14 hours ago, Ardaedhel said:

In fairness, this is a very tall order for any list. Part of the reason I keep getting away with my cheesy shrimps locally is the prevalence of bomber heavy play from multiple other top players. It's difficult enough to address either kind of threat individually--both simultaneously is really tough.

Not to say I think it can't be done. Just that it's not a challenge limited to ISDs.

In fairness, I agree with your assertion. But these are the dominant and played poles of the meta atm.

13 hours ago, Grey Mage said:

Ive run double ISD, and its fun. It has its issues.... but its fun, and you win some, you lose some, like any other list.

Does it have issues? Of course. It should.

If you dont take a fighter screen there should be consequences. If you dont take activations there should be consequences. If you dont take significant ship based weaponry there should be consequences. If you dont take objectives that your fleet is well equipped to handle/exploit, there should be consequences. Thats the nature of of choice.

To often I see people wanting the advantages without the consequences. When I fly double ISD I know Im going to be out activated and that I have to plan for that to be the case. That its going to be far less forgiving. That the squadrons I take in my remaining 100pts are going to need to either deal significant antiship damage or be a **** good fighter screen, but that I probably cant afford to mix the two.

Balanced should never be confused with without drawbacks.

2 ISD1 with SFO, OE x17
3 Gozanti + Moffy J
Howl + 4 Ints.
Variables defined: 5 activations, 5 total AA squadrons. What's the whole thing about no fighters and not enough activations?

12 hours ago, shmitty said:

How do you define or know something would auto-lose?

i have my own internal definition of that, but curious of yours.

Shmitty, what's your definition?

Technically auto-lose would mean a 0% chance of winning. But that of course doesn't happen. If we define normal gameplay as being typically tight and competitive with understanding of the meta, then it also means occasionally, someone does something bone-headed over the course of a high number of sampled games.
So, in practical terms, I'd say anything lower than 25% in an otherwise 50-50 game is "auto-lose". As reference, Sc2, Wows, Magic, Xwing, LoL tend to consider above 55-45 split to be imbalanced, and things near 60-40 require immediately game-errata action.

Now, those are statistics, which require someone wonderful like @shmitty and @Baltanok to collect (thank you).

However I can also make a point about ease of play or avenues to victory. Strategies with multiple paths to victory, assured victory through inaction (immediately defined advantage), "cheesy" ease of execution as opposed to extreme difficulty in learning the counter, high durability of battle plan, or too high ability to switch tactics are considered easier to play, and can also be considered too powerful or not good for the game even if they do not technically exceed numerical levels of winrate imbalance.

----

(Spoiler: squadrons)

In the case of squadrons, you have multiple paths to victory: alpha strike with FCT->Yavaris, or picking off small ships, or winning on points based on the 3 squadron point scoring objectives, and MoV advantage (well played squadrons tend to win high, and can usually afford a low score or tie towards late tournament). I'll not argue the middle ones. The battle plan is very durable: 3 transports tend to mean some escape in a skirmish, avoiding tabling, squadrons in mass are very hard to take out, and require usually a similar fleet of mass squadrons tooled for AA. Its ability to change tactics based on how many ways it can win also makes it easy to take on all comers.
Basically, do you agree with this : a good Rieekan 134 rebel bomber list (has some AA and intel) plus Yavaris 2AA dice and 5 activations is good and has plays against ANY list it comes against. It has no viable tier1 meta hard counter and no inherent weaknesses.
As opposed to fleets with even 1 large ship: Need upgrades to make relevant damage, usually weak to squadrons, has a lower activation count. Immediately simply by existing it has inherent 2 tier1 meta weaknesses.
You could also say that MSU has some inherent weaknesses: easy to lose ships, susceptible to squadron alpha strikes on small ships. However, these can be mitigated by proper activation control. Activation control is a particularly powerful tool in Armada.

12 hours ago, Beatty said:

Now I have to ask a question, not insinuating anything but is it the list or the player that makes the list you faced unbeatable? If you are saying an average player can take and win a tournament with it please post the list so we can pick it apart and help others defend against it. If it's the player there isn't too much that can be done because this game favors the player's abilities over list building more than any Miniature Game I've played.

Definitely the list archetype. I play both the squadron list and the test list in testing swaps, self-testing and casual games. I also have variants of the squadron lists that include weakness mitigations from observed games. The squadron list wins, sometimes via face-rolling tabling, via running away on points, or just in overall points and lost squadrons and ships combined. That's why I don't care for the hard counter lists that can technically table the list, if they can't overcome points lost or that the hard counter list is so AA gimpy that it loses firepower for regular games against other opponent you inevitably face.

That said, the player who concocted the original is really really good, and most of the locals have a hard time beating him. Hey, winning regionals is a thing.

I actually like to think I'm not an under-average list builder, I'm pretty meta-cognizant and avoid traps, I've pointed out macro theories on objectives (points vs non-points) and game balance. But I can't get anything in the list building phase to help mitigate this issue, and I've pointed that out other lists people share too, that they do not have enough AA to not be at an inherent disadvantage to mass squadrons first wiping out their cover, then going full bombing run.

9 hours ago, Blail Blerg said:

Shmitty, what's your definition?

I suppose for me auto-win can't be determined until deployment. I can in my head picture two fleets that could each be in an auto-win position over the other based on how they deploy.

If both fleets are skillfully constructed it would be exceptionally rare that I would think one fleet would auto-win over the other based on looking at the fleet lists alone. Now, a poorly constructed fleet vs a well tuned fleet is different, but I don't think that is your point.

Fleet construction can give an advantage to one side based on matchup, but I have come to think of practice as being more important than fleet construction.

9 hours ago, Blail Blerg said:

Technically auto-lose would mean a 0% chance of winning. But that of course doesn't happen. If we define normal gameplay as being typically tight and competitive with understanding of the meta, then it also means occasionally, someone does something bone-headed over the course of a high number of sampled games.
So, in practical terms, I'd say anything lower than 25% in an otherwise 50-50 game is "auto-lose". As reference, Sc2, Wows, Magic, Xwing, LoL tend to consider above 55-45 split to be imbalanced, and things near 60-40 require immediately game-errata action.

That seems fair. Fortunately, I don't know that anything in the game right now quite meets that criteria. I was pretty convinced Demo and Rhymer needed a nerf during the Wave 2 Regionals. I discussed it with one of the FFG Devs and they were looking at it at the time, but it apparently didn't meet their internal criteria for a hard nerf. We know they have one based on IA and X-Wing. From talking with them, if they look and see one particular build and its known counter as dominating the top tables they would consider it. 4x4 and Rebel Sabs in IA being the best example of this.

It wasn't discussed, but they seem to obviously have a lower threshold for soft-counters. I think that is there preferred method as well. JJ as an attempt to improve VSDs as an example.

In looking at the recent data, I don't see any one strategy in Armada as being nearly dominate enough to have their attention.

9 hours ago, Blail Blerg said:

Basically, do you agree with this : a good Rieekan 134 rebel bomber list (has some AA and intel) plus Yavaris 2AA dice and 5 activations is good and has plays against ANY list it comes against. It has no viable tier1 meta hard counter and no inherent weaknesses.

As opposed to fleets with even 1 large ship: Need upgrades to make relevant damage, usually weak to squadrons, has a lower activation count. Immediately simply by existing it has inherent 2 tier1 meta weaknesses.
You could also say that MSU has some inherent weaknesses: easy to lose ships, susceptible to squadron alpha strikes on small ships. However, these can be mitigated by proper activation control. Activation control is a particularly powerful tool in Armada.

I don't agree. It has a serious inherent weakness in that it has only one real threat, the squadron ball. The Pelta and Yavaris have very limited firepower against ships, only one strong arc each, and nothing to boost their firepower. The squadrons are a serious threat, but are not insurmountable. This fleet really wants to go straight ahead. It will slaughter anything that flies straight towards it. It is highly dependent on Squadron commands so putting it in a position where a squadron command in a sub-optimal choice hurts it greatly.

Regardless of fleet makeup, I would fight it by threatening both flanks. It reduces the firepower of the ships further, makes coordinating the squadrons tougher, and can dilute the squadrons or put them out of position. Even a light ship like a TRCR90 or Arquitens can kill the Pelta or Yav from the flank and cannot totally be ignored.

10 hours ago, Blail Blerg said:

2 ISD1 with SFO, OE x17
3 Gozanti + Moffy J
Howl + 4 Ints.
Variables defined: 5 activations, 5 total AA squadrons. What's the whole thing about no fighters and not enough activations?

I win alot of games with a similar list, whats the whole thing about it not being viable?

48 minutes ago, shmitty said:

It h as a serious inherent weakness in that it has only one real threat, the squadron ball. The Pelta and Yavaris have very limited firepower against ships, only one strong arc each, and nothing to boost their firepower. The squadrons are a serious threat, but are not insurmountable. This fleet really wants to go straight ahead. It will slaughter anything that flies straight towards it. It is highly dependent on Squadron commands so putting it in a position where a squadron command in a sub-optimal choice hurts it greatly.

Regardless of fleet makeup, I would fight it by threatening both flanks. It reduces the firepower of the ships further, makes coordinating the squadrons tougher, and can dilute the squadrons or put them out of position. Even a light ship like a TRCR90 or Arquitens can kill the Pelta or Yav from the flank and cannot totally be ignored.

Thank you for this. An Ozzel MSU list with Demo hitting the same turn that the TIEs come in and engage the squadrons should crack the Yavaris or Pelta, giving you a chance to run with everything else. Heck, Ackbar doom pickle rolling up the sides can email on a lot from long range.

1 hour ago, geek19 said:

Ackbar doom pickle rolling up the sides can email on a lot from long range.

From: [email protected]
To: [email protected]

Subject: Damage

33 minutes ago, Ardaedhel said:

From: [email protected]
To: [email protected]

Subject: Damage

I had to sit and think to figure out the dice for a second. There's a lot happening there.

11 minutes ago, Madaghmire said:

I had to sit and think to figure out the dice for a second. There's a lot happening there.

MC80C, Ackbar, EA, Defiance , CF. Or maybe an Opening Salvo or Most Wanted in there.

I definitely did not just copy-paste some icons randomly into a graphic.

3 minutes ago, Ardaedhel said:

MC80C, Ackbar, EA, Defiance , CF, Most Wanted. Or maybe an Opening Salvo in there.

I definitely did not just copy-paste some icons randomly into a graphic.

My process went MC80A, Ackbar, Defiance, CF and either EA/or High Cap Ion Turrets

9 hours ago, shmitty said:

In looking at the recent data, I don't see any one strategy in Armada as being nearly dominate enough to have their attention.

I've sliced the data in a number of ways for my own amusement, trying to break archetypes out by points spent. I came up with a threshold for each fleetbuilding lobe (Large/Medium, mostly MSU, Heavy Squads), and if a fleet was over the threshold in 1 area, and not the others, I assigned it to that lobe. I also considered fleets that did not exceed any of the thresholds as "balanced" and made them the central node. (Picture 4 triangles, making up a larger triangle, like the Hojo clan crest)

Squad-heavy lists are more likely to go to the top tables, but it isn't as biased as the Rhymerball archetype was in wave 2. 120+ squad pt lists made up about 25% of the lists that fit one of my 4 categories, and were 37% of top 4, 16% of bottom qtr. compare to Wave 2 rhymerballs: 18% overall, 37% top, 10% bottom.

As an example, the lists that fell into the MSU lobe were evenly balanced between top4 & bottom quarter (18.5% overall, 18% bottom, 17.5% top 4)

Big ship & "balanced" nodes were below expected, but again, not as strongly as W2:Rhymer.

the imbalance quotients (% above expected: top + % below expected bottom) were:

W5:Squad 84%

W5:MSU: -3%

W5:Large: -41%

W5:Balanced: -37%

For reference:

W2:Rhymer 150%

@Baltanok That's some really cool data, though I've never seen it addressed as percentages of taken vs top4.

Is it possible to generate winrates for archetypes?

16 minutes ago, Baltanok said:

I've sliced the data in a number of ways for my own amusement, trying to break archetypes out by points spent. I came up with a threshold for each fleetbuilding lobe (Large/Medium, mostly MSU, Heavy Squads), and if a fleet was over the threshold in 1 area, and not the others, I assigned it to that lobe. I also considered fleets that did not exceed any of the thresholds as "balanced" and made them the central node. (Picture 4 triangles, making up a larger triangle, like the Hojo clan crest)

Squad-heavy lists are more likely to go to the top tables, but it isn't as biased as the Rhymerball archetype was in wave 2. 120+ squad pt lists made up about 25% of the lists that fit one of my 4 categories, and were 37% of top 4, 16% of bottom qtr. compare to Wave 2 rhymerballs: 18% overall, 37% top, 10% bottom.

As an example, the lists that fell into the MSU lobe were evenly balanced between top4 & bottom quarter (18.5% overall, 18% bottom, 17.5% top 4)

Big ship & "balanced" nodes were below expected, but again, not as strongly as W2:Rhymer.

the imbalance quotients (% above expected: top + % below expected bottom) were:

W5:Squad 84%

W5:MSU: -3%

W5:Large: -41%

W5:Balanced: -37%

For reference:

W2:Rhymer 150%

Nice, you gots some stats chops, sir, and I'm loving it.

Out of curiosity, was your criteria for the Large category "had any large ship" ? That would make sense at first blush, but would also define "balanced" as "did not have a large ship" , which seems... not quite optimal.

Cool stuff @Baltanok and an interesting way to establish some fleet categories. I know percentage of points spent on various areas is something I plan to get into the next iteration of the worksheet.

What did you use as cutoffs?

50 minutes ago, Ardaedhel said:

Nice, you gots some stats chops, sir, and I'm loving it.

Out of curiosity, was your criteria for the Large category "had any large ship" ? That would make sense at first blush, but would also define "balanced" as "did not have a large ship" , which seems... not quite optimal.

23 minutes ago, shmitty said:

Cool stuff @Baltanok and an interesting way to establish some fleet categories. I know percentage of points spent on various areas is something I plan to get into the next iteration of the worksheet.

What did you use as cutoffs?

I picked 120+ for squads, 110+ for large/medium ships, 144+ for small combatants. Fleets that met 2 criteria were excluded, while fleets that met zero were balanced.

Those were picked to try to make the populations in each lobe about the same size, and was roughly +1 standard deviation above average.

I'm not satisfied with the large ships cutoff, since it makes 1 ISD a "large ship" list, while 1 MC80 is a "balanced" list. I will redo it with 111+ pts being a "large ship" list. That would be 2 medium or large ships, while 1 large ship lists would be "balanced."

(ships were also calculated at the cost of their cheapest variant, since we don't have granular enough data to get more precise than that.)

Since "balanced" and "large ships" groups had similar performance, it probably won't make much difference, but it's probably worth the effort.

30 minutes ago, Baltanok said:

I'm not satisfied with the large ships cutoff, since it makes 1 ISD a "large ship" list, while 1 MC80 is a "balanced" list. I will redo it with 111+ pts being a "large ship" list. That would be 2 medium or large ships, while 1 large ship lists would be "balanced."

Imbalance quotients:

W5:Squad 91% (124+)

W5:MSU: 11% (153+)

W5:Large: -44% (144+) (the next larger amount of big ship points, or the 56 most big ship heavy lists)

W5:Balanced: -42%

So, moving the cutoffs a little further out made some small changes to increase the MSU & Squad "Imbalance quotients" at the expense of Balanced & Large/Medium ship lists.

The biggest complaint about squadron/flotillias is how they almost need each other and you cannot use one without the other, furthermore not taking any will leave you at a serious disadvantage since flotillias have the cheapest squadron value per point. I think when the Delta cruiser comes out with its budget 4 squadron value you might see less flotillias for the Rhymer ball and see more Carrier ships.

1 hour ago, Marinealver said:

The biggest complaint about squadron/flotillias is how they almost need each other and you cannot use one without the other, furthermore not taking any will leave you at a serious disadvantage since flotillias have the cheapest squadron value per point. I think when the Delta cruiser comes out with its budget 4 squadron value you might see less flotillias for the Rhymer ball and see more Carrier ships.

I think the data shows that any fleet you have should have a flotilla.

--

Baltanok: Why 120+ points for squadrons btw? Is it such that there aren't that many that go from say 100 to 120? What about tracking for over 100 points in squadrons + at least 1 instance of BCC? That sounds like a bomber archetype. (Although you can make an argument for mixed here, but that at over 100 points of bombers and not AA, bomber damage is key for that list). I think it would be nice to see your assumptions a bit more clearly explained.

I still also see that in both cases in wave 5, taking 120+ bombers has a 87% or 91% higher rate of prevalence in top40. That's still a pretty huge swing. Do we really want a game that you have a much higher chance of winning if you always take 100+ squadrons and only 300 for ships and upgrades, or just MSUs?

1 hour ago, Blail Blerg said:

Baltanok: Why 120+ points for squadrons btw? Is it such that there aren't that many that go from say 100 to 120? What about tracking for over 100 points in squadrons + at least 1 instance of BCC? That sounds like a bomber archetype. (Although you can make an argument for mixed here, but that at over 100 points of bombers and not AA, bomber damage is key for that list). I think it would be nice to see your assumptions a bit more clearly explained.

I am separating the archetypes based on how they differ in points spent from the "average" fleet. There are 263 fleets in the "full data" tab that I can calculate the total points spent on squadrons, Med/large ships, and small combatants. Trying to adjust for individual upgrade cards increases the chance of my judgement skewing the conclusions. Generally, try to make your criteria as simple & objective as possible.

The median (i.e 131st or 132nd largest point spend) is 93 points in squads, 96 in medium/large ships, and 98 in small combatants. I wanted to make sure the cutoff between a "balanced" fleet and a "squadron-heavy" fleet is somewhere between that value, and the maximum value of 134 squad points. So, I could pick 110, 120, 130, whatever I wanted. It's just a matter of saying "I want the top X% of squadron lists" and figuring out what point-spend matches that number. Or saying I want 40 fleets in each lobe, and what point spend is needed to break that out.

Then I have to figure out the cutoffs for the other two archetypes for a similar percentage, and start filtering the lists to find just the ones I want for each group.

But, the bigger I make X%, the more lists get picked up by more than one category, and are therefore discarded, and the smaller I make X%, the fewer lists get put in each lobe except "balanced." That would result in small group populations, and make conclusions less reliable due to small sample size.

Overall, there are going to be some lists that are right on the border between two archetypes, and different people will disagree on where the cutoffs should be. But unless there is a massive step-change in performance at some break-point, it shouldn't matter too much if I move the cutoff a few points in either direction. The conclusions should be similar.

I'm not sure why you'd discard lists that do fit into more than 1 category. That seems like it muddles the picture.

One of those Rhymerball archetypes you used to find was Rhymer + Demo lists in wave 2.

Come wave3/4 you could actaully run a 6MSU Demo + Rhymerball in full plus a hefty bid. And this list is really strongly both archetpyes.

Anyways. So. what do you conclude from the data?

I discard multi-archetype lists because they presumably fly differently than a "pure" archetype, and might show different distributions than the pures. Also, very few of them. I think I have discarded 10-ish lists for fitting into multiple archetypes out of 260-ish. Given that there are 3 hybrids, none of them have enough population to support analysis.

Wave 2 data is what it is, I'm not spending a lot of time to derive the archetypes from first principles again.

Conclusions are hard, but I think I'll say something like: Anti-squad play could take a small boost without too much danger of major disruption. MSU's are about right. Note that some combo's take some time after introduction to really hit the battle-mat, so we may already have the tools we need, but just aren't using yet. (Ten Numb, for example)