Forum adverts like this one are shown to any user who is not logged in. Join us by filling out a tiny 3 field form and you will get your own, free, dakka user account which gives a good range of benefits to you: No adverts like this in the forums anymore. Times and dates in your local timezone. Full tracking of what you have read so you can skip to your first unread post, easily see what has changed since you last logged in, and easily see what is new at a glance. Email notifications for threads you want to watch closely. Being a part of the oldest wargaming community on the net. If you are already a member then feel free to login now.

Updated.

 mrhappyface wrote:
Updated.

Well. The Tau are certainly starting to make a comeback.



Last three sets, in reverse chronological order:

Tournament over the weekend:

Chaos Space Marines VS Orks

Chaos Space Marines VS Tau

Chaos Space Marines VS Blood Angels

Good to see the fishes in blue are climbing up.

What is up with the Emprah's witches?
Sisters are fabulous.
Twice the ratio of the nearest competitor.

I figure it must be experience,
as any Sisters player must have been using the same army for what, ten years? 15? Has it been 20 years already?
Holy geez I feel old.

It may also be that only a small number of players are in on it in this dataset. Meaning we cannot really say it accounts for variance in skill or opponent type. The skew is real.

The average number of games per army is 78.6. I'd say any army with less than 1/2 of that (39.3) had too few games to be considered for now

The most obvious thing you'd want to check before trying to actually use this data for anything is whether there's a bias in favor of the reporting player. Maybe players who are likely to report games -- and in general probably just people active on this forum -- are significantly better players than average or tend to bring more competitive lists, etc. In this case most of what you'd actually see in the data is just which armies reporting players want to use.

For example, I've played maybe 15 games of 8th at this point. I've won them all. I could report these matches, but there's probably not much information here about the strength of Eldar or Blood Angels or Imperial Guard or Sisters or Necrons, or the weakness of the armies I played against. In several cases I won because I clearly outplayed the other person or because they made huge and obvious mistakes. In other cases there were army mismatches where my lists were simply much better, but this has largely not been because I have better options available to me but because my opponents have not made as good a use of their own options. I've played against several Eldar armies using multiple big units of Dire Avengers each. I've played against lists using lots of Conscripts but where the Conscipts weren't actually defending anything, with the rest of the list including Terminators that appeared far away from the Conscripts and could then be ignored.

This suggests the second big issue that's likely to make this data useless. Nobody really cares whether the game is in some on-average sense balanced. We care about whether it's balanced at particular skill levels and in particular contexts. We want new players to be able to pick up the game and have enjoyable, balanced matches. We want expert players to be able to do the same, at least when they're not trying to break the game with incredibly cheesy lists. And maybe we even want balance for tournaments where we expect people to bring the cheesiest lists they can think of. Maybe Space Marines dominate the top end because 5 Stormravens and Guilliman are hard to beat, while losing horribly with every other composition. Maybe they end up with a roughly even win rate because of this. That's still a huge problem. Ultimately, average win rates are not very helpful for understanding this stuff.

 djones520 wrote:

Audustum wrote:
Custodes Vs. Genestealers: Genestealers.

4 first turn charges from 80 Purestrains after getting seized hurts.

Yeah, I got a major beef with an army that can spam first turn assualts, can ignore moral, can look out sir, and the slew of other special rules those guys get. Hell their "Guardsmen" are base LD 9, so good luck causing morale failures for them.

Switching to my Imperial stuff tomorrow, the Eldar have just constantly dissapointed me so far.

Yeah, the pain is very real. That said, GSC has like nothing else going for it so I'm not sure they'd even be usable without the gimmick.

 jeff white wrote:
What is up with the Emprah's witches?
Sisters are fabulous.
Twice the ratio of the nearest competitor.

I figure it must be experience,
as any Sisters player must have been using the same army for what, ten years? 15? Has it been 20 years already?
Holy geez I feel old.

There's also, like, 4 of us.

While it's less pronounced now than it was on the first day, I'd hazard a fair number of the reported Sisters games are from the same small group of people reporting our own games, because there's not that many of us out there.


I've played, and won, 14 or 15 matches so far, half of them as Sisters of Battle. I can only assume my sisters-in-arms have contributed similarly.


So, I think we're strong, but I think results are also heavily skewed by the players [and lack thereof] playing Sisters.

 jeff white wrote:
What is up with the Emprah's witches?
Sisters are fabulous.
Twice the ratio of the nearest competitor.
I figure it must be experience,

That and the fact that they're 2 times better than marines for points. It doesn't take that much experience to make a comparison and come to a conclusion that they're somewhat overbuffed atm. Not broken but definitely competitive.

Dionysodorus wrote:
The most obvious thing you'd want to check before trying to actually use this data for anything is whether there's a bias in favor of the reporting player. Maybe players who are likely to report games -- and in general probably just people active on this forum -- are significantly better players than average or tend to bring more competitive lists, etc. In this case most of what you'd actually see in the data is just which armies reporting players want to use.

This would be a useful consideration, if we were tracking actual W/L rates for a campaign or something. But we are trying to compare inter-army competitiveness here, meaning that skillful players and effective lists are exactly what we want to compare here.
I would kinda expect regular dakkanoughts to be above average,, as we are essentially a giant bloc of mathammer nerds.


Automatically Appended Next Post:

Dionysodorus wrote:
Nobody really cares whether the game is in some on-average sense balanced. We care about whether it's balanced at particular skill levels and in particular contexts. We want new players to be able to pick up the game and have enjoyable, balanced matches. We want expert players to be able to do the same, at least when they're not trying to break the game with incredibly cheesy lists. And maybe we even want balance for tournaments where we expect people to bring the cheesiest lists they can think of. Maybe Space Marines dominate the top end because 5 Stormravens and Guilliman are hard to beat, while losing horribly with every other composition. Maybe they end up with a roughly even win rate because of this. That's still a huge problem. Ultimately, average win rates are not very helpful for understanding this stuff.

Knowing these numbers is exactly how you would go about achieving this ideal. We are not looking to become the GRAYTEST PLAYURZ EVAR, we're trying to identify which armies are on the upper and lower bounds of power, while tracking how well everyhting congregates towards a central power level. You need numbers for that. The data isn't exactly descriptive or 100% reliable, but it is better than nothing, and fulfills our objective here.

 koooaei wrote:
That and the fact that they're 2 times better than marines for points.

Err, no, Battle Sisters aren't "two times better".

Space Marines have huge advantages over Sisters when it comes to their statline. Those four points difference pay for +1s, +1t, ATSKNF, and +1WS. And, soon, it'll also pay for chapter tactics, as well. If a Tactical Marine squad charges a Battle Sister squad, it's all over for the BSS-- that's how huge the difference is in their statlines. Even a Tactical squad can outfight us in melee. A scout squad with combat knives could devastate almost any of our squads in melee.

Actually I'd argue Sisters were probably overcosted until this edition. Right now they're able to play their to their favored (short-ranged shooting) playstyle efficiently. And that's literally all they're able to do. This is the same playstyle Sisters players have been using for 20 years; in fact, barring having more units than previous editions due to points costs, we're literally playing the exact same lists we've been playing since 3e Codex: Witch Hunters and to a lesser extent 2e Codex: Sisters of Battle.

A lot of those wins are commander spam keep that in mind. Just like Flyrant lists were good for a time, but soon power creep put the nail in that coffin. Once commander spam is taken care of it's probably over for the Tau.

 Gamgee wrote:
A lot of those wins are commander spam keep that in mind. Just like Flyrant lists were good for a time, but soon power creep put the nail in that coffin. Once commander spam is taken care of it's probably over for the Tau.

let's hope that the commander spam is taken care of while simultaniously getting buffs to the 2/3 of the stuff in the book (overpriced units and gear). I feel your pain. Ork.

 koooaei wrote:

 Gamgee wrote:
A lot of those wins are commander spam keep that in mind. Just like Flyrant lists were good for a time, but soon power creep put the nail in that coffin. Once commander spam is taken care of it's probably over for the Tau.

let's hope that the commander spam is taken care of while simultaniously getting buffs to the 2/3 of the stuff in the book (overpriced units and gear). I feal your pain. Ork.

Exactly I want the rest of the army reworked a little so it's viable to bring Tau without having to resort to the commander spam list. I always felt your pain. I had a friend who played orks and nids all 7th ed. Poor orks being this bad for this long.

 Selym wrote:

Dionysodorus wrote:
The most obvious thing you'd want to check before trying to actually use this data for anything is whether there's a bias in favor of the reporting player. Maybe players who are likely to report games -- and in general probably just people active on this forum -- are significantly better players than average or tend to bring more competitive lists, etc. In this case most of what you'd actually see in the data is just which armies reporting players want to use.

This would be a useful consideration, if we were tracking actual W/L rates for a campaign or something. But we are trying to compare inter-army competitiveness here, meaning that skillful players and effective lists are exactly what we want to compare here.
I would kinda expect regular dakkanoughts to be above average,, as we are essentially a giant bloc of mathammer nerds.


Automatically Appended Next Post:

Dionysodorus wrote:
Nobody really cares whether the game is in some on-average sense balanced. We care about whether it's balanced at particular skill levels and in particular contexts. We want new players to be able to pick up the game and have enjoyable, balanced matches. We want expert players to be able to do the same, at least when they're not trying to break the game with incredibly cheesy lists. And maybe we even want balance for tournaments where we expect people to bring the cheesiest lists they can think of. Maybe Space Marines dominate the top end because 5 Stormravens and Guilliman are hard to beat, while losing horribly with every other composition. Maybe they end up with a roughly even win rate because of this. That's still a huge problem. Ultimately, average win rates are not very helpful for understanding this stuff.

Knowing these numbers is exactly how you would go about achieving this ideal. We are not looking to become the GRAYTEST PLAYURZ EVAR, we're trying to identify which armies are on the upper and lower bounds of power, while tracking how well everyhting congregates towards a central power level. You need numbers for that. The data isn't exactly descriptive or 100% reliable, but it is better than nothing, and fulfills our objective here.

This is innumerate nonsense. This data is essentially useless. Yes, it is "better than nothing". It is far, far worse than a competent person playing a few games and seeing how units feel to them, supplemented by lots of mathhammer.

All statistics is best interpreted through the lens of expertise and experience, this is no different. And it is being interpreted through that lens.

Dionysodorus wrote:
This is innumerate nonsense. This data is essentially useless. Yes, it is "better than nothing". It is far, far worse than a competent person playing a few games and seeing how units feel to them, supplemented by lots of mathhammer.

Then is it your belief that Imperial Guard aren't top level this edition? Do you think that Sisters are bad too? Do you believe that Tau and Orks are in fact great armies this edition? And if you don't believe any of that then hasn't the data shown quite clearly which armies are doing well and which aren't?

By saying that this data is useless you are denying what is so blatantly true and supported by this data.

 mrhappyface wrote:

Dionysodorus wrote:
This is innumerate nonsense. This data is essentially useless. Yes, it is "better than nothing". It is far, far worse than a competent person playing a few games and seeing how units feel to them, supplemented by lots of mathhammer.

Then is it your belief that Imperial Guard aren't top level this edition? Do you think that Sisters are bad too? Do you believe that Tau and Orks are in fact great armies this edition? And if you don't believe any of that then hasn't the data shown quite clearly which armies are doing well and which aren't?

By saying that this data is useless you are denying what is so blatantly true and supported by this data.

No, this is also nonsense. No one with any expertise in statistics could think this.

I think Guard are top tier. I think Sisters are very strong. I think Tau and Orks are pretty weak. This data does not provide good evidence that any of that is true, though. Just because it lines up with your (and my) prejudices doesn't mean that it's solid -- that's awful reasoning. It is absolutely not true, and frankly is kind of embarrassing, to say that my contention that this data is useless is the same as disagreeing with some of what it seems on its face to show.

I mean, in my earlier post I went over a couple possible issues with the data. You'll recall that one problem I brought up is that maybe dakka-ites tend to win, and so maybe this is just showing which armies dakka-ites want to play, which will probably be biased towards the ones that they think are good. So then it'd be totally unsurprising for this (garbage) data to end up reflecting the forum's consensus about which armies are good. There are of course many other possible sources of bias.

 mrhappyface wrote:

Dionysodorus wrote:
This is innumerate nonsense. This data is essentially useless. Yes, it is "better than nothing". It is far, far worse than a competent person playing a few games and seeing how units feel to them, supplemented by lots of mathhammer.

Then is it your belief that Imperial Guard aren't top level this edition? Do you think that Sisters are bad too? Do you believe that Tau and Orks are in fact great armies this edition? And if you don't believe any of that then hasn't the data shown quite clearly which armies are doing well and which aren't?

By saying that this data is useless you are denying what is so blatantly true and supported by this data.

Welcome to the world of pain I had in 7th trying to show Tau was only mid tier other then riptide wing, stormsurge spam, and some formations. Data and logical analysis means nothing to some people. I will make one point for my opponent since he is doing rather poorly in the debate himself. While this data shows which armies are weak it doesn't really specify how they are weak. That would take more in depth analysis and more time. Thankfully as more and more battle reports come in we will be able to get a better picture.

Some like the Tau are well understood where their problem lies. Others like the Deathwatch is a little harder just due to the small sample size and player base. So far the rough consensus seems to be the army is simply a little too expensive for some things.

Edit
Okay I seen your new post. You do make a point. The data is not professional level of data that will survive scrutiny from a university professor. At least right now. Dakka's sample size is more or less equalling other websites sample size as well and we're seeing a fairly similar distribution in actual tournament winners ect ect. Where your point is most valid is not the obvious extremes but the middle of the ground since it's easier for bias to creep into the data when it's all a lot closer in value.

And yet we have no other kind of metric, outside tournament results, to see if X army is doing better than army Y, overall.

I'd rather take heed of this info than what happens only in my meta (which is much less representative).

The only real problem that can arise from this is if someone is trying to intentionally dilute the data with false reporting or not reporting the losses. The data follows the general trends everyone expected so that isn't likely.

The only thing that is a LARGE outlier is the Sisters result. How many unique sisters players have reported these results? For instance I know I only compromise ~12% of the GK results.


Automatically Appended Next Post:

 Gamgee wrote:

 mrhappyface wrote:

Dionysodorus wrote:
This is innumerate nonsense. This data is essentially useless. Yes, it is "better than nothing". It is far, far worse than a competent person playing a few games and seeing how units feel to them, supplemented by lots of mathhammer.

Then is it your belief that Imperial Guard aren't top level this edition? Do you think that Sisters are bad too? Do you believe that Tau and Orks are in fact great armies this edition? And if you don't believe any of that then hasn't the data shown quite clearly which armies are doing well and which aren't?

By saying that this data is useless you are denying what is so blatantly true and supported by this data.

Welcome to the world of pain I had in 7th trying to show Tau was only mid tier other then riptide wing, stormsurge spam, and some formations. Data and logical analysis means nothing to some people. I will make one point for my opponent since he is doing rather poorly in the debate himself. While this data shows which armies are weak it doesn't really specify how they are weak. That would take more in depth analysis and more time. Thankfully as more and more battle reports come in we will be able to get a better picture.

Some like the Tau are well understood where their problem lies. Others like the Deathwatch is a little harder just due to the small sample size and player base. So far the rough consensus seems to be the army is simply a little too expensive for some things.

Edit
Okay I seen your new post. You do make a point. The data is not professional level of data that will survive scrutiny from a university professor. At least right now. Dakka's sample size is more or less equalling other websites sample size as well and we're seeing a fairly similar distribution in actual tournament winners ect ect. Where your point is most valid is not the obvious extremes but the middle of the ground since it's easier for bias to creep into the data when it's all a lot closer in value.

Gamgee, we both know everybody hated Tau because of the playstyle and didn't care about the W/L data.

Dionysodorus wrote:
No, this is also nonsense. No one with any expertise in statistics could think this.

I think Guard are top tier. I think Sisters are very strong. I think Tau and Orks are pretty weak. This data does not provide good evidence that any of that is true, though. Just because it lines up with your (and my) prejudices doesn't mean that it's solid -- that's awful reasoning. It is absolutely not true, and frankly is kind of embarrassing, to say that my contention that this data is useless is the same as disagreeing with some of what it seems on its face to show.

I mean, in my earlier post I went over a couple possible issues with the data. You'll recall that one problem I brought up is that maybe dakka-ites tend to win, and so maybe this is just showing which armies dakka-ites want to play, which will probably be biased towards the ones that they think are good. So then it'd be totally unsurprising for this (garbage) data to end up reflecting the forum's consensus about which armies are good. There are of course many other possible sources of bias.

Anyone with actual experience in statistics would know you're talking rubbish.

The data set has problems certainly, there are biases as you say, but statistics is never about perfect data sets and if you actually had experience in statistics you'd know that. This data set is comprised from the experiences of probably under 100 Dakkanaughts and so you'd expect that the data from these players and the forum posts to coalesce, that I do not deny. However, the community outside of Dakka also reports that IG and SoB are doing great and Orks and Tau have been screwed again. The only outliers to this data spread come in the form of highly competitive tournaments where the only comparisons to be made there are between the effectiveness of Conscripts VS Stormravens VS Razorwings VS Horrors.

Saying that this data is garbage is just wrong and shows you have a very narrow understanding of how prejudice and biases are dealt with in real world data sets.

 Vector Strike wrote:
And yet we have no other kind of metric, outside tournament results, to see if X army is doing better than army Y, overall.

I'd rather take heed of this info than what happens only in my meta (which is much less representative).

I mean, you probably shouldn't do either. If you don't feel competent to come to reasonably certain conclusions from your experience playing and watching games, reading battle reports, working out the math, etc. -- and really most people who do feel competent won't be -- you should embrace uncertainty. "We don't have anything better to go on so we have to use this" is basically never a good approach, unless for some reason you have to make a decision. You should simply adopt a Bayesian worldview and perhaps slightly update the probability you attach to certain propositions being true as new evidence comes in. The evidence presented in this thread is extraordinarily weak, and so you should not update your beliefs very much because of it.

 Quickjager wrote:
The only real problem that can arise from this is if someone is trying to intentionally dilute the data with false reporting or not reporting the losses. The data follows the general trends everyone expected so that isn't likely.

The only thing that is a LARGE outlier is the Sisters result. How many unique sisters players have reported these results? For instance I know I only compromise ~12% of the GK results.


Automatically Appended Next Post:

 Gamgee wrote:

 mrhappyface wrote:

Dionysodorus wrote:
This is innumerate nonsense. This data is essentially useless. Yes, it is "better than nothing". It is far, far worse than a competent person playing a few games and seeing how units feel to them, supplemented by lots of mathhammer.

Then is it your belief that Imperial Guard aren't top level this edition? Do you think that Sisters are bad too? Do you believe that Tau and Orks are in fact great armies this edition? And if you don't believe any of that then hasn't the data shown quite clearly which armies are doing well and which aren't?

By saying that this data is useless you are denying what is so blatantly true and supported by this data.

Welcome to the world of pain I had in 7th trying to show Tau was only mid tier other then riptide wing, stormsurge spam, and some formations. Data and logical analysis means nothing to some people. I will make one point for my opponent since he is doing rather poorly in the debate himself. While this data shows which armies are weak it doesn't really specify how they are weak. That would take more in depth analysis and more time. Thankfully as more and more battle reports come in we will be able to get a better picture.

Some like the Tau are well understood where their problem lies. Others like the Deathwatch is a little harder just due to the small sample size and player base. So far the rough consensus seems to be the army is simply a little too expensive for some things.

Edit
Okay I seen your new post. You do make a point. The data is not professional level of data that will survive scrutiny from a university professor. At least right now. Dakka's sample size is more or less equalling other websites sample size as well and we're seeing a fairly similar distribution in actual tournament winners ect ect. Where your point is most valid is not the obvious extremes but the middle of the ground since it's easier for bias to creep into the data when it's all a lot closer in value.

Gamgee, we both know everybody hated Tau because of the playstyle and didn't care about the W/L data.

Yet now the Tau are even more gunline than ever. Oh the irony. You are right Tau hare was not rational and I've seen many dumb reasons proposed to nerf them. The loud minority made it sure seem like the Tau were more broken than the old fish of fury days. Ah well time to rectify the situation then and I have hope GW will do this.

 mrhappyface wrote:

Dionysodorus wrote:
No, this is also nonsense. No one with any expertise in statistics could think this.

I think Guard are top tier. I think Sisters are very strong. I think Tau and Orks are pretty weak. This data does not provide good evidence that any of that is true, though. Just because it lines up with your (and my) prejudices doesn't mean that it's solid -- that's awful reasoning. It is absolutely not true, and frankly is kind of embarrassing, to say that my contention that this data is useless is the same as disagreeing with some of what it seems on its face to show.

I mean, in my earlier post I went over a couple possible issues with the data. You'll recall that one problem I brought up is that maybe dakka-ites tend to win, and so maybe this is just showing which armies dakka-ites want to play, which will probably be biased towards the ones that they think are good. So then it'd be totally unsurprising for this (garbage) data to end up reflecting the forum's consensus about which armies are good. There are of course many other possible sources of bias.

Anyone with actual experience in statistics would know you're talking rubbish.

The data set has problems certainly, there are biases as you say, but statistics is never about perfect data sets and if you actually had experience in statistics you'd know that. This data set is comprised from the experiences of probably under 100 Dakkanaughts and so you'd expect that the data from these players and the forum posts to coalesce, that I do not deny. However, the community outside of Dakka also reports that IG and SoB are doing great and Orks and Tau have been screwed again. The only outliers to this data spread come in the form of highly competitive tournaments where the only comparisons to be made there are between the effectiveness of Conscripts VS Stormravens VS Razorwings VS Horrors.

Saying that this data is garbage is just wrong and shows you have a very narrow understanding of how prejudice and biases are dealt with in real world data sets.

I can only assume that you're just lying about your familiarity with how statistics is used in the real world. Your approach is far, far below acceptable real-world standards in political polling, product quality control, determining the effectiveness of new medical procedures, and really any investigation into any interesting question.

Dionysodorus wrote:

 mrhappyface wrote:

Dionysodorus wrote:
No, this is also nonsense. No one with any expertise in statistics could think this.

I think Guard are top tier. I think Sisters are very strong. I think Tau and Orks are pretty weak. This data does not provide good evidence that any of that is true, though. Just because it lines up with your (and my) prejudices doesn't mean that it's solid -- that's awful reasoning. It is absolutely not true, and frankly is kind of embarrassing, to say that my contention that this data is useless is the same as disagreeing with some of what it seems on its face to show.

I mean, in my earlier post I went over a couple possible issues with the data. You'll recall that one problem I brought up is that maybe dakka-ites tend to win, and so maybe this is just showing which armies dakka-ites want to play, which will probably be biased towards the ones that they think are good. So then it'd be totally unsurprising for this (garbage) data to end up reflecting the forum's consensus about which armies are good. There are of course many other possible sources of bias.

Anyone with actual experience in statistics would know you're talking rubbish.

The data set has problems certainly, there are biases as you say, but statistics is never about perfect data sets and if you actually had experience in statistics you'd know that. This data set is comprised from the experiences of probably under 100 Dakkanaughts and so you'd expect that the data from these players and the forum posts to coalesce, that I do not deny. However, the community outside of Dakka also reports that IG and SoB are doing great and Orks and Tau have been screwed again. The only outliers to this data spread come in the form of highly competitive tournaments where the only comparisons to be made there are between the effectiveness of Conscripts VS Stormravens VS Razorwings VS Horrors.

Saying that this data is garbage is just wrong and shows you have a very narrow understanding of how prejudice and biases are dealt with in real world data sets.

I can only assume that you're just lying about your familiarity with how statistics is used in the real world. Your approach is far, far below acceptable real-world standards in political polling, product quality control, determining the effectiveness of new medical procedures, and really any investigation into any interesting question.

Until you can prove the proper usage of statistics and why you are superior at them this point is merely an attack on the opponent and getting very close to getting personal.

This is getting way too heated. I'm gonna go over there.

See ya in the conscripts thread!