So I guess I'm going to be the one to revive this post....

Because reasons

I'm trying to work out the system they use to cost ships out. It's now wave 5, there's plenty of ships and squads to gather data from. With all the brilliant minds I've seen interacting on here there's no reason it can't be figured out. I saw a few of the older posts using damage % and spread sheets, but the numbers just seemed too far off to truly be accurate.

That said, here's what I came up with.

Instead of using a more advanced formula, I figured points may be more literal, as in, a sum of all the positive and negative qualities. I took the CR90B and added all the features together to see if I could reach 39 points.

4hull+1antisq+1com+1SQ+2eng+3DFtokens+8bluedice+7shields+4speed+12yaw+3upslots-2blankyaw

4+1+1+1+2+3+8+7+4+12+3-2=44

The number was so close I figured maybe I just didn't equate the negative values correctly. Maybe blank yaws were worth -2 instead of only -1

4+1+1+1+2+3+8+7+4+12+3-4=42

I also then considered that maybe upgrade slots had no value.

4+1+1+1+2+3+8+7+4+12-4=39

This got me to 39, while I'm aware that didn't make it necessarily correct, it at this point seemed accurate...

So I looked at the CR90A. If you consider every red die an upgrade at +1 from blue, then it adds +5 to the cost total which matches.




Satisfied that I may have had a fairly accurate base I moved to an Imperial ship, the Glad I to test if it was still accurate

I came up with

5hull+1antisq+2com+2SQ+3eng+3DFtokens+14bluedice+14diceupgrades+8shields+3speed+6yaw-2blankyaw

5+1+2+2+3+3+14+14+8+3+6-2=59

It was very close so I looked at the differences between this and the corvette to see if I could spot a reason for the point difference.

The only thing I could think was that base cost of black and red might be calculated differently than an upgrade from blue, so I took the 28 points of dice and upgrades out and reentered values to give black dice a value of +2 and red a value of +1 this brought the total to 25 instead of 28 as shown here

5+1+2+2+3+3+22+3+8+3+6-2=56

This puts the values spot on, so I checked it against the Glad II

At which point I figured, an upgrade from blue to red dice had previously been calculated as +1 so it would make sense that the upgrade from black to red dice would count as +2

The only puzzling aspect seems to be the anti squad upgrade of 1 blue die costing an extra 2 points, maybe when dice are added to anti squad they add at black for 1 point and then it was upgraded to blue for 1 point?

If accounted this way it seems to add up




It was at this point I figured I'd try my luck against the mighty Imperial I Class Star Destroyer, see if this formula could possibly still add up.

11hull+3antisq+3com+4SQ+4eng+4DFtokens+4bluedice+14blackdice+8Reddice+12shields+3speed+5yaw-2blankyaw

11+3+3+4+4+4+4+14+8+12+3+5-2=73

37 points short. which didn't make much sense to me.

However, if you take the large base into account, maybe that's a factor. Maybe they use that as a point buffer? So i checked this card for any possible points I missed, as 37 seemed an odd point buffer, and it occurred to me that it was the first card I'd tested to feature two of the same upgrade slot? Maybe that was worth a point or two? At this point I'm willing to venture 2 (we'll test it against others later) Making a large base count for a solid 35 points

35+11+3+3+4+4+4+4+14+8+12+3+5+2-2=110

So Again, I wanted to check my system against the "upgrade" ship the ISD II

7 dice get upgraded from black to blue at +1 and one from black to red at +2 then one last point to upgrade the extra offensive retrofit into a defensive retro fit upgrade for an even 10 points between the 2



Since the Imperials only have 1 large base ship, there's really no way to test that buffer value against another imp ship, so I tested against the mc80s

Liberty:
35buffer+8hull+2antisq+3com+2SQ+4eng+3DFtokens+8bluedice+7Reddice+11shields+3speed+4yaw+2 Doubleslot-4blankyaw

35+8+2+3+2+4+3+8+7+11+3+4+2-4=88 off by 8

Homeone:
35buffer+8hull+3antisq+3com+4SQ+4eng+4DFtokens+10bluedice+8Reddice+15shields+2speed+3yaw

35+8+3+3+4+4+4+10+8+15+2+3= off by 7





This is all I've really had time to work on, anyone have any ideas? corrections?

First, there simply isn't a magic formula. It simply cannot work, because ships are far more than the sum of their parts. There are positive and negative synergies between stats (e.g. black dice are almost worthless to a slow, poor-maneuvering ship - and defensive retrofit slots are much more valuable for big expensive ships), as well as things that are just impractical to account for in a formula (for example, the width of the arcs, or the friendliness of the maneuver chart). You can of course come up with decent approximations, but they will always need manual adjustment and balancing.

Having said that, yeah, of course someone's done this properly before: https://drive.google.com/open?id=0B62s92eVIoaPZk05RFlLNGdpT1U

6 minutes ago, DiabloAzul said:

First, there simply isn't a magic formula. It simply cannot work, because ships are far more than the sum of their parts. There are positive and negative synergies between stats (e.g. black dice are almost worthless to a slow, poor-maneuvering ship - and defensive retrofit slots are much more valuable for big expensive ships), as well as things that are just impractical to account for in a formula (for example, the width of the arcs, or the friendliness of the maneuver chart). You can of course come up with decent approximations, but they will always need manual adjustment and balancing.

Having said that, yeah, of course someone's done this properly before: https://drive.google.com/open?id=0B62s92eVIoaPZk05RFlLNGdpT1U

Only about 1000000x better than what I came up with. Thanks man.

Aeons ago when the customs community here really existed, there were a pair of formulas for ships that were used, with our rule for new costs being "Run both; take the higher unless it's ridiculous; probably add a custom tax; playtest and adjust, when in doubt up". I've never seen Diablo's Google form before but it looks right. If you want I can post the links to the BGG forum topic where they were, or the MATLAB code I wrote them into so I didn't have to to do math, just enter figures. @DiabloAzul is 100% correct though, the formulas don't account for vagaries like dice synergies (both upgrades and maneuver), arc setups, or anything else that can't be numerically deduced. Now while it's perfectly possible to develop a formula including those vagaries into its calculations, that is a spline curve no human mind can compute, and even it will get things wrong. When I first developed the Sorannan class Star Destroyer, the formulas spat out figures thirty points to either side of the cost I gave it, because it's an edge case ship. By my figuring we'll have enough data points for reasonable estimates of that equation set around Wave 8-10, but it will never be perfect, so take any formula's answer with a grain of salt, even if it can perfectly guess every FFG release you can correct it to.

1 minute ago, GiledPallaeon said:

Aeons ago when the customs community here really existed, there were a pair of formulas for ships that were used, with our rule for new costs being "Run both; take the higher unless it's ridiculous; probably add a custom tax; playtest and adjust, when in doubt up". I've never seen Diablo's Google form before but it looks right. If you want I can post the links to the BGG forum topic where they were, or the MATLAB code I wrote them into so I didn't have to to do math, just enter figures. @DiabloAzul is 100% correct though, the formulas don't account for vagaries like dice synergies (both upgrades and maneuver), arc setups, or anything else that can't be numerically deduced. Now while it's perfectly possible to develop a formula including those vagaries into its calculations, that is a spline curve no human mind can compute, and even it will get things wrong. When I first developed the Sorannan class Star Destroyer, the formulas spat out figures thirty points to either side of the cost I gave it, because it's an edge case ship. By my figuring we'll have enough data points for reasonable estimates of that equation set around Wave 8-10, it will never be perfect, so take any formula's answer with a grain of salt, even if it can perfectly guess every FFG release you can correct it to.

Yeah I'd appreciate any relevant links.

Gives me a chance to read up

2 minutes ago, Darth Sanguis said:

Yeah I'd appreciate any relevant links.

Gives me a chance to read up

*Wields Ancient Art of FFG Forum Necromancy

https://boardgamegeek.com/thread/1371685/how-cost-customs-and-variants-point-formulas-and-o

I can give you the MATLAB code too if you want, but I never got my own version fully online, so it's just an automated version of the formulas listed above. MATLAB also has a fairly odd syntax, so the raw code probably won't help you a whole lot either, though it's yours if you want it.

Just now, GiledPallaeon said:

*Wields Ancient Art of FFG Forum Necromancy

https://boardgamegeek.com/thread/1371685/how-cost-customs-and-variants-point-formulas-and-o

I can give you the MATLAB code too if you want, but I never got my own version fully online, so it's just an automated version of the formulas listed above. MATLAB also has a fairly odd syntax, so the raw code probably won't help you a whole lot either, though it's yours if you want it.

Nah, I'll make due with what's here. It looks like more people have put considerably more effort and thought into this than I did myself. Looks like there's plenty here already. Thanks though, you guys are always so helpful!

I have to admit I posted the earlier link without double-checking... it's not the same spreadsheet I used in the past for the Armada Shipyards customs. Back then I used this one: https://docs.google.com/spreadsheets/d/1Ucyc_ZVSri89giVoLNj72ER36rZg85lOnMpUO5L32mA/edit#gid=38434546

10 minutes ago, GiledPallaeon said:

the MATLAB code I wrote them

Seems to be facts fitting theories instead of theories fitting facts

Just now, Tirion said:

Seems to be facts fitting theories instead of theories fitting facts

Adjusting theories until I could establish facts... which, with as many variables as there were, probably wasn't going to happen.

OH MY GOD. THE CRABBY PATTY IS PEOPLE... PEOPLE!!!!

Just now, Drasnighta said:

OH MY GOD. THE CRABBY PATTY IS PEOPLE... PEOPLE!!!!

I was just about to make this joke lol

What? Soylent green not good enough for you?

1 minute ago, idiewell said:

What? Soylent green not good enough for you?

That depends, is it actually made by harvesting things from the Sea Floor, or not?

37 minutes ago, Drasnighta said:

That depends, is it actually made by harvesting things from the Sea Floor, or not?

Never been curious as to what human tastes like? Cannibalism is quite natural.

5 minutes ago, Undeadguy said:

Never been curious as to what human tastes like? Cannibalism is quite natural.

Why do you assume I'm still curious?

1 hour ago, Drasnighta said:

That depends, is it actually made by harvesting things from the Sea Floor, or not?

Oh it's actually made by "harvesting" things all right...

You know what the Single Most disappointing thing in Food marketing right now, is?




...

Available in:

Powder.
Drink.
Food Bar.


NOT GODDAMN CHIPS.

SO FREAKING DISAPPOINTING.

4 minutes ago, Drasnighta said:

You know what the Single Most disappointing thing in Food marketing right now, is?




...

Available in:

Powder.
Drink.
Food Bar.


NOT GODDAMN CHIPS.

SO FREAKING DISAPPOINTING.

"Plant based"


Why would I want to eat my food's food?

There used to be a company that sold human-flavored tofu called "HuFu".

1 minute ago, Democratus said:

There used to be a company that sold human-flavored tofu called "HuFu".

I'm genuinely surprised that's not still around... seems like there'd be a huge market for it.... unless it tasted crummy...

1 hour ago, Darth Sanguis said:

I'm genuinely surprised that's not still around... seems like there'd be a huge market for it.... unless it tasted crummy...

I'm going to guess it did NOT taste like chicken...

9 minutes ago, idiewell said:

I'm going to guess it did NOT taste like chicken...

Almost universally its described as being more Pork-like.

Hence the very term, Long-Pork.

Which gave us "Long" as a Title.

Didn't it...

Didn't it, Mr. Silver?