34 minutes ago, Sciencius said:

Very interesting analysis, unlike MajorJugglers pure mathemathical jousting models, you are simply letting the tournament data "speak" for themselves which is in some way the best. Thank you for this. I do have a few constructive comments:

1. I think you should rename the thread title to "New Idea for Point Costs - A Data- and Percentile-Driven Approach"

2. I noticed that practically all pilots in your spreadsheet have a reduced cost - practically all of them. That strikes me as odd - If every pilot gets a price-reduction then nothing is achieved and simply leads to inflation. I am wondering if the data is driven by your "scarcity-boost"? Would you kindly re-upload the spreadsheet, but this time, mark if the price-adjustment comes from adjustment or from scarcity bonus.

3. This has been mentioned, but data should also be weigthed by tournament-participant-size being top-dog in a 64-man tournament is a totally different beast than winning your 8-person laugh-it-up-fuzz-ball tournament kabooze.

1. Maybe so. The original title is kind of click-bait-y. The real reason I named it that way is because I am by no means a statistician of any kind and I did this purely for my own amusement.

2. 60 pilots go up, 47 stay the same, 220 go down. This does mean that about 2/3 are getting a price reduction, which is not insignificant and hadn't really occurred to me before. This could lead to a sort of inflation, but the intent was to bring every pilot to the average level, and it seems the ubiquity of the most efficient pilots had made 2/3 of pilots suboptimal. It would be difficult to counteract this under my model. If the model were to succeed in making every pilot playable, does it matter that you might get slightly more in a list than you could before? Perhaps it does. I could have used the Academy Pilot as the benchmark instead of the overall average, but I feel the AP has had its results skewed by Howlrunner. In terms of adding an indicator of whether the change came from performance or scarcity, that's awfully hard to do because all of them are affected by both except the ones where there was no data that simply got the maximum scarcity boost. If you download the excel spreadsheet, you can see all the data for yourself.

3. I considered both. In most cases it hardly made any difference (neither Kylo's nor Guri's costs would be changed by such a reflection, nor would the 5 Cavern Angels Zealots) and I found the extra data was necessary. Additionally, I personally believe it's a little bit of a trap to price things according to their use by the best of the best players; it's a personal ideology thing, but I feel like there are things that excellent players should get better use out of than the average person. For example, Soontir, Vader, and Whisper would have gone up more under that system because they tend to be played very well at a higher level. Similarly, the Ups wouldn't have gotten as much of a break because top players are great at slow-rolling, action management, and range control. In the end, I don't think it's that big a deal and changing it would require manually re-entering 360 percentages, which I'm not terribly eager to do. This was intended as a fairly rough exercise anyway.

12 hours ago, Sciencius said:

Very interesting analysis, unlike MajorJugglers pure mathemathical jousting models, you are simply letting the tournament data "speak" for themselves which is in some way the best. Thank you for this. I do have a few constructive comments:

1. I think you should rename the thread title to "New Idea for Point Costs - A Data- and Percentile-Driven Approach"

2. I noticed that practically all pilots in your spreadsheet have a reduced cost - practically all of them. That strikes me as odd - If every pilot gets a price-reduction then nothing is achieved and simply leads to inflation. I am wondering if the data is driven by your "scarcity-boost"? Would you kindly re-upload the spreadsheet, but this time, mark if the price-adjustment comes from adjustment or from scarcity bonus.

3. This has been mentioned, but data should also be weigthed by tournament-participant-size being top-dog in a 64-man tournament is a totally different beast than winning your 8-person laugh-it-up-fuzz-ball tournament kabooze.

Both approaches are very useful! Ideally the developers should be both predicting ship value ahead of time, and then also looking closely at tournament results after the fact. Alex has said that they do look at tournament results, but probably not in any sort of mathematically rigorous way like this.

The third leg is to get quantitative data to tune your models, from actual games. I.e. how much more often do phantoms get to shoot due to their decloack shenanigans? This directly affects their cost efficiency, which hopefully is accounted for correctly in the models.

Usually when a ship is at the extremes of tournament usage, you can go back and look at its cost efficiency and that will tell you why. Sometimes you might make a wrong assumption for a certain ship, cost it based on that, and then people figure out that it is either under/overcosted, and it shows up in tournament results accordingly.

23 minutes ago, MajorJuggler said:

Both approaches are very useful! Ideally the developers should be both predicting ship value ahead of time, and then also looking closely at tournament results after the fact. Alex has said that they do look at tournament results, but probably not in any sort of mathematically rigorous way like this.

The third leg is to get quantitative data to tune your models, from actual games. I.e. how much more often do phantoms get to shoot due to their decloack shenanigans? This directly affects their cost efficiency, which hopefully is accounted for correctly in the models.

Usually when a ship is at the extremes of tournament usage, you can go back and look at its cost efficiency and that will tell you why. Sometimes you might make a wrong assumption for a certain ship, cost it based on that, and then people figure out that it is either under/overcosted, and it shows up in tournament results accordingly.

I wholeheartedly support this. And I am sorry if it sounded like I devalued your work MJ, I certainly do not, I found your "jousting"-math very interesting and very important. The main problem with a pure tournament-data driven approach, is the pricing of the pilots rarely being played (for what ever reason) i.e. the scarcity-bonus as used here.

12 hours ago, ClassicalMoser said:

1. Maybe so. The original title is kind of click-bait-y. The real reason I named it that way is because I am by no means a statistician of any kind and I did this purely for my own amusement.

2. 60 pilots go up, 47 stay the same, 220 go down. This does mean that about 2/3 are getting a price reduction, which is not insignificant and hadn't really occurred to me before. This could lead to a sort of inflation, but the intent was to bring every pilot to the average level, and it seems the ubiquity of the most efficient pilots had made 2/3 of pilots suboptimal. It would be difficult to counteract this under my model. If the model were to succeed in making every pilot playable, does it matter that you might get slightly more in a list than you could before? Perhaps it does. I could have used the Academy Pilot as the benchmark instead of the overall average, but I feel the AP has had its results skewed by Howlrunner. In terms of adding an indicator of whether the change came from performance or scarcity, that's awfully hard to do because all of them are affected by both except the ones where there was no data that simply got the maximum scarcity boost. If you download the excel spreadsheet, you can see all the data for yourself.

3. I considered both. In most cases it hardly made any difference (neither Kylo's nor Guri's costs would be changed by such a reflection, nor would the 5 Cavern Angels Zealots) and I found the extra data was necessary. Additionally, I personally believe it's a little bit of a trap to price things according to their use by the best of the best players; it's a personal ideology thing, but I feel like there are things that excellent players should get better use out of than the average person. For example, Soontir, Vader, and Whisper would have gone up more under that system because they tend to be played very well at a higher level. Similarly, the Ups wouldn't have gotten as much of a break because top players are great at slow-rolling, action management, and range control. In the end, I don't think it's that big a deal and changing it would require manually re-entering 360 percentages, which I'm not terribly eager to do. This was intended as a fairly rough exercise anyway.

Thank you for replying.

1 Haha, nice - and who cares about you not being a "real" statistician, as long as it is clear what method you have used.

2. Thank you for doing this and I am sorry I did not have the time to count the different positive and negative adjustments. Those numbers seems more reasonable, but perhaps also more akin to what ships we see played?

You could add two colums in your spreadsheet one for each of the two types-adjustments, in addition to the sum the final adjustment, but again only a suggestion to satisfy my own curiosity, because I like your approach and want to understand the data.

3. Interesting, what you are telling me is that your have performed several sensitivity analysis, and arrived at very similar results - that is good, as that tells us that the data is sound.

ADDED UPGRADES!

https://docs.google.com/spreadsheets/d/1fZqj7rroGGcAPio285FJ7qOnluaUtixxs508FJq1pwQ/edit#gid=404016399

I know the real points are coming in 6 days but I have no patience. Plus mathematical modeling can be fun!

They're on the second sheet. This took a long time as the data was quite difficult to parse. Changes by as much as 100% or more are common at very low point values. I didn't want to mess with scaling values (and which ones do that shouldn't or don't that should, etc.), so I just took the starting values for Initiative 4, Medium Base, 2 Agility and projected flat from there. I figure if we're going to have a constant value for each, this should be the place to reference. Some things will be more or less efficient (notably shield and SR), but with the changes to the pilots it shouldn't be horribly abusive. Honestly a lot of the changes have worked out much more synergistically than I had expected.

Major changes I don't like this time:

• Leia only up to 4 (still much better)

• Ezra only down to 16 (sadness abounds)

• Advanced Sensors up to 12 points, and FCS and Advanced Optics get a 50% hike. Across the board Techs and Sensors are hurting (but hey, data)

• Juke up to 8, also notably Crack, Heroic, Marksmanship, Fanatical, Preadator, Outmaneuver all go up a point, Lone Wolf and Selfless up by 2

• No cannons went down, and HLC and Tractor went up a point each

• Saw, Maul, Phasma, and Nien go up one each

• Moldy Crow at 19

• BB-8 costs same as a generic BB astromech

Things I do like:

• 10 point Barrage, 8 point Prockets, 5 point ESC (so 8 are still legal), MUCH cheaper other missile options (2-4 points each!)

• Free Chopper and Vizago, 1 point Rebel Lando and Baze, 2 point Novice Technician, Scum Chewy, Hera, and ISB and Freelance Slicers make for some very interesting options

• More expensive Delta 7-B but cheaper CLT seems right on track. Also 1 point Grappling Struts

• Free R5-TK and Greedo (not sure what I think of free skilled bombardier), cheaper named gunners across the board, Luke up to 29 and 5th Brother to 11.

• Quiz down to 11 and Kallus, Cassian, Jabba, and Tarkin down to 4 (same as Leia or Tactical Officer)

• 1 point Instinctive Aim, 2 point Predictive Shot, 3 point Brilliant Evasion, Heightened Perception same, SR, Hate, and Battle Meditation all up a smidge.

• Illicits actually usable again (Cybernetics, RCC, and CD at 3, ID and Feedback Array at 2, DMS at 1)

But go check them out for yourself! I wish I could squad-build with my own points but it would be kind of cumbersome without YASB or an emulator.

On that note, any volunteers to make a YASB emulator so we can play around with fake points and exploit the obviously terrible loopholes until real points come in 6 days?

IG88-B gets less of a points reduction than.....

all the other ones?

It's definitely an interesting exercise, thank you for putting it together and sharing! Going through some of the stuff and realizing how it kicks me right in the bias was quite the feeling. On one hand, it looks like it would totally neuter some of my favorite meta lists right now, on the other some of the super fun jank that wouldn't see play under "normal" points predictions probably would.

For example, last night I played Mace with Sense, 7b and R2 along with Saesee and Lumi with 7B and R2 (200 on the dot). That list goes up 13 points to 213, meaning I'd have to lose a 7B (which in my mind would make the list unplayable) or 2 R2's and Sense (also likely unplayable). This makes me sad, because 3 I4 jedi don't seem overpowered and it was fun.

On the other hand, a random jank list that looks super fun would have dropped by 15 points, giving me all the toys I wanted plus room for a few additional ones.

You'd have to figure that with an exercise like this, you are essentially looking to drop the power level of the average list played. This means that random fun jank list actually might be able to compete, which is neat.

With these prices you could get:

• 5 TIE Aggressors with Ruthless (because it's free), Homing missiles, Cluster Missiles, and an Ion Cannon Turret each.

• 5 A-Wings with Daredevil and Intimidation and both Ion and Homing missiles (edit: forgot they only have 1 slot; make 2 Arvel and Jake instead of the ions)

• Marauder Kath with VTG escorted by 4 Binayre pirates, all equipped with Feedback Array. Also room to swap one for Yushyn to remove the disarms if you prefer.

• 6 Blue Squadron Torrents with Dedicated, Cluster Missiles, and Synchronized Console

• 8 Trade Federation Drones with Grappling Struts and Discord Missiles (Or any number with ESC instead of both)

• 5 generic inquisitors with Cluster Missiles and Instinctive Aim (Repo + Focus then launch missiles without worrying about lock? Sounds nice).

Probably a lot more interesting stuff with named pilots but it's tricky to list build from just a list.

I've really enjoyed this and I really hope when the Official app is better, all tournaments can have lists submitted via the app to a searchable official database (List juggler/fortress style) including outcomes and points for each match, maybe even with a player ID.

That would provide an insane amount of playerbase data from which an exercise like this might help bring balance to the game.

Too bad I don't see FFG, sorting out their app or setting up an official List/tournament fortress any time soon.

12 minutes ago, Scum4Life said:

Too bad I don't see FFG, sorting out their app or setting up an official List/tournament fortress any time soon. 

They have said they pay attention to user-created compilations though, including here on the forums but also things like ListFortress. Also note that ListFortress specifically does not violate any of their IP policy, our at least could easily be modified so not to violate it.

6 minutes ago, ClassicalMoser said:

They have said they pay attention to user-created compilations though, including here on the forums but also things like ListFortress. Also note that ListFortress specifically does not violate any of their IP policy, our at least could easily be modified so not to violate it.

Here's to hoping then

I sort of wish in the official squad building app custom feature ha given us the ability to tweak the point costs ourselves. Just mark increases as red and decreases as blue and a reset to put them back to their official point costs.

Fenn Rau 3p MORE?!?! 😤

7 minutes ago, Vector Strike said:

Fenn Rau 3p MORE?!?! 😤

Almost all I6s go up. I got the same response on Anakin.

The fact is that they’ve consistently done very well at the competitive level. Joy Rekkoff and Kad Solus can’t say the same.

Wedge and Vader get bigger bumps still and Boba actually doesn’t come down. Even rebel Han goes up significantly. Of I6s only Midnight, Dengar, and non-Rebel Hans come down. I5s have a similar trend, though Guri going down a point seems bizarre. AS Guri still comes out to a point more though.

19 minutes ago, ClassicalMoser said:

I5s have a similar trend, though Guri going down a point seems bizarre. AS Guri still comes out to a point more though.

I figure some of that is, for as popular as Guri is, and as strong as she can be when flown right, it's really easy to fly bad with her (I speak from experience ) . I'd guess that her winrate is lower than you'd expect for a ship this widespread.

14 hours ago, ClassicalMoser said:

With these prices you could get:

• 5 TIE Aggressors with Ruthless (because it's free), Homing missiles, Cluster Missiles, and an Ion Cannon Turret each.

• 5 A-Wings with Daredevil and Intimidation and both Ion and Homing missiles (edit: forgot they only have 1 slot; make 2 Arvel and Jake instead of the ions)

• Marauder Kath with VTG escorted by 4 Binayre pirates, all equipped with Feedback Array. Also room to swap one for Yushyn to remove the disarms if you prefer.

• 6 Blue Squadron Torrents with Dedicated, Cluster Missiles, and Synchronized Console

• 8 Trade Federation Drones with Grappling Struts and Discord Missiles (Or any number with ESC instead of both)

• 5 generic inquisitors with Cluster Missiles and Instinctive Aim (Repo + Focus then launch missiles without worrying about lock? Sounds nice).

Probably a lot more interesting stuff with named pilots but it's tricky to list build from just a list.

Discord Missiles are three-dot limited, I believe.

Other than that, none of them sound ridiculous. I can see missiles being cheaper - I think I agree with the original price increase of homing missiles but 4 seems more sensible than 5.

Thanks for this; it's a very interesting read.

I got curious about Special Forces Gunner last night, but it was too late to get into it then. The gunner TIE/sfs have always been a pretty interesting alternative to standard 3-red mainline jousty medium fighters. While naked Zetas and Omegas go down a point, Special Force Gunner goes up one. Net no change in the final product. And yet, many other equivalent ships (T-65, T-70, Belbullab, Kihraxz, Starviper---and even the Imdaar Phantom!) all see some of their generics go down in price (although B-Wings and ARCs see a slight increase in their cheapest generic).

So my supposition is that this increase is ALL Quickdraw. And I'm pretty confident I'm right.

QD: 68 squadrons, 29.0% average percentile. Backdraft: 5 squadrons, 14.3%. Omega: 4 squadrons, 6.55%. Zeta: 1 squadron, 7.3%. Meanwhile, the percentile for SF Gunner on BD, Omega, and Zeta is lower than the percentile for the ships themselves (22%, 12%, and 9% respectively), which probably indicates that they're *worse* with the gunner. Low sample size, but still.

I wouldn't call it a *flaw* in the model, to be sure. It's a different way of seeing. But again, what fascinates me about it is the various decisions it makes which aren't too humanlike. A human might see Guri and think "she needs to be more expensive" but the data finds something else. A quick look at percentiles: Guri herself: 24.3%, Advanced Sensors on her: 26.2%, Fire Control System: 21.0% (again low sample size). It's easy for a human to see how powerful Guri's upper limit is. Data doesn't back that up with a lot of wins. TIE/sfs are the other way. Data shows them to be fairly weak, except Quickdraw. The gunner--stapled to QD--goes up along with QD, but the generics get left behind.

So in total, this seems like an interesting tool. Does Guri need a nerf? Probably not, really. But the full build would still get one, with Shield Upgrade and Advanced Sensors going up, maybe also Afterburners (which isn't exactly stapled on). Low-upgrade Guri gets buffed, though. Meanwhile, you'd probably want a designer looking over the data to decide: Quickdraw rather than SF Gunner is the issue. Hold it steady, or maybe even reduce SF Gunner by 1, and add two more points to Quickdraw themself. The net result to QD builds--up from 55 to 57 with just the Gunner is the same, but allows a Gunner Zeta or Gunner Omega to drop 2 points, since they aren't well-performing ships.

I figure such frequently-paired upgrades are perhaps harder for a model to parse. Juke and Phantoms. Quickdraw and Special Forces Gunner. Aethersprites and Delta 7B. Y-Wings and Veteran Turret Gunner. The strength of the ships and the strength of closely-associated upgrades are certainly going to be strongly correlated. I don't know how much final build price adjustments wind up double-dipping because of it.

19 minutes ago, theBitterFig said:

So in total, this seems like an interesting tool. Does Guri need a nerf? Probably not, really. But the full build would still get one, with Shield Upgrade and Advanced Sensors going up, maybe also Afterburners (which isn't exactly stapled on). Low-upgrade Guri gets buffed, though. Meanwhile, you'd probably want a designer looking over the data to decide: Quickdraw rather than SF Gunner is the issue. Hold it steady, or maybe even reduce SF Gunner by 1, and add two more points to Quickdraw themself. The net result to QD builds--up from 55 to 57 with just the Gunner is the same, but allows a Gunner Zeta or Gunner Omega to drop 2 points, since they aren't well-performing ships.

I figure such frequently-paired upgrades are perhaps harder for a model to parse. Juke and Phantoms. Quickdraw and Special Forces Gunner. Aethersprites and Delta 7B. Y-Wings and Veteran Turret Gunner. The strength of the ships and the strength of closely-associated upgrades are certainly going to be strongly correlated. I don't know how much final build price adjustments wind up double-dipping because of it.

Yeah, an interesting tool (maybe as a more objective starting point) is all it's really intended to be.

And yes, the frequently-paired items is tricky and a known issue with the system. The ones you mention (especially Anakin and 7B), plus combos like Howl + Iden or Howl + generics tend to balance out the point change rather than putting it all on the source of the list's power level (I think Howl should go up quite a bit and Iden and the generics should come down by a comparable amount). I was surprised that SR got an increase, but I think that's mainly due to its use on I3 Inquisitors. 17 actually seems not horrible if we do need a flat cost.

On the other hand, the generic Y-Wings didn't go up at all because really the abuse is just as prevalent on Loks (which also saw only a very minor increase). VTG at 9 for the double-tap might be an overreaction but is probably closer to right than 6. If the devs want it to work on bow ties, they have to price it differently for those. At any rate, overall increase of Juke Sigmas going to 56 is reasonably in-line with some of the devs' more potent nerfs. Maybe there's something to that?

1 hour ago, ClassicalMoser said:

Yeah, an interesting tool (maybe as a more objective starting point) is all it's really intended to be.

And yes, the frequently-paired items is tricky and a known issue with the system. The ones you mention (especially Anakin and 7B), plus combos like Howl + Iden or Howl + generics tend to balance out the point change rather than putting it all on the source of the list's power level (I think Howl should go up quite a bit and Iden and the generics should come down by a comparable amount). I was surprised that SR got an increase, but I think that's mainly due to its use on I3 Inquisitors. 17 actually seems not horrible if we do need a flat cost.

On the other hand, the generic Y-Wings didn't go up at all because really the abuse is just as prevalent on Loks (which also saw only a very minor increase). VTG at 9 for the double-tap might be an overreaction but is probably closer to right than 6. If the devs want it to work on bow ties, they have to price it differently for those. At any rate, overall increase of Juke Sigmas going to 56 is reasonably in-line with some of the devs' more potent nerfs. Maybe there's something to that?

Grey Bomber in Rebels did go up 1 point. Rebel Gold stayed pat, but Scum generic Y-Wings went down. But like, this makes perfect sense! If Rebels want a Y-Wing spam list, or if anyone wants it in Hyperspace, this is how it gets accomplished, with Grey Bombers. Meanwhile, Scum Y-Wings are just a worse option than Scurrgs--within the context of Scum, which is necessarily the extended context.

The five most popular Y-Wings (Drea, Dutch, Grey, Horton, Norra) all went up, #6 in Gold holds, everything else dropped.

I did a "biggest winners and losers" sort of thing comparing my model of the costings with the actual new points.

Long story short I think Han and Luke are very good pilots and some of the best upgrades are Kanan , Afterburners, and gunner Luke. Ezra gunner got a bigger break than I gave him and AS, LW, Barrage, and Sense didn't get the nerfs my model predicted. Pretty much ALL the Jedi are undercosted now, and the clones and droids are right on the mark or 1 under.

Dash and Kanan are very sad, as are most of the VCX and YT2400 pilots. I can't fathom how on earth the devs expect them to be viable. It makes absolutely no sense. Dash might be good at 90 points but probably still wouldn't take his title. Leebo didn't get his crew slot back. There's literally nothing going for them. Zero. Not in firepower, not in survivability or support or neat tricks. The VCX might see some use (but not Kanan), but the YT-2400 just got benched for another 6 months. How on earth did the devs not see that it got used in literally THREE OP lists this entire season! And Leebo and the Wild Space Fringer in ZERO! Oh, and Dash got a mean percentile of 17, so it's not like even his most hardcore fans are able to do well with him. Ezra Bridger pilots are also pretty sad, but the Sense/Cassian combo is interesting (read: broken) and might save them. That remains to be seen. With our luck they'll see that combo and up poor Ezra for it.

Scum is also still very sad. The aces didn't get nerfed that badly, but the large bases and crew/gunners didn't get buffed enough at all. The Lancer and Firespray look to be in a pretty good place and Teroch didn't get nerfed at all. Illicits staying mostly the same is definitely the worst thing, though Contraband Cybernetics will be a very decent and not-broken go-to (I hate go-tos but it was probably the only good one anyway).

But the main thing is that EVERYONE loses because the single-slot target-lock-based missiles STAYED THE SAME! WHY? They're all exactly 2 points overcosted. They're seen very little and when they are it's always a suboptimal use of points. FFG, FIX THIS.

And Guri went up instead of down, as your model said it should have been done.

21 hours ago, Vector Strike said:

And Guri went up instead of down, as your model said it should have been done.

My model makes almost everything that did happen a victory and almost everything that didn't happen a tragedy.

Got bored so I did it again. Now that S2 of 2019 is well under way and there's a significant amount of data from the early tournament scene, I wanted to see what's doing well and what's not, so I took the Meta Wing data and plugged it back into my model.

Due to the small sample size and out of desire to curb runaway predictions, I've tempered the final result by taking the average of my last prediction and this prediction. I expect to update the numbers sporadically as the season goes on.

If you're interested, you can find the new costs under the 3rd and 4th tabs here. Wave 5 items not included as they didn't exist yet and have no data anyway.

https://docs.google.com/spreadsheets/d/1fZqj7rroGGcAPio285FJ7qOnluaUtixxs508FJq1pwQ/edit?usp=sharing

TL;DR: I5s are all still too cheap and I4s too expensive. Generics generally have too much of a points divide between the two in almost all cases, and mid-initiative named pilots are all over the board. Cardinal and Dormitz need to lose 3 points while Tavson gains 2, and the YT-2400 still receives the maximum possible reduction in all cases.

Concussion missiles at 6 are probably fine, but clusters go down to 4 and homing and ion to 3. Ion cannon down to 4. DBM down to 5 and barrage up to 9.

Other interesting stuff as well.

4 hours ago, ClassicalMoser said:

barrage up to 8.

🤨 Barrage is 8 and has been since the July update.

30 minutes ago, Hiemfire said:

🤨 Barrage is 8 and has been since the July update.

Meant 9

1 minute ago, ClassicalMoser said:

Meant 9

K.

21 hours ago, ClassicalMoser said:

TL;DR: I5s are all still too cheap and I4s too expensive. Generics generally have too much of a points divide between the two in almost all cases, and mid-initiative named pilots are all over the board. Cardinal and Dormitz need to lose 3 points while Tavson gains 2, and the YT-2400 still receives the maximum possible reduction in all cases.

Concussion missiles at 6 are probably fine, but clusters go down to 4 and homing and ion to 3. Ion cannon down to 4. DBM down to 5 and barrage up to 9.

Other interesting stuff as well.

Yup, most of that sounds pretty much bang on. Especially 5's and 4's being mostly under and overpriced respectively.

Edit for evidence, here is the trend for initiative of the top 50 pilots from meta-wing so far.

After nerfs to Rebel Beef and Phantoms, Init 4 is being squeezed out. I think I1 and I6 are just pulling from a smaller pool overall.