How to make your own KoW/Warpath army list!

[darkPrince010]

As suggested by scarletsquig

In 10 easy steps!

1) Firstoff, have a base idea of what you want. Kind of a no-brainer, but it really can help you figure out what might be a better choice for filling certain unit slots.

2) Come up with unit names and what sizes you want them to be available in. Keep in mind you MUST have at least 1 unit that can be at least a Regiment/Section, as H/M, Armor, War Machines, and Aircraft require one to be used. Honestly, the more the merrier here as figuring out the point cost of a Section/Regiment is very easy if you already know the value for a Troop/Team

3) Figure out an appropriate statline (Sans Special Rules). Keep in mind that Flying/Jump Troops automatically changes Speed to 10 in 1st edition. Don't forget also that there are no official army lists that have a 2+ to hit (By to-hit I mean Melee/Ranged for KoW and Hit for WP), nothing in KoW that has a 7+ or better Def (At least as far as I know...), and at most you should only have around 2 shots/melee attacks per figure in the unit (ie a Team shouldn't have 25 shots or 25 melee attacks).

4) The points. It helps to find a model from an official rulelist that has as close of a statline as you can, and modify from there. I haven't made an army list in a while, so my memory may be a bit fuzzy here, but this should be approximately right (These are all assuming you're figuring it out for a Troop/Team or single other model. DON'T figure out a section/regiment or platoon/horde without figuring out the cost for a Troop/team unit first, even if the troop/team is not part of your final army list):

(Note: These all assume you're giving it better stats then the comparative unit. If it's worse, simply reduce it by that many points instead)

Speed (Average: 5): +10 points if there is significantly more movement (Fairly arbitrary, something like 3 -> 6 or 7 for example. Ignore this for units with Fly/Jump Pack)
Hit/ Melee (Average 4+): +10 points for every point of difference (ie, 4+ -> 3+ is +10 pts, 4+ -> 5+ is -10 pts)
Ranged (Average 5+): Same as for Hit/Melee
Def (Average 4+): Same as for Hit/Melee
Attacks/Firepower (Average 1/model in unit): +5 pts for every 5 additional attacks (Keep in mind all KoW and WP pts are in multiples of 5, so when in doubt, round up)
Range (Average 24): +10 pts for every additional 12 in range
Nerve (Average 10/12): +10 pts for every additional 1/1 (Nerve can be simplified to a single number, and then adding that number to 8/10 to find the nerve stat. This is important for calculating the Nerve of a unit of a larger size like a Troop -> Regiment)

Special: Can buy X special weapons in warpath does not cost additional points.

5) Special Abilities: The addition of all special abilities is +10 points, except for the exceptions mentioned below. In the case of a Special Rule with (X), it counts as +10 pts per X:
Breath Weapon/Flamer: +10 pts per 2X (ie, Flamer (4) would be +20 Pts)
Fly/Jump Pack: +20 Pts
Craven/Yellow-Bellied: -10 pts.
Blast: +10 pts per 1d3, +20 pts per 1d6 (So Blast (2d6) would be +40 pts)
Lumbering/Shambling: No cost
Transport: +10 pts per 5X (ie Transport (20) would be +40 pts)
Regeneration/Bloodbath: +20 pts

6) Army Special Ability: Every army gets an additional Special Ability that usually affects all their units. This is worth 10 pts as mentioned in step (5), and is not calculated into the cost of each additional unit (You can choose an ability or abilities worth more than 10, but then you have to start factoring that in to unit cost).

7) For Sections/Regiments and Platoons/Hordes (Note that this is different for KoW and WP, so don't mix them up!):

KoW:
Troop/Half Regiment -> Regiment: Points = (2 * Troop) - 10. Nerve = Troop Nerve * 2 (See step (4) about figuring out your single Nerve stat)
Regiment -> Horde: Points = (Regiment * 2) -10. Attacks = Attacks * 2 (Including doubling any Breath Weapon/Flamer/Zap! special ability). Nerve = Regiment Nerve *2

WP:
Team -> Section: Points= (Team * 2) -10. Attacks = Team Attacks * 2. Firepower = Team Firepower * 2. Nerve = Team Nerve * 2.
Section -> Platoon: Points= (Section * 2) -10. Attacks = Section Attacks * 2. Firepower = Section Firepower * 2. Nerve = Section Nerve * 2.

8) Available unit upgrades: For WP, these are usually BFGs or special melee weapons. Usually, the number a unit gets is 1 for Teams, 2 for Sections, and 4 for troops.
For KoW, Banners cost 15 points and Musicians cost 10 points.
For other unit upgrade ideas, I'd advise looking at existing army lists. Usually, if you'd be reducing a stat by an equal number of points that the upgrade would raise it by, make it a free optional upgrade. Otherwise charge a number of points equal to the difference in cost between the lost stats and the upgrade.

9) BFGs and Special Weapons in Warpath: For a BFG, again find a similar existing BFG and apply steps (4) and (5) to find the new cost. For melee weapons, do the same but count each extra die rolled as +10 pts (Instead of +10 per 5 as per the guidelines in step (4)).

Do this for everything, keeping in mind most:
KoW armies have 6-7 infantry units, 1-2 Cavalry units, 1 Large Infantry unit, 1 Large Cavalry unit, 2-3 War Machines, and 7-8 Heroes/Monsters
WP armies have 6-7 infantry units, 2-3 Ordnance units, 3-4 Armor units, 1 Aircraft unit, 3-4 Heroe/Monster units, 4-5 BFGs, and 2-3 Special Melee Weapons.

10)And you're done!

Please keep in mind this is a very rough outline, and will need playtesting in order to make sure it is costed appropriately. But be sure to have fun with it, and please let me know of any comments, feedback, or criticisms in the comments!

[scarletsquig]

Great writeup! This is a really good start, and has got me thinking. Quite interesting that the same approach is used for both Warpath and KoW. It does seem like there are enough similarities there to make the same sort of approach usable for both.

A few thoughts...

For starts, percentage-based costing for the special rules would be better. A 200 pt unit with Fly is way more deadly than a 50 pt one, and should pay more for it, rather than it being a flat 20 points. This design mechanic is evident in the heroes/monsters section of the official lists where the more expensive heroes pay more for their steeds.

Some sort of scaling on the basic stats might be in order too, for example, De 6+ is really powerful (compared to De 5+ it halves the amount of damage you take, whereas the move from De 4+ to 5+ only reduces it by a third).

So, something like

De 2+ : -25 points
De 3+ : - 10 points
De 4+ : + 0 points (average)
De 5+ : +10 points
De 6+ : +25 points

.. might work out better for building the core of the unit, and even then, this is still flawed.

Modifications to De might need to be based on Nerve, otherwise it comes down to the same issue as Fly.. where you could make a unit have 100 nerve and then only pay a measly +25 points for De 6+. There is a certain level of "stats dependant on other stats" that should probably be worked in.

So, De is intrinsically linked to Nerve, since both are "defensive" stats. Me and Ra are linked to At, but not to each other, and Sp is entirely seperate and is a universal boost to all of the other stats.

Meaning that De could then be something like (reducing Ne down to it's "single stat" form, so average of 2):

De 2+ : - (12.5*Ne) points
De 3+ : - (5*Ne) points
De 4+ : + 0 points (average)
De 5+ : + (5*Ne) points
De 6+ : + (12.5*Ne) points

.. meaning that the crazy theoretical 100 Ne Unit would at least have to pay 1250 points for its De 6+ instead of 25.

I'm thinking of a layered approach for calculcating the basic statline:

1. Determine Ne and At. Flat +10 points per Nerve, +1 point per At.
2. Determine De, points based on Ne.
3. Determine Me and Ra, points based on At.
4. Determine Sp, points based on the entire points value of the statline so far.
5. Determine special rules, points based on relevant parts of (or all of) the statline.

Finally, how many points would be charged for the "average" statline? Need that core in place, to base the rest of the modifications off of.

Ideally, the acid test of the entire system should be taking the formula, plugging in the stats from any unit in the existing rules and ending up with a reasonably close match for the points values across the board, erring on the "slightly overpriced" side of things if there is any imbalance.

Edit: Just going to add some thoughts on Me and Ra, they're really easy since you can determine their points based on the odds of hitting. This isn't based on the "core statline" approach, but could easily be changed to fit it).

Me 2+: +((5/6)*(3*At)) points
Me 3+: +((4/6)*(3*At)) points
Me 4+: +((3/6)*(3*At)) points
Me 5+: +((2/6)*(3*At)) points
Me 6+: +((1/6)*(3*At)) points

Copy+paste for Ra?

Basically, just taking the At value and plugging in the d6 roll odds to determine the points. Actually, I should probably extend this to De too. As for rounding up to the nearest five, that can be ignored until the end of all the calculations.

[darkPrince010]

I like it. I may be able to try plugging in the values tonight, but worst case scenario I'll get something figured out sometime tomorrow and post how well your initial version works (Personally it seems like that'll end up with far better balance than my flat point buy would). Thanks!

[scarletsquig]

Cool. I think if we can get the basic statline working, and matching up pretty well to all units in the official lists that don't have any special rules, then that will be a good groundwork for moving on to the special rules afterwards.

I'm wondering about the "average statline" approach, and if it is going to hold up. It's fantastic for modifying existing points values and writing a balanced list that way, making minor tweaks to whatevers there, but I'm wondering if starting from zero on all stats might be better, or possibly from a core minimum of Ne 2. I can't recall any unit in either KoW or Warpath that has a value lower than 8/10,so that could be a good solid default that costs, say 5 points. Ne is also the only stat (along with Sp?) that doesn't appear as "-" in one or more profiles, so starting from zero would allow these kinds of units to exist.

I'll definitely make the flash app for this purely because it will allow for a much more complex approach if the computer can handle all the number crunching and all the person has to do is just work out what stats and rules they want.

[mattjgilbert]

I think starting from an average line is the better approach.

You should also consider scaling the value of skills too, based on the stat line and combination of other skills (without overcomplicating it which will be a challenge).

Personally at this stage I'd focus on KoW as it's the more developed in terms of game design and army balance.

[Baragash]

I created a spreadsheet model that allows me to do this. I took a unit of 10 Elf Bowman as the base unit for the entire system, and kept their starting points of 90.

Then I broke all the stats down and analysed their effectiveness in reference to the Bowmen unit.

Of course the weighting is arbitrary, and I gave everything equal weighting to begin with, which is debateable, I can't remember where it sits now.

So I divided into offensive and defensive stats, and with equal weighting that means 50% of the points go to each.

Offensive was broken down into Movement, Shooting and Melee (again equal weighting) so each is effectively 17% of the points (33% of 50%).

Defence was broken down into damage received in Melee, damage received from Shooting and Nerve.

Move: so if a unit is 50% faster than the Bowmen, it will upweight the points by 50% (so in the example as it stands the Elves pay 15pts for their move, the hypothetical unit pays 22.5pts).

Shooting: I evaluated each unit's average damage in each 12" range band (0-12", 12-24" etc), against all four defensive values (3+, 4+, 5+ and 6+) but gave double weighting to 4+ and 5+ for being more common. As before the assigned base points from the Elves is up- or down-weighted based on the comparative % effectiveness. A Melee unit scores zero and receives 0 points. Note that magic gets rolled into here too as each outcome can be evaluated as a shooting attack causing X "damage".

Melee: Same method as Shooting basically, but without the range band issue.

Both Defensive versions were done in the same way using damage received instead of damage caused, and weighting 4+ and 5+ enemy to hit stats as double compared to 3+ and 6+.

Nerve was evaluated on the basis of the number of damage points needed to be scored such that a 7+ would Rout the unit (ie it becomes a more common outcome than not being routed). IIRC I weighted that 75% damage to Rout against 25% damage to Waver as Wavering is only a 2 point band in most cases.

Then I layered the impact of all the special rules into the formulas, bar a few like Inspiring which don't really translate into maths, though I have an idea how to proxy some of them.

The upshot is that it confirmed my view of the KoW points system as being some weird Warhammer pricing left over based on unit size and not ability.

Some examples from my system (KoW/my analysis):
90/90 Elf Bowmen Troop
160/100 Elf Bowmen Regiment
60/65 Elf Spearmen Troop
60/65 Dwarf Ironclad Troop
110/75 Dwarf Ironclad Regiment
70/80 Dwarf Ironwatch Troop (Xbows)
95/85 Dwarf Ironwatch Troop (Rifles)
245/165 Dwarf Ironwatch Horde (Rifles)
60/70 Orc Ax Regiment
60/70 Orc Greatax Regiment
85/110 Orc Morax Troop

[scarletsquig]

^ Interesting approach, although in my case I'd be looking at whether I took the right approach to writing the formula as opposed to assuming that Unit Size is a major factor.

- Equal points distribution between the stats will result in way too many points assigned to At, since it is the largest stat on the profile. This may be why your approach works well for troop units, but breaks down for large units. Speed is also really important, and Me and Ra really should depend on AT, just like De should be linked to Ne! I really don't think attempts to cut the statline up like a pie are going to work since there are too many dependencies, and Sp in the KoW rules is clearly determined as a % of the rest of the statline, if the hero mounts are any indication.
- The range weighting is an interesting one. It's definitely a more complex stat to handle, and Ra seems much more valuable on the table than Me since you get to make more attacks with it during each game. I was thinking 4*At at least to plug into the Ra modifier, and reducing that back down to 3* for 12" range troops, adding 1 to the multiplier for each 12" range band, for something like: Ra: +((d6 to hit odds)*((Range/12 + 2)*At)) points.
- I like the efforts to calculate overall damage causing potential, not sure about the whole double weighting for 4+ and 5+ due to them being common though, this approach doesn't gel well with a formulaic approach since it takes into account the stats of other units rather than the statline on it's own. That's a whole extra layer of meta-complexity that I hadn't considered, and it's effectively trying to do the job of playtesting.

Any chance of posting the formulae, uploading the spreadsheet to google docs or something? I'd look over the approach more, but can't comment too much without seeing it. The formulae you came up with for the special rules would be interesting to see - percentage modifiers for the base stat or something else?

I think starting from an average line is the better approach.

What statline should be considered average, and how many points should be charged for it? How would be best to handle negative points reductions from it, such as "-" Ra for all Melee units, and "-" stats for a lot of units without getting a result that potentially goes below zero points for a minimum-statlined unit? I'm thinking of the low end here, like army standard bearers and other small units.

- If starting from a "zero stats" approach gave very similar end result as an average stat approach, could it be considered just as valid? After all, the starting point really doesn't matter since any average statline approach can be converted into a zero-stat approach just by adding in some fancy exceptions.

Me could be changed to (using an average statline approach, with the same "based on a d6 roll" approach):

Me 2+: +((2/3)*(3*At)) points
Me 3+: +((1/3)*(3*At)) points
Me 4+: +0 points
Me 5+: -((1/3)*(3*At)) points
Me 6+: -((2/3)*(3*At)) points

.. which should work out the same? Way, way way messier though since you have to start using 50% as your zero-point for the calcs which makes the maths one hell of a headache compared to my nice clean 1/6, 2/6, 3/6 etc. calcs for Me that can be used by starting from a (not unreasonable) assumption that a unit pays zero points for the ability to do zero damage. It's just a pain, and I'd much rather start from zero while hashing out the formulae and then convert to an average-based system at the end if it helps to iron out any issues.

Agree with dealing with KoW first, I was already running with that assumption. Also, the intent is to generate reasonably accurate points values that err on the side of costing a little bit too much. I'd be pretty happy with a system that can match up to the actual KoW profiles without any of the points going under the official numbers and with most of them ending up about 5-20% over.

[Baragash]

- Equal points distribution between the stats will result in way too many points assigned to At, since it is the largest stat on the profile.

That's not what I did or what I said I did. (This is also why a linear regression only works if you create a base conversion of the stats first).

Speed is also really important, and Me and Ra really should depend on AT, just like De should be linked to Ne! I really don't think attempts to cut the statline up like a pie are going to work since there are too many dependencies, and Sp in the KoW rules is clearly determined as a % of the rest of the statline, if the hero mounts are any indication.

I have based it on the stats that are linked together. I don't personally believe that the hero mount costs are particularly indicative of a formulaic approach on Alessio's behalf, he doesn't like mathhammer....

This may be why your approach works well for troop units, but breaks down for large units.

....and he wouldn't have ended up with the cost unit scaling he did if he had used such an approach.

The reason it breaks down for large units is because the KoW points values use an arbitrary and obviously (at-a-glance obviously) illogical scale.

10 man unit -> 20 man unit = basically double the cost for a couple of nerve points
20 man unit -> 40 man unit = basically double the cost for double the attacks and a few nerve points

There is no logical defence in favour of a 20 man unit being twice as effective as a 10 man unit. It's both counter-intuitive and mathemtically illogical/impossible. A 40 man unit is broadbly speaking twice as effective as a 10 man unit. The results I get are intuitively logical as it prices a 20 man unit at 10-20 points more than a 10 man unit and a 40 man unit at about double, which is what a casual glance at the stats would suggest is the expected result.

Which is why this

Agree with dealing with KoW first, I was already running with that assumption. Also, the intent is to generate reasonably accurate points values that err on the side of costing a little bit too much. I'd be pretty happy with a system that can match up to the actual KoW profiles without any of the points going under the official numbers and with most of them ending up about 5-20% over.

is not achieveable.

- I like the efforts to calculate overall damage causing potential, not sure about the whole double weighting for 4+ and 5+ due to them being common though, this approach doesn't gel well with a formulaic approach since it takes into account the stats of other units rather than the statline on it's own.

Because the model calculates effectiveness the entire model takes into account the stats of other units. Adjusting the weighting simply tells the model that a particular combination occurs more often in the game than others.

It's inherently impossible to create a formulaic approach regardless of complexity without meta-gaming because you can't possibly determine the value of a stat/stat combination without reference to the range of those existing in the game. All weighting does is establish the parameters of your base case.

Any chance of posting the formulae, uploading the spreadsheet to google docs or something? I'd look over the approach more, but can't comment too much without seeing it.

I'll have a look at Google docs at some point.

The formulae you came up with for the special rules would be interesting to see - percentage modifiers for the base stat or something else?

Everything is calculated on effectiveness, so all it does is adjust the formula if that special rule's checkbox is ticked. Crushing Strength is an easy example, because it calculates the effectiveness against the four Def values, if the box is checked it just changes the % chance of success (by a % dependent on the value of CS).

[scarletsquig]

I have based it on the stats that are linked together. I don't personally believe that the hero mount costs are particularly indicative of a formulaic approach on Alessio's behalf, he doesn't like mathhammer....

I asked him in person, and he said that yes, there is a formula used for generating points that incorporates a "core" profile and modifys it from there (so, the approach that the fan list authors have been using is in fact a really good one since it is a direct extension of what Alessio does, only easier because it benefits from having hundreds of base profiles to choose from rather than one).
Naturally, this gets tweaked for balance, so any entirely formulaic approach is not going to be 100% accurate, and for list building, experience, playtesting and a good eye for balance will create far better lists than any formula, but I disagree with your statement that it is impossible to even come close, which is all that is really needed for a unit builder... I'm thinking along the lines of
"Someone has a FW Basilisk in their army, wants to make some rules. Type stats into a program, and within 2 seconds, there's a reasonably-costed, probably slightly overpriced points value determined for the critter that will do the job for someone who wants to use the model in a friendly game and have it munch some elves."

I'm not trying to create a shining example of perfect computer-generated game design, just to get close enough, which is what the old VDR rules for 40k did an excellent job of (after they were updated to cut out the cheese).

Anything can be modelled with mathematics.

There is no logical defence in favour of a 20 man unit being twice as effective as a 10 man unit. It's both counter-intuitive and mathemtically illogical/impossible. A 40 man unit is broadbly speaking twice as effective as a 10 man unit. The results I get are intuitively logical as it prices a 20 man unit at 10-20 points more than a 10 man unit and a 40 man unit at about double, which is what a casual glance at the stats would suggest is the expected result.

Or your system simply isn't working out (not to dismiss your efforts, it sounds like you've taken a very thorough and internally logical approach). If your 20-man units are only 10 points more than that indicates that not enough importance is being given to Ne or At, since those are really, really important. They could even be important enough to be percentage modifiers and I could be taking the complete wrong approach by basing De off Ne instead of the other way around (in fact, considering how the army lists look, this is probably the case, increases in Ne and At really jack up the points).

I agree that the approach taken is not logical according to the damage output/ damage taken approach that you've taken, but nevertheless, KoW is pretty well-balanced so Alessio definitely did something right somewhere along the lines, and I don't agree with you that he's a bad designer who doesn't like maths. In fact, I have the exact opposite opinion. Any work that he did for GW should not be taken into account, because there are far too many external pressures involved when working for that company. If GW wants to sell a new plastic unit of UberMarines with UberFists riding UberDinos to kids, then it's a safe bet that the designer will have been told "make these overpowered so we can sell loads of them".. the CEO of GW has outright publically stated that rules only exist to sell models, so I flat-out disregard any criticism of the relative skill of individuals working for the company when it comes to rules balance. Every ex-GW designer I've seen has produced far superior work after they left the company and got to run with their own ideas, and the people working for Mantic are no exception.

As for the meta-game balancing that is not a concern of mine at all. Only making sure that the formula generates stats that are roughly in-line with those found in the rulebook. The balancing for the game has already been done in terms of "army lists exist, and have been well-tested", so anything that correlates with them should also be effectively balanced.

You think that it's impossible, I completely disagree. That's all. Can't say whether I'm right or wrong on that point until I've given it a go*. I might even end up drawing the same conclusions.

*Which might take a while tbh, I'm really busy right now, and this was just a random idea that I thought might be kinda nice as a thing for KoW players to have fun with.

[Jonathan Faulkes]

Mmmm. Simultaneous equations! 6 unknowns with a final result (points), just need to choose your equations carefully. My suggestion would be stick to solve the equations for troops, regiments and hordes seperatly.

[orcsbain]

This is EXACTLY the sort of thing that I was thinking of to include in the book (and thus hopefully point Alessio towards).

A "official" way to include all my/the odd, weird and OOP minatures that are available out there, in FRENDLY (that is NON Tournament) games.

Perhaps if it can be nailed here it can just be looked over and tweaked to get into the book. Wlouldnt that be nice, to see some fan generated gaming rules in an official release.

keep it going guys and perhaps NS can pass on a linky to the man himself

ttfn

[Max Jet]

Wlouldnt that be nice, to see some fan generated gaming rules in an official release.

Like the new nerve system?

I just want to say: Careful people! All creature generating rules suffer from very, very bad balance. I know this, god knows how I abused the creature design rules for Tyranids. It is impossible to forsee all the holes and loops and to this date the KoW rules move as close towards balance as possible for a wargame, that isn't chess.

I will keep an eye on this too, but just be carefull what you wish for.

[scarletsquig]

Mmmm. Simultaneous equations! 6 unknowns with a final result (points), just need to choose your equations carefully. My suggestion would be stick to solve the equations for troops, regiments and hordes seperatly.

Yep, both you and Baragash might be totally right about there being a huge and weird disconnect between troops/ regiments /hordes that just throws everything overboard.

I'd like to give a totally formulaic approach that gives great importance to Ne and At a shot first before giving in and admitting that it just doesn't make any sense and having to factor in Unit Size into things.

Only if that fails will I go for an approach that involves either taking 3 different formulae (horrible) or calculating Regiments and Hordes as a percentage of Troops (even more horrible).

I just want to say: Careful people! All creature generating rules suffer from very, very bad balance.

Yep, it's going to be awful, so the general idea is to deliberately overcost so that most of the variance falls on the underpowered side of the line.

I'm going to start with Dwarf Ironclads as the test case. They don't have any of the special rules or missile troops that really mess with things. Keeping it dead simple for now. Missile troops and special rules are a headache that I'll save for later.

Edit: Okay, here's my initial scribbling attempt at getting this to work (spreadsheet here, although you'll need to download it to play with it for now):

Step 1. Base profile, starts with 5 points for a statline of Ne 2 and zero on everything else.

+1 point per At.
+ 10 points per Ne over 2.

Step 2.

Me = + ((d6 odds to hit) * 3 * At)) points.
De = + ((d6 odds to be not damaged) * 5 * Ne)) points.

Step 3.

Total everything up.
Speed = +5% points for every Sp point.


Here's what I get when plugging the dwarf ironclads into that:

Ironclad Troop: 59
Ironclad Regiment: 107
Ironclad Horde: 249

Actual points in the KoW rules are 60/110/210.

So far, it is going a little high on hordes, but is working out pretty well!

Needs a ton of refinement obviously, and is probably going to fail hard when I test it on other units (which will allow me to identify the offending stats), but I think it's a good indication that it is not impossible to make this happen, it's just that Ne is considered really, really freaking important in the rules, and that needs to be factored in.

Edit: Giving it a spin with Dwarf Ironguard and Berserkers, a couple more melee units without special rules. My values on the left, official values on the right:

Ironguard Troop: 31/35
Ironguard Half-Reg: 65/65 (Woo!)
Ironguard Regiment: 113/130

Berserker Troop: 69/50
Berserker Half-Reg: 130/90
Berserker Regiment: 233/170

Not bad, so far. Next step will be to compile a list of all non-ranged, no special rule units in the game and cost them using the formula to identify the parts that need changing.

For instance, the Berserkers kinda indicate that I'm placing too much weight on At currently, or that I'm charging too much for low-De units.

[mattjgilbert]

The step up from one unit size to the next in terms of points cost is artificial in KoW. The best approach is probably to NOT try and invent some all-purpose formula for calculating all unit sizes in one go but fix on ONE and then apply the same scaling logic that the system has. That is: (roughly) double -10 points for each size increase. That's what the official lists do and how all fan lists have been done. I don't see the point in trying to figure out something else which will require a large amount of brain power for little gain.

[Baragash]

I agree that the approach taken is not logical according to the damage output/ damage taken approach that you've taken, but nevertheless, KoW is pretty well-balanced so Alessio definitely did something right somewhere along the lines, and I don't agree with you that he's a bad designer who doesn't like maths. In fact, I have the exact opposite opinion. Any work that he did for GW should not be taken into account, because there are far too many external pressures involved when working for that company. If GW wants to sell a new plastic unit of UberMarines with UberFists riding UberDinos to kids, then it's a safe bet that the designer will have been told "make these overpowered so we can sell loads of them".. the CEO of GW has outright publically stated that rules only exist to sell models, so I flat-out disregard any criticism of the relative skill of individuals working for the company when it comes to rules balance. Every ex-GW designer I've seen has produced far superior work after they left the company and got to run with their own ideas, and the people working for Mantic are no exception.

Just because I have a low opinion of the points values in KoW, should not be taken to think I have a poor opinion of Alessio as a games designer, broadly speaking I like KoW.

Or your system simply isn't working out (not to dismiss your efforts, it sounds like you've taken a very thorough and internally logical approach). If your 20-man units are only 10 points more than that indicates that not enough importance is being given to Ne or At, since those are really, really important. They could even be important enough to be percentage modifiers and I could be taking the complete wrong approach by basing De off Ne instead of the other way around (in fact, considering how the army lists look, this is probably the case, increases in Ne and At really jack up the points).

Well At is not relevant as the vast majority of units receive no additional At between those two sizes.

You can do a simple experiment to demonstrate my point. Pick any 20 man unit. Pick 2 10-man units of the same type, say, Elf Bowman. The 2 10-man units cost almost the same as the 20-man unit (180 vs 160). Except for the same points, the 2 units have double the attacks (and in this particular case, the 20-man unit's Nerve advantage only lasts for 5 points of damage, after that the fact the 2 units effectively have twice as many wounds).

You could look at Ironclads in the same way, and playing out the example would show a higher contrast as throwing twice the number of attacks at Def 5+ will make a huge difference.

The issue doesn't permeate the whole system. The jump from 5 to 10 man units seems broadly speaking about right, the jump from 20 to 40 is conceptually about right but it's out of sync because the jump from 20 to 10 isn't right, and it's that jump that breaks any attempt to have an over-arching system.

The step up from one unit size to the next in terms of points cost is artificial in KoW. The best approach is probably to NOT try and invent some all-purpose formula for calculating all unit sizes in one go but fix on ONE and then apply the same scaling logic that the system has. That is: (roughly) double -10 points for each size increase. That's what the official lists do and how all fan lists have been done. I don't see the point in trying to figure out something else which will require a large amount of brain power for little gain.

Well you can just run a math system across 10-man units and then apply the KoW scaling system.