Jump to content
Stay On The Leader

A picnic in the sunshine with cake for everyone

Recommended Posts

If r3-A2 was 3 points, it wouldn't ruin it.  R3 needs to be 1 point. and all the junk astromechs should be reprogrammed into better ones. I think r3-a2 is pretty balanced in that it causes stress to receive stress, but yeah, maybe it should be a once per round thing.

 

Share this post


Link to post
Share on other sites

If Stressbot was only once a turn, you're going to need to also level the playing field by nerfing all ships and situations where single stressing basically does nothing. Double-stress SHOULD be bad for you, it's meant to stack stress so you have to take stress effects and can't just shed them because you've got a wealth of green moves, or are running a build where a single dot of stress doesn't hurt you. I'm pretty sure people don't necessarily run double stress with the R3-A2 for "fun" but because you sometimes need to drop two stress to actually have your stress control count. 

Share this post


Link to post
Share on other sites

Stress normally hurts mobility and action economy, therefore.

When you are flying a toaster, you don't mind stressing the toaster, but you are still hoping your opponent is flying something that actually cares that much about being stressed, something that can change as the meta changes after the nerfs. Well, TLT still allows it to hurt several ships tho and be relevant even after having stacked a bunchload of stress.

Edited by DreadStar

Share this post


Link to post
Share on other sites
3 minutes ago, UnitOmega said:

If Stressbot was only once a turn, you're going to need to also level the playing field by nerfing all ships and situations where single stressing basically does nothing. Double-stress SHOULD be bad for you, it's meant to stack stress so you have to take stress effects and can't just shed them because you've got a wealth of green moves, or are running a build where a single dot of stress doesn't hurt you. I'm pretty sure people don't necessarily run double stress with the R3-A2 for "fun" but because you sometimes need to drop two stress to actually have your stress control count. 

I think that's part of the problem, because when you look at other stress options they're hugely limited (range/arc) while R3-A2 is not just 'any range, in arc or not' but also by far the easiest way to double-stress.  When dealing stress becomes a strong strategy it becomes very difficult to justify playing anything but R3-A2 as it's MILES ahead of other options.

Share this post


Link to post
Share on other sites
2 hours ago, Voitek said:

Same issue as usually always with the List Juggler data:

The entire analyse is based on TOP RESULTS DATA. Top8/16 lists in every tournament is the only what goes into consideration and populates the algorithm.

What about mediocre and low-performing results? 

Lets take a look at Bwing. position 27th of 40. Mediocre to Low. Yet Blair managed to get to the final game with TWO of them! On the same tournament, we probably had 5+ defenders and equal number of jumpamsters in the finals. But there was probably about 50 defenders in the tournament, and 20-30 jumpmasters. Which ship is best among those 3 mentioned? Well clearly not Bwing, according to the data! And it probably should be, if 2 of maybe 10 Bwings being in use in the entire tournament made it to top2!

It's obviously not the fault of the list juggler's algorithm - it's all about the limited data we are getting. But it clearly shows that this data is not some Ultimate Truthsaying Tool, and that is how people seems to get this data - as a undoubtful gospel. Well, how about nope!

1 hour ago, Stay On The Leader said:

Except this data is CERTAINLY not that.

It's neither a measure of popularity nor a measure just of Top-8 cuts.  It's a measure of success drawn from a data pool that is weighted overly-heavily towards Top-8 cuts.  If anything a ship that is hugely popular is at a disadvantage in this ranking as they can't ALL win.

54 minutes ago, Makaze said:

It's a fairly basic chart with a very clear statement about where the data came from and SotL has been very open with answering questions about his methodology. If you don't like the dataset then tough, it's the best we have available. If you don't like the methodology then suggest a better one. But quibbling about the semantic difference between "ranking" vs. "popularity in top 8 cuts" and demanding massive and nuanced disclaimers is just ridiculous.

No, it is a very big deal.  Because if we are weighing the entire study to the top of the lists, it's like discussing the Middle Ages based only on what the nobility were doing.

 

 

1 hour ago, FlipmodeSH said:

It is simple enough to have a quick discussion of 'How much data is in that section?' and let people draw their own conclusions. This all seems rather personal.

1 hour ago, GreenDragoon said:

Wow, seems to have struck a nerve for some...

42 minutes ago, Timathius said:

Thank you for writing this out. The only issue I had was the way it was presented with no real reference to the actual data used. 

27 minutes ago, Sunitsa said:

I see lots of abstract criticism in this thread but a suprising lack of example on why you are criticizing.

Well, don't dump a table without text would be my moral of the story here.

Share this post


Link to post
Share on other sites

If R3-A2 is the easiest to double-stress, that does probably reflect poorly on other stress mechanic combos - since it's in arc already and doesn't trigger on double-tap secondary weapons, so you need an ability like Gunner or the Y-Wing title to do it. 

But at the same time, I think that makes it kind of a chicken and egg thing? (Maybe that's not the right metaphor). R3-A2 is the easiest to double-stress, so people take it because they need to double stress in order to actually make their stress control actually "control" in some cases. And even now with Inspiring Recruit there's a pretty cheap way to counter double-stress. So, do we hope for ships, upgrades or errata to make the double-stress game easier over all, so people have more options? Do we keep double stress rare and difficult because of how strong it is, and accept that some builds will naturally ignore normal stressing? Does something need to be done about ships which are flexible enough to ignore a single stress token one way or the other? 

That's probably a more technical discussion which needs its own thread though. 

Share this post


Link to post
Share on other sites

Jumpmaster has maintained spot 1 since its release despite 3 attempts to nerf it.

Can FFG just ban that ship already (Maybe keep Dengar)? There isn't any quick errata that can fix it. It is just a broken chaise. Scum has enough good other stuff the faction wouldn't die without the toilet seat, in fact it would probably help bring list diversity. 

Share this post


Link to post
Share on other sites
57 minutes ago, Darth Meanie said:

No, it is a very big deal.  Because if we are weighing the entire study to the top of the lists, it's like discussing the Middle Ages based only on what the nobility were doing.

Yes, that'll remain true until everybody enters all the data for all the tournaments, unfortunately :D

It's even worse for matchup data. Very few TOs actually enter each match individually, for understandable reasons.

Share this post


Link to post
Share on other sites

InTo answer the tinfoil hat brigade:

1) there is no data manipulation on my part other than to remove ships played on vassal before their release date.

2) I've no dog in this fight, there's no ship I want to show is too good or too weak.  I didn't know what the result would and was interested to find out.

3) The reason I ran this was to see if it could be usefully added to my buying guides as an indicator of his value of ships change over time.

Share this post


Link to post
Share on other sites
51 minutes ago, Darth Meanie said:

 it's like discussing the Middle Ages based only on what the nobility were doing.

Technically, that's pretty much the way history is discussed, sadly. Look at the conception most people have of Caesar, as a "tyrant" and a "dictator."

Of course, who called him that? Members of the contemporary nobility like Cicero who were massively wealthy landowners taking advantage of the slave economy Caesar was trying to reform, which was what had made him so popular with the masses in Rome, and exactly why he was so feared by the Senators who assassinated him. I mean, that's not to paint Caesar as some kind of altruistic saint, because he was certainly ambitious and scheming. But most people's understanding of history really only plays at the top level of politics, because those were the people who shaped the events at the time. For large portions of history, the records which survived were the accounts by people who were influential at the time, or worked *for* people who were influential. It took centuries for a more nuanced look at the fall of the Roman Republic before modern historians started to really be able to do contextual analysis of the time period. It's why Gibbon's Decline and Fall of the Roman Empire was once considered and authoritative text for that period, but 200ish years later, it's mostly just a curiosity for casual readers and a reference book for historians because of all the primary sources he had collected at the time.

 

Though, that's tangential to your point you were trying to make (sorry, I love history), which I agree with. A chart with no data and no analysis is going to be misleading, or at best incomplete and foster discussion and argument. This chart the OP provided was a curiosity, at best. And, like all things that are curious, people talked about it.

Share this post


Link to post
Share on other sites
4 minutes ago, Stay On The Leader said:

InTo answer the tinfoil hat brigade:

1) there is no data manipulation on my part other than to remove ships played on vassal before their release date.

2) I've no dog in this fight, there's no ship I want to show is too good or too weak.  I didn't know what the result would and was interested to find out.

3) The reason I ran this was to see if it could be usefully added to my buying guides as an indicator of his value of ships change over time.

You have never been accused of any of those things. All that was asked is that you preface the data you presented, and say how thin of a data pool it was based on. 

 

You have been taking these things very personal. It is not meant to be. 

 

Share this post


Link to post
Share on other sites
9 minutes ago, haslo said:

Yes, that'll remain true until everybody enters all the data for all the tournaments, unfortunately :D

It's even worse for matchup data. Very few TOs actually enter each match individually, for understandable reasons.

The main problem is TOME. It doesn't care about lists and most TOs use it. 

Share this post


Link to post
Share on other sites
4 hours ago, Stay On The Leader said:

Post-nerf, that's correct.

But Aggressors have been in terminal decline over ever expansion of the last 18 months, so it's reasonable to assume their position is correct.  Or at least that if it's materially changed it's not been reflected yet in the events that have been played.

As one of the early top tier BroBot players from 2015 (Worlds 2016 #20, 6-2), I can attest that BroBots don't have the tools and cost efficiency any more to consistently compete with tier 1 squads. I have moved on from them for a long time now. If you fly them wonky then you can still catch people off guard, but they just don't have the brute force anymore to either punch through or withstand the insane number of dice that are now getting tossed around.

[edit]: By Worlds 2016, BroBots were clearly no longer tier 1 based on the empirical evidence (in my opinion). Only one of the 41 squads that went 6-2 was BroBots. There were several that were 5-3, but that's... only slightly better than average. That includes Phillip Booth, who Top 4'ed in 2015 and is an exceptional player. He did go on to win a Regionals this season, but overall I still don't think they're top tier anymore. They're still good, but not great. Qualitatively, I think they have gone down quite a bit since the 2016 Regionals season (see below), which was later reflected in the 2016 Worlds results.

 

Some general thoughts about the metrics. I prefer the following data analysis method as it is far more informative:

  1. Only look at squads that made the cut, or equivalent proportion of players. (i.e. use the top ~10% - 15% of a System Open rather than just the Top 8).
  2. Weight all results by points spent, not number of ships. Weighting by number unfairly punishes expensive ships, and unfairly presents cheap ships as more prolific than they really are. What you really want is how much of the stuff on the table is "X".
  3. Weight all results by attendance. It's OK to use store championships or kit tournaments, they just get weighted lower.
  4. Compute both the rate of occurrence (making into the cut or equivalent), AND the performance after reaching this minimum threshold.

This leads to two big advantages:

  1. The above can be used to make a scatter plot along 2 axis that contains much more information and is much more informative at-a-glance than just distilling each contestant down to one variable.
  2. Pruning out results below the Cut, and weighting by points spent and attendance gives a significantly cleaner picture of what the top-end meta really looks like.

Here is an example of what such a scatter plot looked like for the 2016 Regionals season. This was evaluating squad archetypes, so points spent was N/A in this case.

 

58a7dc8096fea_2016Regionals.thumb.png.32a8207865f719ed863113a37445067b.png

 

The data presentation can be cleaned up a little, with the clustering of low appearance / performance squads resulting in overlapping text, but you get the idea. I emailed the guy who wrote the Meta Analyzer tool all this, but I don't know if it will go anywhere. I could do it myself, except right now I don't want to:

  1. Learn SQL.
  2. Learn all the other languages to implement this on a web page.
  3. Host a website.

The List Juggler data is all there publicly, so anyone else can go create their own "meta analyzer" if they are capable and motivated. If anyone wants to be a code-monkey and implement the above (and potentially more, there are more ideas), I'm happy to help with the formulaic side of things.

Edited by MajorJuggler

Share this post


Link to post
Share on other sites
1 minute ago, Timathius said:

Still no good is it hosted somewhere? or trying to upload direct?

Was trying to upload direct. Copy - paste shows it in the post submission but not after I post. Still figuring out the new forum.

Edited by MajorJuggler

Share this post


Link to post
Share on other sites
19 minutes ago, MajorJuggler said:

The data presentation can be cleaned up a little, with the clustering of low appearance / performance squads resulting in overlapping text, but you get the idea. I emailed the guy who wrote the Meta Analyzer tool all this, but I don't know if it will go anywhere.

I don't know either, but when I start compiling graphs, this one will be at the top of the list :)

Yeah, anybody can build their own analysis of the Lists Juggler data. And meta-wing is OSS as well, contributions are very welcome if anybody feels like it.

Share this post


Link to post
Share on other sites
1 hour ago, Stay On The Leader said:

InTo answer the tinfoil hat brigade:

1) there is no data manipulation on my part other than to remove ships played on vassal before their release date.

2) I've no dog in this fight, there's no ship I want to show is too good or too weak.  I didn't know what the result would and was interested to find out.

3) The reason I ran this was to see if it could be usefully added to my buying guides as an indicator of his value of ships change over time.

Come on over to my new thread. . .you can chew me out and vent a little steam. ;)

1 hour ago, VaeVictis said:

Technically, that's pretty much the way history is discussed, sadly. Look at the conception most people have of Caesar, as a "tyrant" and a "dictator."

Of course, who called him that? Members of the contemporary nobility like Cicero who were massively wealthy landowners taking advantage of the slave economy Caesar was trying to reform, which was what had made him so popular with the masses in Rome, and exactly why he was so feared by the Senators who assassinated him. I mean, that's not to paint Caesar as some kind of altruistic saint, because he was certainly ambitious and scheming. But most people's understanding of history really only plays at the top level of politics, because those were the people who shaped the events at the time. For large portions of history, the records which survived were the accounts by people who were influential at the time, or worked *for* people who were influential. It took centuries for a more nuanced look at the fall of the Roman Republic before modern historians started to really be able to do contextual analysis of the time period. It's why Gibbon's Decline and Fall of the Roman Empire was once considered and authoritative text for that period, but 200ish years later, it's mostly just a curiosity for casual readers and a reference book for historians because of all the primary sources he had collected at the time.

 

Though, that's tangential to your point you were trying to make (sorry, I love history), which I agree with. A chart with no data and no analysis is going to be misleading, or at best incomplete and foster discussion and argument. This chart the OP provided was a curiosity, at best. And, like all things that are curious, people talked about it.

It was the best analogy I could come up with. :)  OTOH, you're right. . .history is told by the winners and those with the power to enshrine it.  Why should X-Wing break the mold? :P

Share this post


Link to post
Share on other sites
9 minutes ago, gamblertuba said:

Today I have learned that different people use the word "meta" to mean very different things.

there is only one meaning:

 

Metagame comes into play in any game where no single strategy is dominant and opposing sides are aware of multiple strategies that can succeed dependent upon opponents' actions. In order to perform at the highest level, it then becomes necessary to think about what your opponent thinks you will do (which may depend on what he thinks you think he thinks he will do, etc.) and to make decisions based on clues regarding what level they are thinking on.

You try to guess what will be popular and/or effective and you find ways to beat that.

 

Edited by Thormind

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...