4 hours ago, player3010587 said:

internet glitched and reeeposted multiple times, lol

i agree

28 minutes ago, jagsba said:

L'ulo is also brought the most, and (no stats to back this up but) probably the most likely to be fit into a 'bad' list. Zari's status as second best suggests there's a lot to be said about 5 A's being good and L'ulo being brought down by worse lists.

Probably a good assumption, yeah. I would like to note how happy it made me that you clearly stated it as an assumption, not a definite truth ♥

46 minutes ago, Brunas said:

Lulo is the worst performing RZ2, outside of whatever the name of the middle generic one is.

Huh.

I expect @jagsbais correct about high frequency being an issue for Lulo, but I'm a bit surprised. Is Poe the best performing T70?

There's a discussion we could have as a community about the merits and weaknesses of outcome based results (data driven) and theoretical results, and how and when one should try to reconcile the two.

1 minute ago, Biophysical said:

There's a discussion we could have as a community about the merits and weaknesses of outcome based results (data driven) and theoretical results, and how and when one should try to reconcile the two.

All the T70s are roughly equivalent, ~24% - some are 23% cut, some 25%, except for snap rolling deep at 30%.

19 minutes ago, Brunas said:

Probably a good assumption, yeah. I would like to note how happy it made me that you clearly stated it as an assumption, not a definite truth ♥

its a conclusion drawn from numbers so it must be true, for example:

7 minutes ago, Brunas said:

All the T70s are roughly equivalent, ~24% - some are 23% cut, some 25%, except for snap rolling deep at 30%.

daddy tubbs has a 100% cut rate (played once, made cut once). Best T70.

7 minutes ago, Biophysical said:

There's a discussion we could have as a community about the merits and weaknesses of outcome based results (data driven) and theoretical results, and how and when one should try to reconcile the two.

data is inherently garbage because people don't understand it and the people that do understand it can manipulate it to support their beliefs. theoretical results are garbage because they're not supported by data.

16 minutes ago, jagsba said:

data is inherently garbage because people don't understand it and the people that do understand it can manipulate it to support their beliefs. theoretical results are garbage because they're not supported by data.

i agree

6 hours ago, RStan said:

I wonder how much of the prequels can be told using only gifs...

All you need is one:

29 minutes ago, jagsba said:

its a conclusion drawn from numbers so it must be true, for example:

daddy tubbs has a 100% cut rate (played once, made cut once). Best T70.

data is inherently garbage because people don't understand it and the people that do understand it can manipulate it to support their beliefs. theoretical results are garbage because they're not supported by data.

11 minutes ago, Brunas said:

i agree

And we revel in our ignorance as thousands of years of progress burns down around us.

34 minutes ago, jagsba said:

data is inherently garbage because people don't understand it and the people that do understand it can manipulate it to support their beliefs. theoretical results are garbage because they're not supported by data.

way to narrowly avoid saying "garbage people".

But I'm a refined enough person to read between the lines.

16 minutes ago, Biophysical said:

And we revel in our ignorance as thousands of years of progress burns down around us.

Life is about survival of the fittest, and numbers and math can't exist without humans for support. They're a parasite, and we should end them.

1 minute ago, Brunas said:

Do what you want, man.

FTFY.

17 minutes ago, Tlfj200 said:

way to narrowly avoid saying "garbage people".

It'd be redundant. Especially when talking about math.

1 hour ago, jagsba said:

Too true. But the base model with a missile is ok, and I’ve found some work for the named pilots. Not as good as the named Strikers, perhaps, but Oddball with Synched Console to pass the lock to a protorp wielding with focus Arc?

I meant all A-wings down [maybe not Lulo, but everything else]. Also free daredevil. Gottagofast.

I will also be lobbying for price reductions and free daredevil for TIE adv. Prototypes, M3-As, TIE FOs, with boost for the last two. Gottagofast

This Message brought to you by the gottagofast lobby for two die ships that go fast.

1 hour ago, Biophysical said:

Huh.

I expect @jagsbais correct about high frequency being an issue for Lulo, but I'm a bit surprised. Is Poe the best performing T70?

I don't think so, it looks like Ello, with Poe actually at the bottom, but I may be bad at graphs.

Confirmed bad at graphs, and also reading. Anyway, we already knew Snap was best because Snap goes fast, and Tubbs lets other go fast more. GOTTAGOFAST

What graphs and stats are you all talking about?

1 hour ago, Biophysical said:

There's a discussion we could have as a community about the merits and weaknesses of outcome based results (data driven) and theoretical results, and how and when one should try to reconcile the two.

They are both parts of the same pie, so they touch, blend and are consumed at the same rate when enjoyed.

1 hour ago, jagsba said:

..data is inherently garbage because people don't understand it and the people that do understand it can manipulate it to support their beliefs. theoretical results are garbage because they're not supported by data.

True about data being stats (onset punch line here), but in the pie analogy, you can tell me it’s apple, but if it looks and tastes like cherry, will be at odds.

As far as the crust, without that, what’s going to hold the fruit portion when it’s served?

Enjoy!

...now where’s my tea?

2 hours ago, Biophysical said:

There's a discussion we could have as a community about the merits and weaknesses of outcome based results (data driven) and theoretical results, and how and when one should try to reconcile the two.

I don't know what you mean here, but I definitely want to. Can you explain? (Note, I stopped reading to reply, so apologies if I missed more transparency in a later post.)

9 minutes ago, Jeff Wilder said:

I don't know what you mean here, but I definitely want to. Can you explain? (Note, I stopped reading to reply, so apologies if I missed more transparency in a later post.)

I don’t know how basic you want to go, but there is a clear and distinct line (that most people don’t notice) between “this thing has done well this many times at this rate” and “this thing has done well and is good.”

We still don’t know if Kassis Trelix is good, and we can’t really know, but we know it hasn’t succeeded and no one has tried. We can infer he is bad, but that is theoretical.

We theoretically know L’ulo is the best Rz-2. Data tells us Zari has a better conversion rate. We can attempt to reconcile that with more theory: people tend to fit L’ulo into more lists (which can be bad) where Zari really only shows up in 5 a-wings (which is definitely decent or better).

Did you guys ever talk about the UK national soccer stats errors? Some guy noticed that the higher the number of passes the more likely they were to score, so they spent decades trying to pass a bunch, not realizing more passes correlated with not turning the ball over. He did the effort to collect the data, drew the wrong conclusion (even though the stats supported it) and everyone just went with it.

8 minutes ago, Jeff Wilder said:

I don't know what you mean here, but I definitely want to. Can you explain? (Note, I stopped reading to reply, so apologies if I missed more transparency in a later post.)

Data means something because it is actually stuff that happened, but we don't know the reason why something happened. Say X squad has a 50% conversion rate. Is it that good? Is it run heavily by excellent players more than others? Is the sample size meaningful?

If it really is that good, is it incredibly obvious? Pre-nerf triple jumps is obviously good. Post deadeye nerf it was largely abandoned, until it was shown to still be awesome. Same for Palpatine. Post deadeye nerf Jumps were always good. Post Palp need Palp was always good. Even if they weren't played they were good. There was an underlying reason that they were good even if people didn't believe they were good until later.

Consequently, it's interesting, for me at least, to think about interactions and under what situations things are good or not good, independent of how they're fielded by the community.

Basically, any one particular source of information on a subject is incomplete.

From whatever angle you approach something, you always need to have ALL the angles covered to be able to draw a definite conclusion.

And we live in quantum times.

So whatever feels right, man.

51 minutes ago, Biophysical said:

 Consequently, it's interesting, for me at least, to think about interactions and under what situations things are good or not good, independent of how they're fielded by the community.

Preferences have a big impact on how things tend to evolve especially when the Meta resets every couple months, becomes less important later on.

Players...

• ...are on high alert looking to try the best possible Turret.
• ...prefer the highest initiative thing that jousts.
• ...defer to lower ship count lists, two to three, ideally two.
• ...avoid low init ships, unless they can individually soak up damage.
• ...badly want at least one good reliable "trick" that involves minimal skill to use.
• ...do not have much patience for anything that needs to avoid shots.

I always have to remind myself that the Meta is what is good against what other players are willing to bring.

Hey so have people had enough games to agree that Gas Clouds are really dumb?

7 minutes ago, Boom Owl said:

Hey so have people had enough games to agree that Gas Clouds are really dumb?

Dumb and good?

I only had them in one game and it went ok. Losing the action is still a bit bad, but I might reverse my initial opinion. Flying 4 health ships, being completely safe in case of accidental clipping is actually really feelsgoodman