I've seen this thread topic come up at least 10 times in the time I've been on this forum, and the consensus has always been "Well, you left out these variables".

So, I figured I'd start by listing all of the possible variables, and their interactions, and we'll figure out the numbers for them later.

Please let me know in the comments if I've missed anything

Attack Dice and Agility Dice

In my model, these are never going to simply be added to the score. They are not stand-alone stats, but rather multipliers.

As a proof of concept, Let's look at a Rookie vs a Bandit. They have the same action bar (no defensive actions), the same Agility value, and comparable Maneuver dials. The missile and the torpedo slots are roughly equal.

The key difference is that the X-Wing had 3/2 of the Z-95's attack dice, and some minor bonuses as well. Lo and behold, the Rookie costs a little more than 3/2 of the Bandit.

Similarly, that +1 Attack explains the similarly large shift between the Outer Rim Smuggler's cost and Chewbacca's.

Hull and Shields

This is going to be a flat statistic, though effected by the Agility dice multiplier. The real trick is going to be determining the difference in price of these two very similar stats. A Shield is crit-proof, and therefore worth more than a Hull, but only 8/33 of the Crits will be double damage (well, 7 are double, and 2 are a 50% of also being double), and Crits aren't that likely to be rolled, so it's going to be a minor issue.

Fortunately, solving this will also end the Shield/Hull Upgrade wars. If Shields are better than 1.33 of a Hull, the Shield Upgrade is still worth taking. If it is less than 1.33, that 1 extra point is getting spent poorly.

Actions

Target Lock is improved by having more Attack Dice, having an Ordnance slot, having Focus, having a High PS, and having a poor dial.
It competes with Boost/Barrel-Roll for the "Action to use when you don't have a shot", and with Focus for offense.

Focus is improved by having more Attack Dice, having more Agility dice, having a Crew slot, and having good mobility.
It competes with all other actions for offensive and defensive purposes.

Evade is improved by having more HP and having fewer Agility dice.
It competes with Focus, Cloak, Boost, and Barrel-Roll for the title of best defensive option.

Boost and Barrel-Roll are improved by having more Attack Dice, more Agility dice, having a Crew slot, having a high PS, and having a poor dial.

They compete with all other actions for offensive and defensive purposes

Cloak is.... hard to talk about.

Upgrade Slots

In my understanding, most of the points for filling an Upgrade Slot is left to the upgrade itself, but a portion of the upgrade's worth is left in the ship itself to fill the slot. This is how the Shuttle wound up costing as much as the X-Wing, despite losing hard in a joust when neither have upgrades.

Modifications allow for the addition of extra HP, the Boost action, the Target-Lock action, extra Agility Dice, and improvements to the Torpedo/Missile slots and the Cloak action.

Torpedos and Missiles require a Target-Lock action, and are improved by having fewer attack dice, and compete with Cannons.

Bombs are improved by having a High PS and good mobility.

Cannons and Turrets are improved by having low attack dice and high survivability.

Crew are improved by everything. The more options you have on your naked ship, the more valuable Crew can be.

Astromechs are improved by having high Hull, high Shields, and lots of non-green 1 or 2 maneuvers.

Systems are improved by having good maneuverability and many action options.

Elite Pilot Talents are improved by everything.

Maneuvers

This one is going to be tricky, as we'll have to establish a combinatory weighting system for mathematically quantifying this.
The variety of Turns and Banks are more important than the variety of Koiograns (past the first), which is more important than the variety of Forwards.
The difference between Green and White isn't as important as the difference between White and Red, which isn't as important as the difference between Red and Non-Entity.

On a previous thread, I showed that the X-Wing and Firespray-31 are the median dials, having the most common numbers of bearings (2 turns left and right, 3 banks left and right, 4 forward options, and a 50/50 on 1 or 2 koiograns) in their most common distances (2-3, 1-3, 1-4, and 4/3-4, respectively) in their most common colors.

Pilot Skill

Having a higher or lower Pilot Skill only matters when comparing two ships.

Against an Academy Pilot, Wedge Antilles might as well have the same PS as a Rookie.

As such, the number of the Pilot Skill isn't as important as the odds of placing it in its proper role. High PS improves attack dice, and Low PS improves Agility dice.
Low PS improves Maneuverability, and High PS improves Boost/Barrel-Roll actions.

Titles and other specific ship quirks

The Falcon and the Firesprays both improve their attack dice substantially with an increased arc. They also do odd things to the importance of dials.

Similarly, being able to double-dip in the EPT pool and removing 2 points for the cost of the missile slot may do a LOT for the A-Wing.

Specific Pilot Abilities

This will be determined on a case-by-case basis. Unfortunately, there's a lot of this to slog through, so picking out exact values and variables will be difficult.

I think your X-Wing Z-95 logic and Lambda X-Wing Logic are off.

The Lambda has twice the health an X-wing does. It loses in jousting thanks to a bad dial. It's point cost is affected by that, and points more to Large bases being a cost reduction. Also, the Lambda can take an X-wing if it dishes out enough hurt before the X-Wing can K-Turn.

The X-Wing Z-95 logic of 3/2 the cost for 3/2 the firepower doesn' count the X-wings dial being a smidgen better(Longer K-Turn, and the extra Hull, which means a lot.

Beyond that The Interceptor costs 1.5 times as much as a tie, and gains a better dial, boost, and an extra attack point, highly suggesting it is not a multiplier in the way you're thinking.

But all in all, you might be right about it being a multiplier. Just not as straightforward as you're saying.

Your crit weighting is incorrect. Minor Explosions are a 37.5% of an additional damage, not 50%. Rolling a [crit] result does not produce an extra damage.

Do not forget base size.

Thank you for this, I'm not sure everything lines up but it's a very interesting read. Please continue.

And the design team laughs another laugh of knowing, we yet again, prove we know nothing... because they don't either!

or that the "barge" in the San Francisco harbor is actually a highly complex computer who's sole job is coming up with appropriate point costs for xwing ships using 198,789,467 variables.

So... You're going to take 20+ input variables and you're going to try and figure out a function that maps these 20 inputs to an output, using 80+ data points. You do not know what form this function has so, you'll make a bunch of assumptions, get a certain form for your function, and, then, you'll get a bunch of constants to populate your function. Finally, you'll compare the output your function gives with your 80+ data points, realize it's close, but not spot on. Meh... chalk it up to rounding errors and FFG fudging the values.

This is gonna go well.

There likely is no formula. FFG costs things by playtesting.

So... You're going to take 20+ input variables and you're going to try and figure out a function that maps these 20 inputs to an output, using 80+ data points. You do not know what form this function has so, you'll make a bunch of assumptions, get a certain form for your function, and, then, you'll get a bunch of constants to populate your function. Finally, you'll compare the output your function gives with your 80+ data points, realize it's close, but not spot on. Meh... chalk it up to rounding errors and FFG fudging the values.

This is gonna go well.

Not rounding errors. I am not expecting any ship to have exactly an integer as its answer. If I get a truly accurate algorithm, it should shift as the metagame shifts. Rounding of fractions of a point is not an error, but it is telling about the quality of the ship in question.

There likely is no formula. FFG costs things by playtesting.

This is actuality how they do it. However, we can discover trends here.

Academy TIEs cost 12 instead of 11 because 9 TIE fighters is too much, but 8 is somehow fine.

Rookies cost 21 instead of 20 because 5 X-Wings is too much, but 4 is fine.

A-Wings were costed at 17 because 6 was thought to be too many, but with the Chaardan Refit, they're showing that they've changed their minds on that.

THAT is the math they're doing with playtesting: What fraction of 100 points is this ship worth? FFG is using intuition to solve this problem, aye, but they're not just tossing numbers in the air and catching them as they fall.

A Delta Pilot costs 30 points. This is because fielding 3 of them with Ion Cannons is apparently okay, but fielding 3 of them with Heavy Laser Cannons is not. Either way, you cannot do 4 of them. The cost of the upgrade slot is real.

3 points worth of Upgrade on each is fine in a BBBB list, but not 4.

And so on.

Academy TIEs cost 12 instead of 11 because 9 TIE fighters is too much, but 8 is somehow fine.

Actually, I prefer to look at it from the reverse: 12 points IS the base cost of a TIE fighter, but you get an extra PS at the low end for one point than you would with more expensive ships. If PS3+ TIE Fighters all costed 1 more point, they would be too expensive (except probably Howlrunner) and you would rarely see most of them played. Spending 8.3% more on your squad to go from PS1 to PS3 is about equivalent to spending 9.5% more on your squad to go from PS2 to PS4 on an X-wing. But there is no way you would want to spend 16.7% more to go from PS1 to PS3.

I don't think the Linear Algebra approach here is really that useful in predicting a non-existent formula for predicting future ship costs, but since you're going down this route I would stick with comparing the PS1 equivalent costs for all of the ships. Just look at the base ship costs.

I think your X-Wing Z-95 logic and Lambda X-Wing Logic are off.

The Lambda has twice the health an X-wing does. It loses in jousting thanks to a bad dial. It's point cost is affected by that, and points more to Large bases being a cost reduction. Also, the Lambda can take an X-wing if it dishes out enough hurt before the X-Wing can K-Turn.

The X-Wing Z-95 logic of 3/2 the cost for 3/2 the firepower doesn' count the X-wings dial being a smidgen better(Longer K-Turn, and the extra Hull, which means a lot.

Beyond that The Interceptor costs 1.5 times as much as a tie, and gains a better dial, boost, and an extra attack point, highly suggesting it is not a multiplier in the way you're thinking.

But all in all, you might be right about it being a multiplier. Just not as straightforward as you're saying.

If it were literrally just the xwing and the lambda, and the xwing pilot knew what he was doing, it would be difficult for the lambda to get more than one shot off (and if there was only one shot, it wouldn't likely be range 1 but maybe) before the xwing could kturn.

If you take away 3 points from an xwing for the hull point it has on a z95, then it's 18 to 12. Also, the xwing dial is not better, there are just as many times with a kturn that you will want to stay closer to a fight and use a 3. Same goes for 1/2 banks having green/white interchanged. You could MAYBE argue that a "little #$%^" like a headhunter might want to run as far away as possible, but the dials themselves are equal.

but the [X/Z] dials themselves are equal.

For ***** and giggles I've been working on my own formulas, mainly for maneuver dials since they seem to be worked on the least. I have no idea what a final scale/range should look like, and I know I'm still way off on the whole since named YT's are currently sitting at the bottom of the efficiency ladder...

BUT in my formula's current iteration, both the X and Z have a maneuver rating of 13.5 (which currently has ships as high as 19.5 and as low as 7), so I've got that going for me which is nice.

I think your X-Wing Z-95 logic and Lambda X-Wing Logic are off.

The Lambda has twice the health an X-wing does. It loses in jousting thanks to a bad dial. It's point cost is affected by that, and points more to Large bases being a cost reduction. Also, the Lambda can take an X-wing if it dishes out enough hurt before the X-Wing can K-Turn.

The X-Wing Z-95 logic of 3/2 the cost for 3/2 the firepower doesn' count the X-wings dial being a smidgen better(Longer K-Turn, and the extra Hull, which means a lot.

Beyond that The Interceptor costs 1.5 times as much as a tie, and gains a better dial, boost, and an extra attack point, highly suggesting it is not a multiplier in the way you're thinking.

But all in all, you might be right about it being a multiplier. Just not as straightforward as you're saying.

If it were literrally just the xwing and the lambda, and the xwing pilot knew what he was doing, it would be difficult for the lambda to get more than one shot off (and if there was only one shot, it wouldn't likely be range 1 but maybe) before the xwing could kturn.

If you take away 3 points from an xwing for the hull point it has on a z95, then it's 18 to 12. Also, the xwing dial is not better, there are just as many times with a kturn that you will want to stay closer to a fight and use a 3. Same goes for 1/2 banks having green/white interchanged. You could MAYBE argue that a "little #$%^" like a headhunter might want to run as far away as possible, but the dials themselves are equal.

And a 3K-turn is more likely to be blocked than a 4(Specifically because of large ships for the most part. The risk of collision on a K-turn makes the longer K-turn better over the course of time. Hands down. The X-wing has a marginally better dial, as I said.

Individual Hull Points are not 3 and shields are not 4. The Upgrades are purposefully overcosted compared to a ships natural numbers.

If they were the Z-95 would cost 14+points. The Lambda would cost 35+, etc.

but the [X/Z] dials themselves are equal.

For ***** and giggles I've been working on my own formulas, mainly for maneuver dials since they seem to be worked on the least. I have no idea what a final scale/range should look like, and I know I'm still way off on the whole since named YT's are currently sitting at the bottom of the efficiency ladder...

BUT in my formula's current iteration, both the X and Z have a maneuver rating of 13.5 (which currently has ships as high as 19.5 and as low as 7), so I've got that going for me which is nice.

Feel free to look up my old threads on that.

By the way, you might be interested to know that the Defender's the first ship with all 12 bank/turn bearings, and that the E-Wing's the first with all 5 forward maneuvers.

To start the reverse engineering, I think you should first try to find a formula using just the Wave 1 ships. That is the group that started it all. You can than later test it against newer ships to see if it holds up. That might be a lot easier than trying to figure out a formula using all ships, including Defenders and Phantoms.

You might even go a long way just using the core set.

And, assuming there is a formula...

There are other possibilities too, maybe agility counts for nothing by itself, but ramps the cost of hull and shields. Eg. Hull = 3 on agi 2 but 2 on agi 1 and 5 on agi 3....

There are many nonlinear permutations possible. Formulae put out have gotten more and more complicated, but I can't believe a formula would be so "complex"

Our regression attempts worked pretty well with few ships, but as more and more come out - out models make less and less sense. Earlier posts suggest balancing through testing, that would help deal with intangible benefits, and explain lack of a simple model. Or such balancing would obfuscate a simpler underlying model.