect. The simple fact that gw2e has a large number of accounts on it does not make that information useful in a statistical sense as you claim. What matters far more than raw size of your sample is whether that sample is representative of your population. Let me illustrate this with an example.

Polling is a big part of politics today. Polls always report the number of people polled, and usually that number is a couple thousand. The key is that those couple thousand people are carefully selected so that they represent the broader population that the poll is trying to examine. Now imagine that the poll only contacted people making over 100k and living in cities. Would that still produce data representative of the population in a country as a whole? Obviously not, no matter how many extra people they polled because their sampling is flawed.

The same issue exists with trying to use gw2e as a source of useful information. We’re unable to control the sample, and this means that we’re also unable to draw useful information from it. What if it turns out that most raiders don’t use gw2e because they check gear in other ways? Now the numbers on gw2e are too low. Or if it turns out that mostly raiders use gw2e because of other reasons. Now the numbers are too high. We don’t have any way of getting that sampling information and that makes gw2e a completely worthless source for the purpose of gauging the gw2 community as a whole.

There’s no evidence to support one type of player is more likely to sign up to that website than another. if you check the various metrics, you’ll see what appears to support this. Until such evidence surfaces, which establishes that there are biases, you can assume that all player segments are equally represented among the websites linked account population.

No, you’ve got that backwards actually. Until evidence surfaces that shows equal player representation, you have to assume that an unspecified bias exists in the sample. You can’t just say that any random sample can be trusted. That’s like saying “we haven’t found any evidence that this unspecified pill won’t fix your illness, so we’ll just assume that it works until proven otherwise.” Do you see the problem with that?

Nope. If there’s no evidence to suggest that one type of player would use that website over another, then there really isn’t any issue to use that source until evidence comes about that proves differently.

It’s funny though as you’re pretty much saying that all sources are biased by default and must be proven otherwise. Kind of similar to saying that all sources are unbiased by default and must be proven otherwise. A little amusing how that works out, huh? You’re innocent until proven guilty. No, you’re guilty until proven innocent.

That site has been available for years and has features that can appeal to all players. There’s no evidence to suggest that there is a biased towards one segment of the population to use the website over others. Does that mean that it’s not possible for there to be? No. But then we’re not going on 100% certainties.

Holy mother, its like talking to a wall. You dont have the tool to prove your assertion only A-Net does, because at minimum you would need the total player of GW2, second you would need only the ones that have a level 80 char, and so on.
I will do what you doing now:
I say that 70% of the players of gw2 that dont raid is registred to gw2efficiency, and only 1% of players that raid are registered, so looking at gw2efficency most of the gw2 players raids. Now prove me wrong, if you cant than this is true.
See how its the burden of the one who brings the data to prove if its true or not. If you cant see it now, than i have nothing more to say to you.

OnizukaBR.8537:

Ayrilana.1396:

Fay.2357:

Ayrilana.1396:

Fay.2357:

So I feel the need to correct you here because what you’ve stated is dangerously incorrect. The simple fact that gw2e has a large number of accounts on it does not make that information useful in a statistical sense as you claim. What matters far more than raw size of your sample is whether that sample is representative of your population. Let me illustrate this with an example.

Polling is a big part of politics today. Polls always report the number of people polled, and usually that number is a couple thousand. The key is that those couple thousand people are carefully selected so that they represent the broader population that the poll is trying to examine. Now imagine that the poll only contacted people making over 100k and living in cities. Would that still produce data representative of the population in a country as a whole? Obviously not, no matter how many extra people they polled because their sampling is flawed.

The same issue exists with trying to use gw2e as a source of useful information. We’re unable to control the sample, and this means that we’re also unable to draw useful information from it. What if it turns out that most raiders don’t use gw2e because they check gear in other ways? Now the numbers on gw2e are too low. Or if it turns out that mostly raiders use gw2e because of other reasons. Now the numbers are too high. We don’t have any way of getting that sampling information and that makes gw2e a completely worthless source for the purpose of gauging the gw2 community as a whole.

There’s no evidence to support one type of player is more likely to sign up to that website than another. if you check the various metrics, you’ll see what appears to support this. Until such evidence surfaces, which establishes that there are biases, you can assume that all player segments are equally represented among the websites linked account population.

No, you’ve got that backwards actually. Until evidence surfaces that shows equal player representation, you have to assume that an unspecified bias exists in the sample. You can’t just say that any random sample can be trusted. That’s like saying “we haven’t found any evidence that this unspecified pill won’t fix your illness, so we’ll just assume that it works until proven otherwise.” Do you see the problem with that?

Nope. If there’s no evidence to suggest that one type of player would use that website over another, then there really isn’t any issue to use that source until evidence comes about that proves differently.

It’s funny though as you’re pretty much saying that all sources are biased by default and must be proven otherwise. Kind of similar to saying that all sources are unbiased by default and must be proven otherwise. A little amusing how that works out, huh? You’re innocent until proven guilty. No, you’re guilty until proven innocent.

That site has been available for years and has features that can appeal to all players. There’s no evidence to suggest that there is a biased towards one segment of the population to use the website over others. Does that mean that it’s not possible for there to be? No. But then we’re not going on 100% certainties.

Holy mother, its like talking to a wall. You dont have the tool to prove your assertion only A-Net does, because at minimum you would need the total player of GW2, second you would need only the ones that have a level 80 char, and so on.
I will do what you doing now:
I say that 70% of the players of gw2 that dont raid is registred to gw2efficiency, and only 1% of players that raid are registered, so looking at gw2efficency most of the gw2 players raids. Now prove me wrong, if you cant than this is true.
See how its the burden of the one who brings the data to prove if its true or not. If you cant see it now, than i have nothing more to say to you.

You do not need to know the exact total number of players to form an adequate sample size. That website contains a much large size than you would ever need for GW2.

It’s impractical to test each and every account to determine if they fall in one segment or another. In fact, you’d have to establish exactly what the segments are. There’s nothing to suggest that website would appeal to raiders more than TP flippers which would appeal more to hoarders. It’s simply a website with various features that can appeal to many players depending on their needs.

Your example is easy to dispute by looking at LI metrics.

gw2efficiency suffers from pulling from a sample of the playerbase who knows about the site. I’m fairly certain that despite being out for years, the general population doesn’t know about it because those looking for a site such as GW2Efficiency are those who might want to work with the Trading Post, and people trying to pull statistics.

It’s not something like the GW2 Wiki which is Anet sponsored more or less. It’s a third party site developed to use GW2 API keys for those curious about the numbers. Casual and I would dare say veteran groups in this game have no use to inquire about such a site.

The only reasonable way to make GW2Efficiency more relevant would be to bring attention to it in game, I would say consistent LA spam of ‘Go to GW2Efficiency to look up how cool your account is!’ might start helping.

Ayrilana.1396:

Exactly what subgroup is being specifically picked? You only have access to the data in aggregate form so it’s not like you can pick and choose.

The subgroup dedicated enough to the game they would want to use such a site. Seriously, that alone is going to impact player behaviour.

The fact that this site has something to offer for all kinds of players does not mean that all kinds of players will be interested. Or, to be more precise, that all kinds of players will be interested equally.

Generally, there are two relevant survey biases, that often mess up survey results if they aren’t accounted for:
1. you make a survey, and analyze the results, but only from those that answered, without considering if the fact that someone didn’t answer might not be relevant to the survey
or (very similar one)
you make a survey at a specific location, not considering whether someone being/not being present in this location could be relevant to the survey in question.

You seem to simply assume that the fact someone registered on and is using gw2efficiency is completely not related to their playstyle. That assumption is completely arbitrary – you don’t have any proof that it is true, beyond your belief, and there are some strong indications that it may not be true.
That alone completely invalidates any statistical results about the whole gw2 community you might get out of that site.

Ayrilana.1396:

IThere’s nothing to suggest that website would appeal to raiders more than TP flippers which would appeal more to hoarders.

True. There’s a lot to suggest however that it appeals less to people that are so casual they don’t really care about their account stats. Do you really think this won’t impact the results?

zealex.9410:

Ayrilana.1396:

Whether it’s large, small, growing, or not growing; really has nothing to do with its accessibility. If players want to get into raids then they can do what a lot of others did and join up with others like them with the goal to succeed. quite. Number of raiders are already making it easier by offering training runs.

consider the easy mode idea.

Easy modes are already here… At least for Wing 4. Try normal mode then switch to challenge and see the difference.

Anet don’t want to clearly say there is an easy & hard mode but in fact it’s already here. Maybe not “easy” more “normal” for, at least, the last boss, but it’s really casual friendly. challenge mode need a bit more focus and reflexion.

Paradoxaly, you want an easy mode, but in fact people will be interested by the easy version of the content and not the normal. So it doesn’t change your issue, that you want more people.

Has Anet ever released actual numbers of their game? Even during GW1? I don’t think they care about releasing data, they either do not want any press from it (good or bad) or believe the data somehow compromises their evil plans of making we stick to the game FOREVEEEEER.

Ayrilana.1396:

So you’re telling me that the player population on that website isn’t a large enough sample size?

You’re telling me that only a very specific segment of the population is attracted to that website? Any evidence to support that?

Sample size is fine. GW2E population is biased towards players who value the services the site offers.

OriOri.8724:

You bring up a good point, but I feel its mostly irrelevant. What you describe is intentionally going out and finding your sample, and cherrypicking people to get a certain narrative out of your findings. That is definitely a problem, but GW2Efficiency does not follow that pattern. That site never went out and decided who would make an account. It never cherrypicked people to include in its statistics, so I don’t feel as if your analogy is the least bit relevant to GW2Efficiency.

Incorrect. You’re arguing that the only type of cherrypicking that matters is the type that’s done intentionally, and that’s not the case. The fact that GW2E didn’t intentionally bias their sample is not proof that the sample isn’t biased. There’s any number of reasons that the sample could be biased, and you could make a pretty good argument for it being biased in essentially any direction.

The primary point is that we have no information about the sample that GW2E shows. When sampling in statistics, you must assume that the sample has an unspecified bias until it’s explicitly proven to be a properly representative sample. This is to prevent your analysis from drawing false conclusions based on bad data. Unless the sample can be proven to be properly representative, it’s useless.

Ayrilana.1396:

It’s funny though as you’re pretty much saying that all sources are biased by default and must be proven otherwise. Kind of similar to saying that all sources are unbiased by default and must be proven otherwise. A little amusing how that works out, huh? You’re innocent until proven guilty. No, you’re guilty until proven innocent.

Sorry, but this isn’t a court of law in the USA. This is statistics, and I’m telling you how it works, not asking for your opinion.

All samples must be assumed to have an unspecified bias unless you explicitly show that the sample is representative of your population. This is how you avoid getting drawing false conclusions from your data. In the realm of science and statistics, you can only report what you know to be true, not what you suspect might be the case. In this situation you have no evidence showing that GW2E is a representative sample, and absence of evidence of the opposite is not evidence of absence. Therefor you must assume that unspecified bias exists within the sample.

Khyan.7039:

zealex.9410:

Ayrilana.1396:

Whether it’s large, small, growing, or not growing; really has nothing to do with its accessibility. If players want to get into raids then they can do what a lot of others did and join up with others like them with the goal to succeed. quite. Number of raiders are already making it easier by offering training runs.

consider the easy mode idea.

Easy modes are already here… At least for Wing 4. Try normal mode then switch to challenge and see the difference.

Anet don’t want to clearly say there is an easy & hard mode but in fact it’s already here. Maybe not “easy” more “normal” for, at least, the last boss, but it’s really casual friendly. challenge mode need a bit more focus and reflexion.

Paradoxaly, you want an easy mode, but in fact people will be interested by the easy version of the content and not the normal. So it doesn’t change your issue, that you want more people.

you make a mistake here
the difficulty of the raid is the same and one the normal mode cm is something you can go if you really want to experience it but sadly its just a one time experience.

The “i want an easy mode arguement” at some points make no sense because ppl say they dont have time to bash their heads against a boss and others say i can find 5-9 ppl which making the encounter easier doesnt nececerally fix the player number issue, if a boss isnt dying in under an hour of practice at raid release then it takes too long and ppl cant give that much time….and because of that ppl supposedly cant find more ppl to raid with.

IndigoSundown.5419:

Ayrilana.1396:

So you’re telling me that the player population on that website isn’t a large enough sample size?

You’re telling me that only a very specific segment of the population is attracted to that website? Any evidence to support that?

Sample size is fine. GW2E population is biased towards players who value the services the site offers.

that site does not offer any service for raiding so its completely irrelevant to the arguements that ppl that use it are most likely raiders or not.

We’ll just have to agree to disagree then. 110 thousand people in the sample size is absolutely huge, especially when you consider the number of players that play GW2. According to ArenaNet themselves, 7 million accounts had been created when HoT launched (link below), even if we assume that number has risen to 8 million, we still have a huge sample size relative to the total population. Absolutely huge. Considering the fact that GW2Efficiency statistics can be deemed a convenience sample since they only count those who came out to GW2Efficiency (I think that’s perfectly fair), the size is huge and is therefore statistically significant.

http://www.ign.com/articles/2015/10/23/guild-wars-2-hits-7-million-players-as-expansion-launches