Space Marines at the Las Vegas Nopen 2021

[Alcyon]

My concern with 9th, comes down to the current meta leaders. They're all armies with great assault options, which makes sense with the emphasis on taking midfield objectives. So how does GW make the shooting based armies like IG & tau able to compete? I'm less worried about power creep than I am about some fractions just not having the tools to succeed with these missions, and its frustrating because armies like Tau and Guard should be able to go in the midfield and be effective.

 

It's OT but I think this is the main issue, yeah. If Tau get some kind of jump-shoot-jump again that'll be a big help, allowing them to shoot something off an objective and then "charge" onto it without having to risk actual close combat and thus being unable to shoot next turn or Overwatch. Some discounting of their overcosted units and buffing of their weapons would help them as well.

 

Guard, I think their best bet is to play hordes with gun backup. They do have the now-cheaper Ogyrns as a pretty solid combat unit especially with their cheap character support (Missionaries, Astropaths etc.) They're not in a great spot but I don't think there are many competitive players championing them right now. I'd like to see a list with lots of Ogyrns, Manticores and Infantry Squads or at least an explanation why those units don't work in IG the way their counterparts do in other lists.

[BLACK BLΠFLY]

I’m sure GeeDub will give Tau a big helping hand when they get new rules but hopefully they won’t be quite as unfun to play against again .

[Karhedron]

I think all the bigger heavier SM vehicles are in a bad spot points wise. None get core without special circumstances which hurts. And most are way over priced for what they bring. ATVs, Storm Speeders, and Dreds, are were its at now. Massed infantry might be the way to go too.

Agreed. Razorbacks are also quite good. They bring a balance of transport and firepower for a pretty modest cost.

[Mandragola]

I don't think the vehicle issue is unique to marines, but it's probably worse for them. The Primaris vehicles tend to have enormous firepower relative to their durability, meaning that all of them are glass cannons. That's not what you want at all in this edition of the game. 

 

I don't think that makes the faction particularly weak though. There are so many marine units that there are still plenty of good ones, regardless. Most codexes have weak units in. What defines an army's strength competitively is how good its best units are, not its worst ones. So Harlequins have those shooty skimmer things whose name I can't remember and they've always been trash, but that's fine because they can spam bikes and troupers.

[Jorin Helm-splitter]

@jorin a few metawatch articles ago the 40kstats guys claimed marines are actually under represented in tournaments compared to the faction amounts. 33% of factions are marines, with 31% of. Still a lot of marine armies, but the majority of opponents weren't marines.

 

I meant they were over represented by tier not the number of players vs. the proportion of fractions that they make up in 40k. For example if you look at their meta watch article from December they have a post codex table comparing armies performance before & after the marine/necron codex release. Tau and Harlequin have a similar number of games played even though Tau are definitely a bottom three army.  There are a lot of reasons for this cost of the army, the painting aspects, and the complexity of the rules.

 

So if you look at Prot's example of the top 500 players basically setting the tier list. I don't think it would reflect it nearly as well as if you put the top 500 magic players or top 500 from a video game in an event. You'd see some lower tier decks, and fighters/heroes but the top tiers would be extremely well represented. For example in early 8th edition Ynnari were really strong but I can't remember them having more players at major event then marines once.   

[techsoldaten]

 

 

@prot yea I totally agree.

 

@techsoldaten you're kind of making a lot of assertions and moving the goal posts. You claim there's not enough tournament data, when theres tons of tournaments. You say a lot of strong players aren't playing, but the Nopen just took place in person and TTS exists; strong players are playing. What amount of data is enough? How many "strong players" need to be playing to say that marine pilots are good enough to count? Why does TTS not count? Why do we need to discard any results from places that have covid under control and life is pretty normal?

 

goonhammer just posted an article today about the current meta, and has white scars as 7th with their new algorithm. Idk if you feel like their sample size of 4000 games in January is representational or not, but it's the most comprehensive data parsing we have.

I'm not moving any goalposts. There's a small subset of competitive players facing each other in something called tournaments.

 

These events are not the same as open events where everyone can face each other.

 

The results mean very little. The differentiation point is not quantitative, it's characteristic. Either you are playing against the best other players have to throw against you or you are not.

 

If you are determined to draw conclusions from the results of a small subset - which I estimate to be lower than 10% of the regular tournament crowd - go ahead. Means nothing, you're just making noise.

Ignores that tts is a thing for the millionith time and can't state what metrics he would deem acceptable for data, also for the millionith time.

 

 

Is the issue quantity or quality? You're asking me to provide some criteria that's not representative of reality.

 

I'm local to the NOVA Open, about half the registrants come from here. Every FLGS is locked down, people are not getting together for games in stores or in private. I'm in a Discord with about 200 other players, this is where people used to arrange games. Nobody is talking about TTS. We are trading advice on travelling with your army to Florida or Texas and (in some cases) relocating. Neither is simple to do.

 

That's one major tournament and I'm assuming it's representative of most places. Are you trying to say that's a bad assumption and most competitive players are not affected by lockdowns? 

 

AFAICT, the only people playing right now are the ones who can afford the costs of travel or live somewhere that's not locked down. That excludes a lot of people.

[SkimaskMohawk]

@jorin fair points for sure

 

@techsoldaten idk what the issue is; you're the one whos saying the data doesn't matter due to lockdowns. I'm personally unsure how you would even track a qualitative measurement of tournament results. Who are these paragons of good 40k? How many of them do you need of them at a tournament for it to be counted? How many quality tournaments need to occur for a satisfactory sample size? You seem to have discarded sheer quantity, but at a certain point you still need to answer the question of "what's the sample size youd accept" .

 

As for tts, I can tell you that the 40k TTS discord has over 40000 people in it. So there's definitely representative action coming from it. It all circles back to what criteria is acceptable for analysis; how do tell who's "the best" and qualified to make games count for you, how many games need to be counted to make the sample size large enough, etc...You can't just sit back and argue against the provided metrics and data without ever giving details of your alternative.

[techsoldaten]

@techsoldaten idk what the issue is; you're the one whos saying the data doesn't matter due to lockdowns. I'm personally unsure how you would even track a qualitative measurement of tournament results. Who are these paragons of good 40k? How many of them do you need of them at a tournament for it to be counted? How many quality tournaments need to occur for a satisfactory sample size? You seem to have discarded sheer quantity, but at a certain point you still need to answer the question of "what's the sample size youd accept" .

 

As for tts, I can tell you that the 40k TTS discord has over 40000 people in it. So there's definitely representative action coming from it. It all circles back to what criteria is acceptable for analysis; how do tell who's "the best" and qualified to make games count for you, how many games need to be counted to make the sample size large enough, etc...You can't just sit back and argue against the provided metrics and data without ever giving details of your alternative.

 

My point is current tournament results are not representative of Codex strength, because large numbers of people are not able to play.

 

I don't need to provide a quantitative measure, it's self-evident. Either you're engaged in general competition or you're not. If half the country can not show up for tournaments, it's not safe to assume the results would be the same for the half that cannot participate. That shouldn't be a controversial idea.

 

A few assumptions are baked into this assertion. I'm assuming the number of people who are participating is much less than half of the people who would otherwise play. I'm assuming Codex strength is a relative measure derived from large numbers of games, it represents collective wisdom more than a quantitative measure. And I'm assuming there's nothing special about the people who are participating, they don't possess insights that could replace the collective wisdom mentioned in point 2.

 

None of this is to say tournament data doesn't matter. I'm saying the datum is less significant because it doesn't measure the same thing it once did. At best, it means put an asterisk by those results, treat them as an indicator, but don't be in a rush to draw conclusions.

 

You keep making this about me and what I would be prepared to accept. I'm pointing out the obvious that these results are apples to oranges compared to what they have historically represented. If you want to make a case that tournament results have not been affected, or that more people are able to participate than I am accounting for, or that Codex strength really is a measure of something else, or that the people playing in tournaments possess magical insights that accurately predict what the outcomes would have been if large numbers of people were playing, go ahead. Happy to take up those points.

 

But stop asking me to tell you what number would make the data the same. Full participation would make the data the same, full stop. There is no arbitrary marker that would reveal the outcomes of an otherwise unrestricted environment. 

 

Part of why I'm strong on this point: I'd intuit, as a measure of Codex strength, Space Marines are a fair bit stronger than current tournament results would suggest. 9th edition has seen a major change in mechanics and identifying the best tactics to adapt will take a lot of trial and error. As more games are played, people will overcome their biases toward certain units and figure out how to adapt.

 

What is it about the 16 / 32 player tournament formats we've been seeing that convinces anyone this process of adaptation is happening? I'd really like to know because I don't see it at all. Some of the lists have surprised me and I do see people trying out ideas.

 

But there's nothing I would call a fully-optimized TAAC power list going on with any faction. Repentia spam included.

[Dudley Nightshade]

IANAS (I am not a statistician), but this disagreement about whether an analysis based on the current rate of participation due to pandemic appears to boil down to whether we have a large enough sample for statistical validity. I don’t know the answer, but maybe someone who does this for a living or understands how Goonhammer or 40kstats generates their results might be able to answer that question. In a simple analysis, a sample size of 4000 would be plenty, but that may not be the case here. More complex analyses require more data. It would only be in the case of tiny populations, though, where 100% participation were required to be considered a valid result.

 

Statistics aside, I have a question to add: if some armies don’t work well for the average player, but become unstoppable for the most skilled players, then what do tiers mean for the average player? Is it even fair to say that an army is top-tier if I probably can't use it well enough to win with it? Or is this actually a myth? Maybe that's more than one question. I guess to ask it differently, who are these rankings for, and what can an average player do with this information? 

[SkimaskMohawk]

IANAS (I am not a statistician), but this disagreement about whether an analysis based on the current rate of participation due to pandemic appears to boil down to whether we have a large enough sample for statistical validity. I don’t know the answer, but maybe someone who does this for a living or understands how Goonhammer or 40kstats generates their results might be able to answer that question. In a simple analysis, a sample size of 4000 would be plenty, but that may not be the case here. More complex analyses require more data. It would only be in the case of tiny populations, though, where 100% participation were required to be considered a valid result.

This is kind of the thing. Sample sizes by definition don't fully cover a population, otherwise it'd be a census. There's always a minimum required to hit the proper confidence level. I find it hard to believe that the exact sample size determination for 40k army performance was the pre-covid player base; that any lower numbers than what was being calculated made the confidence level too wide.

 

I also don't really buy that the "elite" gamers need to be omni-present to determine a data set. Pre covid there were so many tournaments going on at the same time that you couldn't have Navanti or Seigler at every GT; theres only so many consistent top winners, so there must have been a broad enough pool of good enough players to make that data valid as well.

 

In both cases you just need a certain amount of data for the math to be valid and I'd trust the goonhammer guy who does data-science for a living to say "hey our sample size isn't the best for this upcoming analysis" like they did with the most recent January numbers on first turn advantage.

[Jorin Helm-splitter]

@Brother Yroc

 

Tier lists at least for fighting games and card games are designed to show which Character/deck would be favored in a matchup of players with equal skill and typically are presented more for the average player. For esport games they may focus more on the tiers for the average pro player instead of the average player.

Edited February 8, 2021 by Jorin Helm-splitter

[techsoldaten]

IANAS (I am not a statistician), but this disagreement about whether an analysis based on the current rate of participation due to pandemic appears to boil down to whether we have a large enough sample for statistical validity. I don’t know the answer, but maybe someone who does this for a living or understands how Goonhammer or 40kstats generates their results might be able to answer that question. In a simple analysis, a sample size of 4000 would be plenty, but that may not be the case here. More complex analyses require more data. It would only be in the case of tiny populations, though, where 100% participation were required to be considered a valid result.

 

There is no statistical validity argument. Either people can freely participate or not.

 

That changes outcomes of matches through selection bias even for people who ARE playing.

 

Again, I'm not saying tournament results are meaningless, I am saying it's not wise to draw conclusions from them. There's no threshold at which they suddenly become 'representative,' you really don't know what would be happening but for this pandemic.

 

The only conclusions you could reach are that they measure how people with wealth / favorable local health policies do with current Codexes.

 

And that's just not interesting.

 

 

Statistics aside, I have a question to add: if some armies don’t work well for the average player, but become unstoppable for the most skilled players, then what do tiers mean for the average player? Is it even fair to say that an army is top-tier if I probably can't use it well enough to win with it? Or is this actually a myth? Maybe that's more than one question. I guess to ask it differently, who are these rankings for, and what can an average player do with this information? 

 

Tiers are a relative measure of strength of Codex options based on collective wisdom. The worst Codex can beat the best Codex on any given day. The likelihood of that happening may be small, but it's always a possibility.

 

Talent is a measure of how well someone can use their army.

 

You could have a top-tier list, be a talentless player, and lose. You could have a bottom tier list, be a highly talented player, and win. Both things happen.

[Alcyon]

I don't think it makes sense to completely discount the current dataset as useless because it doesn't match the full pre-COVID playerbase. I think Wings @ 40kStats would be far more cagey about making predictions or insights based on this dataset if he shared your view, and he is a statistician/data scientist. 

 

Your caveats are important to keep in mind, and I don't think we should ever take meta analysis as the one true gospel, but to completely reject it is another whole kettle of fish. Sometimes I do find myself staying off the B&C or 40k forums generally when the hype or melodrama about balance goes overboard, but there's certainly some value in the dataset we do have and I think broadly it matches people's intuitions on codex strength right now.

[Karhedron]

And then you get counter-meta lists which are not inherently strong but are perfectly tuned to be effective against whatever is currently ruling the meta.

 

The Space Wolves "Circle the Wagons" list with 5 Impulsors from the end of 8th was a good example of this. It did not look particularly strong on paper but used a combination of rules very effectively to create a novel sort of Deathstar.

[techsoldaten]

I don't think it makes sense to completely discount the current dataset as useless because it doesn't match the full pre-COVID playerbase. I think Wings @ 40kStats would be far more cagey about making predictions or insights based on this dataset if he shared your view, and he is a statistician/data scientist. 

 

Your caveats are important to keep in mind, and I don't think we should ever take meta analysis as the one true gospel, but to completely reject it is another whole kettle of fish. Sometimes I do find myself staying off the B&C or 40k forums generally when the hype or melodrama about balance goes overboard, but there's certainly some value in the dataset we do have and I think broadly it matches people's intuitions on codex strength right now.

 

I've never said current tournament results are useless. I have said drawing conclusions from them about Codex strength is unwise.

 

The simplest way to understand my concerns with selection bias is to consider the results from the Las Vegas Nopen. It's a 16 player tournament, most of the ones I'm aware of right now are 8 / 16 players.

 

The SoB list that won - where was the hard counter to Repentia spam? 

 

Of all the lists I saw for that tournament, only the BA and one of the Ork lists looked like it could have handled mid-range shooting and melee. They were both eliminated without facing Sisters. This was an elimination tournament, no one was accumulating points that determined the winner at the end.

 

The question I'm asking myself: if there were double the number of players, or it was a progressive - not an elimination tournament - how would that have affected standings? Would SoB have won if they had to face the Orks one round? What if the lists were more polished, if people had been playing them longer and more frequently?

 

Dunno. But I wouldn't take it for granted the results would be the same with more players, a progressive format, or more polished lists. I can just feel the absence of a meta in the videos.

[Jorin Helm-splitter]

@Techsoldaten

 

If you can't draw broad conclusions about army strength from the data, what can you draw conclusions about from it?

[BLACK BLΠFLY]

It’s hard to draw serious conclusions from Nopen since it wasn’t the same players piloting their own lists each round and some of the entrants aren’t top tourney contenders.

[techsoldaten]

@Techsoldaten

 

If you can't draw broad conclusions about army strength from the data, what can you draw conclusions about from it?

 

Would not say it's safe to draw conclusions from the data.

 

This is not a record of competitive outcomes from the total population of players, or a subset thereof. This is a segregated result set limited to players who can afford to travel or live in places with permissive public health policies. A large number of people who would otherwise be playing are excluded, and that matters.

 

Further, the data does not include results from large tournaments. Even ITC recognizes those outcomes matter more than small regional ones, and I've heard people complaining about tournament formats in other boards. While I don't like criticizing TOs, there's allegations of gamesmanship that should not be ignored if you are considering what relative power levels look like.

 

If you want to use the results as an indicator of relative Codex strength, fine. That's probably safe until vaccinations are more widespread and lockdowns ease up.

 

But I'd expect the relative strength of Codexes to change dramatically once people start getting in more games. If nothing else, the meta will shift towards countering SoB / Slaanesh Daemons / Harlequins because that will be the starting point for 9th edition with a lot of players.

[SkimaskMohawk]

But the both blood angels got eliminated by the sororitas lists lol.

 

I will agree that the nopen isn't a good meta snapshot, but it does show the relative ease of piloting various armies and how unfamiliarity with an opponent can cause big problems.

[Jorin Helm-splitter]

It’s hard to draw serious conclusions from Nopen since it wasn’t the same players piloting their own lists each round and some of the entrants aren’t top tourney contenders.

 

 

I thought we were talking about the data that 40k stats, and other sources have put together. The Nopen results just brought up the "what tier are marines" discussion. 

[BLACK BLΠFLY]

I was going by the title of the OP .

[Dudley Nightshade]

Thanks all. I didn’t quite communicate my question about tiers properly, and I realize that it's actually quite off topic for this discussion. Sorry to derail. If it really still bugs me after further thought I’ll post something in Amicus, but I’m kind of still wrestling with how to communicate it. 
 

For the stats thing, again, I lost track of the specifics of this topic. Based just on the Nopen results, it does seem unlikely that we could draw any meaningful conclusions about the relative strength of the codexes.

[Jorin Helm-splitter]

@Brother Yroc

 

I wouldn't let a question about tiers frustrate you, they're meant to be a quick guide not some absolute.

 

I was going by the title of the OP .

Sorry I thought you were responding to me cause your post was right after mine. I just wanted to make it clear that the Nopen isn't data I would put much value on either (games were entertaining though).

 

The OP really makes this a really broad topic, but it does make it harder to follow.