Awesome job!!! Do you have this written out in a pdf, would be great to be able to read it in that format!!! if it is here somewhere in the thread I must have missed it.

Negative, there is no PDF. Keep your eyes peeled for MathWing 3.0 though. (Disclaimer: there is no ETA). General updates in v3.0 include:

• charts comparing all the pilots' absolute jousting values vs pilot skill
• per-wave meta breakdown
• better modeling of many pilot abilities
• stat line dependent action economy
• effects of autothrusters and TLT incorporated
• regeneration
• more fundamental research and behind the scenes updates to make all of the above (and more) work.

The MathWing 2.0 process in this thread is completely "open source" in that the methodology is 100% documented, and the numbers can be exactly recreated. For IP reasons, I will not be revealing the specific inner workings of MathWing 3.0, although I will still keep the 2.0 methodology archived here.

Looking forward to the Kihraxz Fighter. No rush, just wanted to check it and realized it isn't out yet...

I had some numbers in an earlier post from Dec 14... I'll just re-post and add a couple more *cough* TIE Defender *cough* that should raise some eyebrows. Occasional color commentary included.

These are absolute jousting efficiencies from in-progress MathWing 3.0. PS greater than 1 (when applicable) is not factored in.

Disclaimer: these are all still using legacy meta assumptions. They will change once I model a wave 7/8 TLT-heavy environment. (i.e. Fel is way better than these numbers show)

TIE-Defender/x7: 102.0% would still be solid w/o white-K

Z-95: 100.0% (reference)

TIE Fighter: 99.3%

Vessery/x7 97.7% worth his points at PS1 w/o a white-K

X-wing + I.A -1pt: 95.6% (3/2/2/4 @ 21) should have been this

B-wing + FCS: 95.3%

Y-wing + TLT + BTL: 94.2%

X-wing + free hull: 94.0% (3/2/4/2 @ 21) my suggestion >1yr ago

TIE Adv + ACC: > 93.9% final number pending

A-wing + Refit: 93.1%

B-wing: 92.5%

X-wing + I.A: 92.3% (3/2/1/5 @ 22)

Whisper + VI/ACD/FCS 92.0% assuming shoots first

TIE-Defender/D + Ion 91.8% nothing to see here, move along...

X-wing + I.A: 91.3% (3/2/2/4 @ 22)

Vader + ATC: 90.4%

Kihraxz: 89.7% better than garbage but still DOA

X-wing: 86.3%

Y-wing + TLT: 85.7% assumes always has a shot

Whisper + VI/ACD 85.6% assuming shoots first

Fel + PtL + SD + AT: >85.5% pending wave 7 TLT meta analysis

Corran + FCS/R2-D2: 83.2% (R2-D2 = +3S) final number pending

E-wing: 78.8% less efficient than Corran, and only PS1

TIE Defender: 77.2% FFG valued 4K about same as a full turret

... mostly I look at the analysis and my eyes glaze over... so I'll trust you on the raw numbers. Give me a good command line interface and I'll figure it out... but formulas... [sigh]

Yup. Programming is easy. Implementing formulas that have already been laid out is slightly more difficult. But starting with an open-ended problem and deducing accurate formulas describing the system -- that's on an entirely different level.

Or as Han would say:

Well, that's the real trick, isn't it? And it's gonna cost you something extra.

I'm enjoying the more editorial comments along with the analysis....

A small word of caution:

If X-wing 3.0 includes better visualisation, that's awesome.

If it includes 'secret sauce' methodology, that's more troublesome. While I haven't ever found the time to properly re-implement the formulas you provide, I've always checked them, and in doing so confirmed that I'm happy with your numbers because I'm happy with your methods.

Secret sauce methodology that second-guesses other peoples math devolves rapidly into 'because I said so', and I'm not sure that's helpful.

Something I've never been clear on: these number are jousting numbers yes? So can't a ship have a lousy rating but still perform like a boss simply because jousting isn't the thing? Put another if two ships had the same cost and rating, but one could arc dodge and the other was a t65 xwing, wouldn't the ship that could arc dodge be objectively better as it could joust just as well and could dodge arcs too.

Something I've never been clear on: these number are jousting numbers yes? So can't a ship have a lousy rating but still perform like a boss simply because jousting isn't the thing? Put another if two ships had the same cost and rating, but one could arc dodge and the other was a t65 xwing, wouldn't the ship that could arc dodge be objectively better as it could joust just as well and could dodge arcs too.

Shh... some people build lists based on these numbers, don't spoil it for the rest of us that play against them!

The idea is that jousting is the most basic mode of gameplay and that a poor jousting efficiency needs to be made up in superior dials, action economy, pilot abilities and so on. If the ship special stuff like repositioning actions then it by definition doesn't have to rely on jousting. Jousting isn't the end all of the game but mathwing does show if a ship is effecient for the points or if it needs its special stuff to be good enough to pay for its points in a list.

Something I've never been clear on: these number are jousting numbers yes? So can't a ship have a lousy rating but still perform like a boss simply because jousting isn't the thing? Put another if two ships had the same cost and rating, but one could arc dodge and the other was a t65 xwing, wouldn't the ship that could arc dodge be objectively better as it could joust just as well and could dodge arcs too.

Look at it this way....What it gives you is an idea of how much the ship needs to make up outside of its statline to be effective. So, let's go with the Ewing at 78% as an example. In order for it to be worth its full points, the "other stuff" has to account for 22% of its cost. Some ships can manage that kind of difference with pilot abilities or upgrade choices, etc...In the case of the E, just Corran Horn with the right upgrade cocktail does it and that's why you see him as compared to the others.

Something I've never been clear on: these number are jousting numbers yes? So can't a ship have a lousy rating but still perform like a boss simply because jousting isn't the thing? Put another if two ships had the same cost and rating, but one could arc dodge and the other was a t65 xwing, wouldn't the ship that could arc dodge be objectively better as it could joust just as well and could dodge arcs too.

You are correct. This is why the T65 is still in a poor place, it's bare % behind the Bwing, but the Bwing can barrel roll AND has more upgrade options.

Something I've never been clear on: these number are jousting numbers yes? So can't a ship have a lousy rating but still perform like a boss simply because jousting isn't the thing? Put another if two ships had the same cost and rating, but one could arc dodge and the other was a t65 xwing, wouldn't the ship that could arc dodge be objectively better as it could joust just as well and could dodge arcs too.

There are tournament-capable ships that are lousy jousters, yes.

The T-65 has the unfortunate problem of only really being able to joust and not being the best at it.

Something I've never been clear on: these number are jousting numbers yes? So can't a ship have a lousy rating but still perform like a boss simply because jousting isn't the thing? Put another if two ships had the same cost and rating, but one could arc dodge and the other was a t65 xwing, wouldn't the ship that could arc dodge be objectively better as it could joust just as well and could dodge arcs too.

There are tournament-capable ships that are lousy jousters, yes.

The T-65 has the unfortunate problem of only really being able to joust and not being the best at it.

I've never noticed before that it is the absolute worst ship that is pretty much relying on just its numbers. Every ship below it has something...even if it's just a white K and BR.

Disclaimer: these are all still using legacy meta assumptions. They will change once I model a wave 7/8 TLT-heavy environment. (i.e. Fel is way better than these numbers show)

TIE-Defender/x7: 102.0% would still be solid w/o white-K

TIE Defender: 77.2% FFG valued 4K about same as a full turret

I've proxied it for a couple of games so I realised it was good (and, just as importantly, that it's fun). I didn't realise it was THAT good.

Has anyone run numbers on ordnance with the new options? Well, I guess Guidance Chip is really the only one that effects dice immediately. LRS just most likely allows you to have a Focus or Barrel Roll to have a shot.

I'd love to see the averages for Proton Torpedoes with GC and no Focus.

Hello Maj Juggler,

Can you explain why Soontir Fel's numbers will improve once you have added TLTs as opponents? Is it because he exploits their doughnut holes (which don't really exist with TLT spam)? My math (not as good as yours) says that the first TLT shot has about a 8% chance (assuming AT, stealth and focus) of doing some damage but once the stealth device is gone you are around 16% (if Soontir still has tokens) and then 40% with just AT. For me the problem is low hit points on soontir and high number of shots (that can often see soontir) from the TLTs.

Doesn't the undervaluing of Soontir show how important maneuver is in X-Wing and why (as you mention up front) you can't just consider these numbers in isolation, unless every one was a turret. As the jousting portion does not last for an entire game (as assumed by your lanchaster law (I think)) but rather just turns 2 to 3ish (unless you have a white k, baby!).

TLT (and other turret attacks) will trigger AutoThrusters for most shots. Since ~30% of the meta by points right now is TLT, ships with AutoThrusters should see a pretty significant relative bump in durability once I pull in the wave 7 meta assumptions.

Fel's numbers are for "turtle Fel" double focus + evade every turn, with the very occasional bump and no actions. Having the option to boost/barrel roll, and being PS9 are all bonuses above and beyond the raw absolute jousting efficiency.

A small word of caution:

If X-wing 3.0 includes better visualisation, that's awesome.

If it includes 'secret sauce' methodology, that's more troublesome. While I haven't ever found the time to properly re-implement the formulas you provide, I've always checked them, and in doing so confirmed that I'm happy with your numbers because I'm happy with your methods.

Secret sauce methodology that second-guesses other peoples math devolves rapidly into 'because I said so', and I'm not sure that's helpful.

I do the analysis primarily for myself, for a variety of reasons. A secondary effect is posting the results publicly. It's a free service that I offer to the community, and certainly not one that I am under any obligation to continue to keep providing. If the public finds it useful, that's fantastic, but the intent is not to try and "convince" anyone "because I said so". You can take it, leave it, dismantle it, improve on it, or ignore it, it doesn't matter to me.

MathWing 1.0 that was posted in March 2014 is still several years ahead of FFG's current design and playtesting capability, judging by recent news articles. I'm assuming their intent, as Alex has stated in interviews, is always to release well balanced ships, neither duds nor overpowered. MathWing 2.0 in this thread applies to a wider range of pilots and was cleaned up quite a bit from early 2014, and is also public.

MathWing 3.0 is significantly more refined and involved, and could be useful for my own potential design endeavors, or professional consulting (for this game or others). If I publish the methodology / results, it will be in the academic literature, (possible but unlikely, I would have to see what journals are even applicable). In the meantime the secret sauce will remain secret. The alternative to publishing the 3.0 results without underlying methodology, is to simply not publish the results at all.

MathWing 4.0, if it ever happens, is a whole other thing and I'm unlikely to externally publish what it does, let alone its secret sauce.

Something I've never been clear on: these number are jousting numbers yes? So can't a ship have a lousy rating but still perform like a boss simply because jousting isn't the thing? Put another if two ships had the same cost and rating, but one could arc dodge and the other was a t65 xwing, wouldn't the ship that could arc dodge be objectively better as it could joust just as well and could dodge arcs too.

Correct. The directly above numbers (although not the numbers on the first page) also do not consider pilot skill or available positional actions.

Any word on the TIE Adv. Prototype yet?

MathWing 3.0 is significantly more refined and involved, and could be useful for my own potential design endeavors, or professional consulting (for this game or others). If I publish the methodology / results, it will be in the academic literature , (possible but unlikely, I would have to see what journals are even applicable). In the meantime the secret sauce will remain secret. The alternative to publishing the 3.0 results without underlying methodology, is to simply not publish the results at all.

Oh wow... never even thought about being able to turn higher level math function of a game into a Thesis! You rock! Just let us know when we need to call you DoctorJuggler instead of Major!

MathWing 3.0 is significantly more refined and involved, and could be useful for my own potential design endeavors, or professional consulting (for this game or others). If I publish the methodology / results, it will be in the academic literature , (possible but unlikely, I would have to see what journals are even applicable). In the meantime the secret sauce will remain secret. The alternative to publishing the 3.0 results without underlying methodology, is to simply not publish the results at all.

Oh wow... never even thought about being able to turn higher level math function of a game into a Thesis! You rock! Just let us know when we need to call you DoctorJuggler instead of Major!

I already received my PhD in Electrical Engineering last year. My research topic had nothing do to with MathWing.

There are many things that you can turn into a thesis or dissertation; the primary requirement of most Universities for a dissertation in a STEM field is a "novel contribution to the academic body of knowledge" and meets the corresponding level of academic rigor. You can get away with significantly less for a Master's level Thesis.

There's a great link here:

http://matt.might.net/articles/phd-school-in-pictures/

There are actually many references to uses and analysis of Lanchester's Equations, although applications to miniature wargaming is relatively limited.

So will you release the 3.0 results soon?

I'd love to see the averages for Proton Torpedoes with GC and no Focus.

I was thinking about this so I enumerated the results

Proton Torps, from an x-wing (or any 3 attack) with GC and no focus:

Average Hits: 1.31

Average Crits: 2.20

Average Total: 3.52

edit:.....

ok i think that's right. I had the worst case (4 blanks) case wrong, but only barely made a difference.

Odds of rolling total hits:

1: 0.39% (all blanks)

2: 6.6%

3: 34%

4: 59%

Pretty spiffy for only needing a TL

So, the Kihraxz fighter is crap? Just asking because I'm thinking getting a pair and fly them alongside an Aggressor.

So, the Kihraxz fighter is crap? Just asking because I'm thinking getting a pair and fly them alongside an Aggressor.

Talonbane's a good ace, if you give him mobility upgrades.

Should've been a point cheaper.

I get the desire for caution, of course, but it's about as solid as the x-wing - not bad for a 3-attack jouster, but those themselves are less shiny than they were now that TLTs are out there. 'Filler firepower' just isn't what it used to be.

So better stick to Y-Wings w/TLT, Unhinged (and BTL-A4)?

Nope, stick to Gold squadron TLT Y-wings.

3 basic, 1 BTL-4 stress-Y

2 points on PS upgrade or a droid.

Joust values are for jousting Turrets are for turreting

Nope, stick to Gold squadron TLT Y-wings.

3 basic, 1 BTL-4 stress-Y

2 points on PS upgrade or a droid.

Joust values are for jousting Turrets are for turreting

Not that easy as a Scum player.

Nope, stick to Gold squadron TLT Y-wings.

3 basic, 1 BTL-4 stress-Y

2 points on PS upgrade or a droid.

Joust values are for jousting Turrets are for turreting

Not that easy as a Scum player.

nothing is easy as S&V player in this cold miserable Rebel-biased world when your ships are still nostalgic about times 4 waves ago.

wave 7. everybody spams TLT, Pushes ace Limits...

S&V: Hey, guys, want to joust?

*sounds of horrible horrible death

wave 8

S&V: Hey, guys, see my new B-wing and PWT?

*sounds of horrible horrible death