Skill Ratings & Point Value - Feedback Wanted

[Xotl]

So, the PV revision is now live on the MUL, but there's one last little bit that needs to be tackled.  Skill Rating PV improvements need to be increased in value.  This is because:

1) They were always off, in that they followed the same linear cost progressions that PV used to, while the effect of an increase is anything but linear.

2) The increase in costs to lights to make them equivalent to heavier units has meant that they disproportionately suffer in a world where any unit can cheaply buy the upgrade needed to counteract a light's one defense: their difficulty to be hit.

So, here's the current cost system:

Code: [Select]

IMPROVED-SKILL PV INCREASE TABLE
    Unit’s Base PV   PV Increase per Rating
0-7                 1
8-12                 2
13-17                 3
18-22                 4
23-27                 5
28-32                 6
33-37                 7
38-42                 8
43-47                 9
48-52                 10*
*Increase PV increase by 1 point for every 5 base PV over 52.
And here's some rough stats for people to work from:

Assuming Medium Range and a +1 def modifier (29% of all units in the game have this), the to-hit chance associated with these Ratings are:
4: 58.33%
3: 72.22%
2: 83.33%
1: 91.66%
0: 97.22%

Assuming Medium Range and a +3 def modifier (16.5% of all units in the game have this), the to-hit chance associated with these Ratings are:
4: 27.77%
3: 41.66%
2: 58.33%
1: 72.22%
0: 83.33%

The question then becomes: what should a revised system look like, bearing in mind the reality of the newly implemented PV revision and the need to keep AS simple?  Ideally, what I'd like is something that scales in cost, just as the effectiveness of the Skill Rating bonus does.  Keeping Skill Rating 3 relatively affordable as it is now is fine, but the climb from there I think needs to matter more, based on how much more effective it is.

To start, I'd like to ask: what level of Skill Rating increase do you actually buy in your games?  I have the feeling people are either going to Rating 3 or Rating 0, but I'd like to see some proof.

Please let other AS players know that this thread exists and point them this way if they're interested in contributing.  Thanks.


EDIT
Link to unit spreadsheet if people want to mess around with stuff:
https://www.dropbox.com/s/px3fu6669e40jzu/Alpha%20Strike%20PV%20-%20Output.xlsx?dl=0

[Hellraiser]

Wouldn't it be easier to just come up with a % figure of the PV.

For example.

Every Point the Skill is lowered is a 33% Point Increase (Rounded or Rounded Up)

The scale list seems pretty low cost compared to the BV chart for BT skills.

33% might be too much, perhaps 25% would be better, but the 20% roughly that I'm seeing below seems a tad bit too low.

[Tai Dai Cultist]

In my local games we have an informal etiquette of not going beyond skill 3, or skill 2 in the singular case of a list's commander.

Inside this informal constraint, I like to buy skill upgrades for any unit that "hits hard" in comparison to what else I may have in the list.  Actually, my high-TMM minxy units rarely get a skill upgrade unless I'm reserving sufficient points to go to 3s "across the board".  (But often even in this case, I'll go with spending some of the 'skill upgrade' points on using C3 units)

Outside of this local etiquette: I'm not sure.  We banned 1s and 0s so early in I actually don't have a lot of experience playing with them.  I'd probably take 1 50PV mech munchkin'd up to skill 0 for 90PV rather than take 2 of that same mech left at skill 4 for 100PV.  And any infantry unit with medium range- go to skill 0 every day and twice on sundays. Probably the same for any artillery I may have brought.


Some personal thoughts: I agree that 3s should be about as affordable as now. And that people are probably willing to spend much more than what's necessary to go to 3 as is.  And that 1s and 0s just break the game in effectiveness.  So a wild-ass peanut gallery suggestion: what if the costs were exponential rather than linear? Keep the "PV increase per rating" but you pay it once to improve once.. and rather than twice to improve twice you pay once AND twice for thrice in total.  Example: 25 PV unit costs 5 per skill upgrade as is. going to 3 remains the same.  going to 2 is calculated as 25+5+5+5.  Taking that in turn to 1 now costs 25+5+5+5+5+5+5, and going all the way down to the ridiculous skill 0 costs 25+5+5+5+5+5+5+5+5+5+5 for a whopping 75 points vs the 45PV it currently costs.

[nckestrel]

The one problem with going by to hit only, is that a unit with Skill 0 is still destroyed just as easily.  You pay 100 PV for a Skill 0, and it dies on turn 1 from concentrated fire and that skill 0 was worthless.

Skill increases the effectiveness of the firepower, but it does nothing for survivability. Treating the entire PV as effective firepower..well, it certainly is important, but it's not everything.

But I do agree the current costs are a bit low.

We might need a minimum PV increase per skill.  (See comments about infantry).

[Tai Dai Cultist]

..Skill increases the effectiveness of the firepower, but it does nothing for survivability. Treating the entire PV as effective firepower..well, it certainly is important, but it's not everything...

I simultaneously agree and disagree with that.  On one hand you're right... but on the other hand the best defense of all is killing the other guy first.

[NeonKnight]

Was thinking on this the other day in a discussion with a friend, and my (our) thoughts were, it is hard to have a linear PV cost increase.

I agree with nckestrel, a dead unit that dies in round 1 is expnsive, but my thoughts are also having played a great many different competitive point based games in the past, that's just the breaks. It's also no different in regular TT where a lucky hits hits the head with a Gauss or a PPC or a TAC followed up by two Gyro or Ammo or 3 engine crits.

But the point I want to try and make here, (and let me preface this with math is NOT my strong suit), is skills in BT are not linear. In D&D Minis for example (an old game I competitively played a lot), the To Hit scale was d20 based. 20 always hit, 1 always missed, and lowering any to-hit roll by a factor of 1 was always just a 5% increase in the to hit odds:

20 = 100%
19 = 95%
18 = 90%
17 = 85%
etc.

BT is not, with the 2d6 as shown by Xotl above, the to-hit chance follows more of a BELL curve. Additionally, using a skill 4  (default) against that +1 TMM (doesn't matter if +2 or +3, the maths still the same), we know the following:

4: 58.33%
3: 72.22% - 13.89% increase
2: 83.33% - 11.11% increase
1: 91.66% - 8.33% increase
0: 97.22% - 5.56% increase

So, it's actually a case of diminishing returns. Improving the Skill from a Skill 4 to a 3 means an increase in to hit chance of almost 14%, but increasing from 3 to 2 is only 11%, and then drops to 8.3% and then 5.6%

I think the PV increase should follow something similar. Say for example something like (Remember I not good at maths, this just an example):

SKILL 4   +0% PV
Skill 3*   +4% PV
Skill 2*   +3% PV
Skill 1*   +2% PV
Skill 0*   +1% PV

* Each Skill increase includes the increase(s) above it.

For Example. to increase from Skill 4 to 3, the PV is increased by 4% (again, just an example), but to increase from 3 to 2, increase it by 4% + 3%, and on and on.

Again, not sure of the final numbers, but there it is.

[Elmoth]

Hitting +1 vehicles and mechs is fairly easy. Hitting fast buggers (+3 or +4) is not. Been there as a GM being hit and not hitting myself. Not a lot, since we are noobs but it certainly happened. The cost hike that we saw recently (and that makes a lot of vehicles more expensive that mechs of lower tonnage) seems to go in that direction

a +1 to hit is roughly a +11-17% on your options to hit for the central numbers in the distribution. Let's say +14% on average (and taking into account that this is a gauss distribution). A +15% cost increase seems legit for each +1 you get out of it. Going on a triangular progression (cost 1, 3, 6, 10, 15, 21...) progression could work as well since it will make going to extremes very expensive.

I would like the later. You can go Veteran if you fancy it at the cost of a few resources. If you want Elite your price goes up rapidly. And legendary status is well, Legendary. In the opposite duirection, performace degrades rapidly as well.

[Valkerie]

So far we have (informally) kept the skill improvement to 3 or 2 to meet PV requirements.  I agree with you Xotl, I would like to see the jump to skill 3 remain affordable, but cost a bit more to make the jump to 2 and so forth. 

[NeonKnight]

So far we have (informally) kept the skill improvement to 3 or 2 to meet PV requirements.  I agree with you Xotl, I would like to see the jump to skill 3 remain affordable, but cost a bit more to make the jump to 2 and so forth.

I agree it should cost a lot to go to Skill 0, but Skill 4 to 3 should be the biggest cost, and then go down.

Again, a skill 4 to 3 see the biggest % in effectiveness (14%), while Skill 1 to Skill 0 is lowest (5.6%)

Each skill increase should build on what came before, otherwise again, all you are doing is making a linear increase

[Scotty]

Above all it needs to stay simple enough to do the operation at the table, and avoid fractions and rounding.

My gut says tie it to damage, since at the end of the day that's what it affects most.  What about (values subject to change) 2x the unit's Medium range damage for 4 to 3 and 3 to 2?  If a unit doesn't have medium, use its Short instead.  Then, from 2 to 1 and 1 to 0, make it 1x the unit's longest range damage bracket.  If a unit has no damage values at all (for Mechs) use its size (looking at you, Strider M) to represent physical attacks.

Hm.  That let's in the relatively small subset of melee specialists that do more in melee tab with guns as a loophole.  I'm not sure I'd be able to nail down the potential language while I'm on my phone.

[NeonKnight]

Above all it needs to stay simple enough to do the operation at the table, and avoid fractions and rounding.

My gut says tie it to damage, since at the end of the day that's what it affects most.  What about (values subject to change) 2x the unit's Medium range damage for 4 to 3 and 3 to 2?  If a unit doesn't have medium, use its Short instead.  Then, from 2 to 1 and 1 to 0, make it 1x the unit's longest range damage bracket.  If a unit has no damage values at all (for Mechs) use its size (looking at you, Strider M) to represent physical attacks.

Hm.  That let's in the relatively small subset of melee specialists that do more in melee tab with guns as a loophole.  I'm not sure I'd be able to nail down the potential language while I'm on my phone.

You Forgot the Ostscout 7K ;)

But I think you are on the right track. 2x Medium Range for Skill 4->3 and 3-> 2, but instead of bringing long range into the mix, leave at Medium Range for 2->1 and 1->0 but have it increase to the simple +1 for each.

For the few outliers (like the Ostscout 7K) that have no damage values use a default value of 1 for Medium range.

Yes, there may be a few 'corner cases' out there but I think a good starting place is work first for what works for 95% of the units first and then worry about the corner cases and see how badly they can be abused.

[Tai Dai Cultist]

I agree it should cost a lot to go to Skill 0, but Skill 4 to 3 should be the biggest cost, and then go down.

Again, a skill 4 to 3 see the biggest % in effectiveness (14%), while Skill 1 to Skill 0 is lowest (5.6%)

Each skill increase should build on what came before, otherwise again, all you are doing is making a linear increase

I disagree completely with this line of thought.

Because going from 1 to 0 might be a small-to-negligible increase in % at one point on the bell curve, it could also be the same % increase as 4 to 3 under different circumstances.  You can't price the skill increases based on only one alignment to the bell curve.. even Xotl's OP uses two alignments (and to reiterate, skill 0 is much more valuable in the second example than it is in the first)

I'm full-in that each skill upgrade should increase in price with more acceleration than a linear increase.  Conversely, I also support skill DECREASES giving back diminishing returns on PV after the initial skill dump.

[NeonKnight]

I disagree completely with this line of thought.

Because going from 1 to 0 might be a small-to-negligible increase in % at one point on the bell curve, it could also be the same % increase as 4 to 3 under different circumstances.  You can't price the skill increases based on only one alignment to the bell curve.. even Xotl's OP uses two alignments (and to reiterate, skill 0 is much more valuable in the second example than it is in the first)

I'm full-in that each skill upgrade should increase in price with more acceleration than a linear increase.  Conversely, I also support skill DECREASES giving back diminishing returns on PV after the initial skill dump.

Sorry, I will disagree with that.

If you and I both agree that a Skill 0 pilot should be 20PV (actual amount does not matter) above a Skill 4 pilot, does it matter how the value got there?

If I say, the steps are:

Skill 4 = 0
Skill 3 = +10
Skill 2 = +5
Skill 1 = +3
Skill 0 = +2

Because the to hit rolls get lower and lower % increases

and you say it should be

Skill 4 = 0
Skill 3 = +5
Skill 2 = +5
Skill 1 = +5
Skill 0 = +5

At the end of the day both charts represent the same 20 point increase, but what I am proposing shows where those same PV increases also have the most ingame effect.

[Elmoth]

Usual to hit numbers are +2 to +5 over your skill rating. That makes going to low numbers significantly more important thanwhat NeonKnight is saying. Since you will be hitting on 5+ or so instead of 9+ or so.

[Tai Dai Cultist]

Sorry, I will disagree with that.

If you and I both agree that a Skill 0 pilot should be 20PV (actual amount does not matter) above a Skill 4 pilot, does it matter how the value got there?

We're not at all on the same page.  But to address this thought:

If I say, the steps are:

Skill 4 = 0
Skill 3 = +10
Skill 2 = +5
Skill 1 = +3
Skill 0 = +2

Because the to hit rolls get lower and lower % increases

and you say it should be

Skill 4 = 0
Skill 3 = +5
Skill 2 = +5
Skill 1 = +5
Skill 0 = +5

At the end of the day both charts represent the same 20 point increase, but what I am proposing shows where those same PV increases also have the most ingame effect.

My position would actually be it should cost something like:

Skill 4 = 0
Skill 3 = +2
Skill 2 = +3
Skill 1 = +5
Skill 0 = +10

(although to be precise I'm saying skill 4 -> 3 should stay the same as it currently is; it's just the upgrades AFTER 3 that need to get successively more and more expensive)

As illustrated by the example, I was disagreeing with your premise that the 1 -> 0 upgrade was less valuable than the 4 -> 3 upgrade.  I'm saying the exact opposite is true.

[Elmoth]

https://anydice.com/program/20

Put the cursor in the "at least" Data option and you'll find the to hit probabilities for any given difficulty.

The thing is that ypu are almost never hitting on a 4+. You are hitting on a 7+ AT LEAST most of the time: +1 because the guy moves 1+ inches (but less than 9); +2 or more if it moves more than that nd so has a higher TMM. And then add medium range here (+2). There you go. At least difficulty 7+ if you have skill 4. Usually you have a difficulty of 8 or 9 because the guys move +2 or +3 (and you are still at med range). These are all very normal to hit numbers.

Now, add a -1 to hit by dropping your skill 1 rank to 3). You are jumping 13% or 16% in your ability to hit. And above that, youhit MUCH more consistently turn after turn. So the increase in the to hit from 4 to 3 is quite massive. 4 to 2 is also massive. 2 to 0 not so much, but still significant.

We are not looking at the ends of the curve like NeonKnight was saying but at the centre of it most of the time. And here, a drop in 1 skill point is quite massive.

[NeonKnight]

Thank you.

That is what I am trying to say.

Going from a SKILL 4 to a SKILL 3 is a huge jump in the odds of hitting.

Going from a SKILL 1 to a SKILL 0, while, yes it is an increase in the odds to hit, it no where as massive a jump as 4 to 3 is.

And that is why I think 4 to 3 should be worth more than a 1-0.

And lets be honest, players will more often play with skill 3 pilots than skill 0 pilots.

[Tai Dai Cultist]

We are not looking at the ends of the curve like NeonKnight was saying but at the centre of it most of the time. And here, a drop in 1 skill point is quite massive.

Indeed.  0 skill is only at the edge of a bell curve if you're presuming minimal penalties to-hit.  If you stop presuming only minimal penalties (because afterall, 4 skill increases are one-for-one cancels to 4 pips of penalties to-hit) the 0 skill is at the top of the bell curve when you're at +7 (+2 range, +2 cover, +3TMM, or etc).

Where the unit PVs presumed minimal and constant penalties to hit, you can't keep that presumption when evaluating skill increases/cancels to penalties.

[Scotty]

Anecdotal but still worth mentioning: I ran my event at GenCon this year using pilot-assigned skills rather than unit-assigned skills.  That is to say, a skill 3 pilot was +5 PV, Skill 2 was +10, etc, because a pilot could easily be shot down and lose their machine, but hop back into another and I had neither the time nor the inclination to prong a million cards or to figure per-unit cost increases in the extremely limited time window between games.

The effect in game was hardly noticeable. Skill 3 pilots were in high demand, Skill 2 pilots were a hard to justify luxury even for an excellent mech,  and Skill 1 or 0 pilots (there were at least five of those between the two Clusters) were practically written off outside of In-Character concerns even with what was effectively a major discount for those skills because the offense even on a good Clan Assault didn't pay for itself in light of the entire game's lack of relative durability.

Sadlerbw, care to confirm?  The 74-point Skill 0 Khan Malvina Hazen in a Shrine could still only do five damage at a time without overheating (and she definitely didn't hit every time) .  Was it worth it?

[Joel47]

Sounds like it's time to drag out my "Alpha Strike Simulator" script that I wrote last year to run large numbers of "combats." It only handles one-vs-one combat, but it handles actual damage tracking and can be set to adjust the range based on the initiative winner (allowing a faster unit to manipulate the range to its advantage). I think I'm going to want to add features for "cover" (allowing the initiative winner a percentage chance of getting cover vs its opponent), "stand" (the ability for a unit to stand still if it's advisable -- that's going to take some thinking), and "woods" (a percentage chance the round's combat is affected by woods.
Script usage instructions, written by Xotl: http://bg.battletech.com/forums/index.php?topic=53612.msg1314147#msg1314147
That thread also has discussion of the algorithms; any specific discussion of those would be best broken off into their own thread (or PM me) so as not to walk all over this thread with coding discussions.

To start with, we should look at where we're at. While we know the old point system is going away, looking at the strong and weak units might provide insight to our changes. For this first run, I selected a number of small units. I arbitrarily decided on 26 points, as it was convenient.

Code: [Select]

[
  {"name":"Hunchback IIC (Standard)", "points":26, "skill":4, "type":0, "armor":3, "structure":3, "weapons":[4, 4, 0], "move":8, "special":["CASE"]},
  {"name":"Spider SDR-7M", "points":26, "skill":4, "type":0, "armor":2, "structure":3, "weapons":[2, 2, 0], "move":16, "special":["ENE"]},
  {"name":"Charger CGR-1L", "points":26, "skill":4, "type":0, "armor":3, "structure":6, "weapons":[2, 2, 0], "move":10, "special":["ENE"]},
  {"name":"Vulcan VT-6M", "points":26, "skill":3, "type":0, "armor":3, "structure":2, "weapons":[2, 2, 1], "move":12},
  {"name":"Falcon FLC-4Nb", "points":26, "skill":3, "type":0, "armor":3, "structure":3, "weapons":[2, 2, 0], "move":12, "special":["ENE"]},
  {"name":"Commando COM-7B", "points":26, "skill":2, "type":0, "armor":2, "structure":1, "weapons":[3, 2, 0], "move":12, "special":["CASE"]},
  {"name":"Panther PNT-9R (Tanaka)", "points":26, "skill":2, "type":0, "armor":3, "structure":3, "weapons":[2, 2, 0], "move":8},
  {"name":"Hussar HSR-300-D", "points":26, "skill":1, "type":0, "armor":1, "structure":3, "weapons":[1, 1, 0], "move":18, "special":["ENE"]},
  {"name":"Commando COM-4H", "points":26, "skill":1, "type":0, "armor":2, "structure":2, "weapons":[2, 2, 0], "move":12},
  {"name":"Wasp WSP-8T", "points":26, "skill":0, "type":0, "armor":2, "structure":1, "weapons":[1, 1, 1], "move":12}
]

Here's the output:

Attacker \ Defender	Hunchback IIC (Standard) (Skill 4)	Spider SDR-7M (Skill 4)	Charger CGR-1L (Skill 4)	Vulcan VT-6M (Skill 3)	Falcon FLC-4Nb (Skill 3)	Commando COM-7B (Skill 2)	Panther PNT-9R (Tanaka) (Skill 2)	Hussar HSR-300-D (Skill 1)	Commando COM-4H (Skill 1)	Wasp WSP-8T (Skill 0)
Hunchback IIC (Standard) (Skill 4)	-----	Spider SDR-7M: 489/441/70(4.5)	Hunchback IIC (Standard): 500/442/58(4.8)	Vulcan VT-6M: 522/367/111(6.2)	Hunchback IIC (Standard): 522/364/114(3.4)	Hunchback IIC (Standard): 725/175/100(2.0)	Hunchback IIC (Standard): 600/285/115(4.8)	Hunchback IIC (Standard): 807/171/22(3.1)	Hunchback IIC (Standard): 662/247/91(2.0)	Hunchback IIC (Standard): 627/313/60(5.1)
Spider SDR-7M (Skill 4)	Spider SDR-7M: 493/428/79(4.5)	-----	Spider SDR-7M: 578/405/17(8.6)	Vulcan VT-6M: 472/471/57(5.4)	Falcon FLC-4Nb: 502/440/58(5.6)	Spider SDR-7M: 495/396/109(3.8)	Panther PNT-9R (Tanaka): 464/429/107(4.2)	Spider SDR-7M: 578/391/31(5.4)	Commando COM-4H: 573/333/94(3.4)	Spider SDR-7M: 626/292/82(4.1)
Charger CGR-1L (Skill 4)	Charger CGR-1L: 480/449/71(4.8)	Spider SDR-7M: 581/398/21(8.8)	-----	Vulcan VT-6M: 563/409/28(8.1)	Falcon FLC-4Nb: 564/415/21(6.7)	Charger CGR-1L: 558/397/45(4.4)	Charger CGR-1L: 515/445/40(5.4)	Charger CGR-1L: 674/320/6(6.4)	Commando COM-4H: 529/421/50(4.4)	Charger CGR-1L: 572/408/20(6.9)
Vulcan VT-6M (Skill 3)	Vulcan VT-6M: 505/377/118(6.3)	Vulcan VT-6M: 478/445/77(5.4)	Vulcan VT-6M: 540/427/33(8.2)	-----	Falcon FLC-4Nb: 463/426/111(6.1)	Vulcan VT-6M: 715/189/96(4.3)	Vulcan VT-6M: 583/295/122(6.3)	Vulcan VT-6M: 812/160/28(3.9)	Vulcan VT-6M: 533/312/155(4.8)	Vulcan VT-6M: 724/163/113(4.3)
Falcon FLC-4Nb (Skill 3)	Hunchback IIC (Standard): 504/392/104(3.4)	Falcon FLC-4Nb: 494/447/59(5.5)	Falcon FLC-4Nb: 547/421/32(6.6)	Falcon FLC-4Nb: 458/449/93(6.1)	-----	Falcon FLC-4Nb: 596/247/157(4.9)	Panther PNT-9R (Tanaka): 425/420/155(3.7)	Falcon FLC-4Nb: 807/174/19(4.5)	Falcon FLC-4Nb: 469/393/138(4.8)	Falcon FLC-4Nb: 687/233/80(5.2)
Commando COM-7B (Skill 2)	Hunchback IIC (Standard): 736/160/104(2.0)	Spider SDR-7M: 482/402/116(3.7)	Charger CGR-1L: 602/342/56(4.3)	Vulcan VT-6M: 700/188/112(4.3)	Falcon FLC-4Nb: 546/265/189(5.0)	-----	Panther PNT-9R (Tanaka): 632/166/202(2.5)	Commando COM-7B: 644/212/144(2.8)	Commando COM-4H: 350/183/467(4.2)	Wasp WSP-8T: 664/75/261(3.6)
Panther PNT-9R (Tanaka) (Skill 2)	Hunchback IIC (Standard): 616/237/147(4.8)	Panther PNT-9R (Tanaka): 512/389/99(4.0)	Charger CGR-1L: 505/451/44(5.2)	Vulcan VT-6M: 600/276/124(6.3)	Panther PNT-9R (Tanaka): 446/410/144(3.6)	Panther PNT-9R (Tanaka): 580/178/242(2.6)	-----	Panther PNT-9R (Tanaka): 931/59/10(3.4)	Panther PNT-9R (Tanaka): 532/284/184(2.5)	Panther PNT-9R (Tanaka): 482/417/101(5.7)
Hussar HSR-300-D (Skill 1)	Hunchback IIC (Standard): 803/169/28(3.2)	Spider SDR-7M: 546/434/20(5.2)	Charger CGR-1L: 690/294/16(6.8)	Vulcan VT-6M: 779/183/38(4.1)	Falcon FLC-4Nb: 828/148/24(4.2)	Commando COM-7B: 645/195/160(2.8)	Panther PNT-9R (Tanaka): 924/62/14(3.3)	-----	Commando COM-4H: 925/41/34(2.6)	Wasp WSP-8T: 456/453/91(3.7)
Commando COM-4H (Skill 1)	Hunchback IIC (Standard): 704/201/95(1.9)	Commando COM-4H: 565/332/103(3.4)	Commando COM-4H: 508/432/60(4.3)	Vulcan VT-6M: 531/327/142(4.7)	Falcon FLC-4Nb: 458/389/153(4.8)	Commando COM-4H: 379/182/439(4.2)	Panther PNT-9R (Tanaka): 563/265/172(2.5)	Commando COM-4H: 919/49/32(2.6)	-----	Commando COM-4H: 411/289/300(4.2)
Wasp WSP-8T (Skill 0)	Hunchback IIC (Standard): 630/318/52(5.1)	Spider SDR-7M: 640/306/54(4.1)	Charger CGR-1L: 586/393/21(6.9)	Vulcan VT-6M: 694/186/120(4.3)	Falcon FLC-4Nb: 660/249/91(5.3)	Wasp WSP-8T: 672/75/253(3.6)	Panther PNT-9R (Tanaka): 504/411/85(5.6)	Wasp WSP-8T: 467/436/97(3.7)	Commando COM-4H: 393/297/310(4.2)	-----

Conclusions:
1. The Hunchback IIC is strong due to it only having to hit once against most foes.
2. The Commando 4H and Panther Tanaka are good high-skill units.
3. The Hussar is absolute trash, and the Wasp 8T isn't much better. The Wasp might fare well with a different range algorithm -- this one is "fast_unit_causes_slow_approach" which... is delving into algorithm discussion. Read the old thread, or the code; it was considered the most accurate simulation last time.
4. I thought the Charger would dominate due to its ability to soak damage, but it only did well against skilled opponents.
5. The Falcon and the Vulcan are excellent. This adds credence to the theory espoused by others in this thread that Skill 3 is the "sweet spot."

Back later -- going to run more tests over lunch using higher PVs. I really want to know how things work out with heavies & assaults.

[NeonKnight]

Instead of comparing a Hunchback IIC at skill 4 versus many other, why not Hunchback IIC versus IIC at differing Skills?

[KCmasterpiece]

Using damage for scale seems best. Maybe something like:

(Medium or short damage + long damage) x (4 - desired skill)

It will be linear but linear is easy.

[sadlerbw]

per Scotty's comment: Skill wasn't everything because killing stuff didn't really win the scenarios (at least for the side I was playing). Bringing a unit that fit the needs of the mission was worth way more than low-skill pilots because it was really hard to kill your way to victory. That said, the low-skill pilots were definitely too cheap because the one chance I had where I could justify bringing the units they were riding in, I brought every low-skill pilot I could find! That command star was slow, but no one who got within medium range of it lived to regret their decision. Even then, I was hoping the low-skill pilots would look like juicy targets to gang up on so they could soak damage and take pressure off the rest of the force. basically, I attribute the relative lack of importance of low-skill pilots on the force and scenario designs, not the pricing of the skills.

Switching gear, I have made a spreadsheet! I'm trying to use it to figure out how to generalize how beneficial skill increases are. It is taking into account, target TMM's from 0 to 5, Skills from 0 to 7, and the distribution of TMM's across all the units with Alpha Strike stats. [EDIT: Xotl got me REAL numbers!]. Using every unit in the game may not be the appropriate distribution, but I want to start there.

For now I assumed the following:

Def Mod 0/-1: 11%
Def Mod 1: 29%
Def Mod 2: 27.6%
Def Mod 3: 16.5%
Def Mod 4: 7.1%
Def Mod 5/6: 1.9%

With that distribution of TMM's, I summed up the average of the changes in to-hit number for a given TMM at S/M/L and then multiplied by the percentage of units with that TMM, then summed up the amounts of to-hit change contributed by each TMM group. I also assumed that medium range is twice as important as short or long (sort of how we look at damage for calculating PV). Given all that, and averaging the changes at Short, Medium, and Long ranges, I came up with the following rough numbers:

Weighted average improvement in to-hit compared to Skill 4

Skill 0: 39.3%
Skill 1: 32.1%
Skill 2: 22.7%
Skill 3: 11.7%
Skill 4: 0%
Skill 5: -11.4%
Skill 6: -21.4%
Skill 7: -29.5%

If that is somewhere close to correct, then the offensive power of a unit ought to go up by the offensive portion of it's PV multiplied by that percentage. I don't know how much of a units PV is attributable to attack power, but lets just say 60% for the time being. At that ratio, I would expect a skill increase to, on average, raise or lower a unit's PV by this amount (rounded to whole percentage points for simplicity):

Skill 0: 24%
Skill 1: 19%
Skill 2: 14%
Skill 3: 7%
Skill 4: 0%
Skill 5: -7%
Skill 6: -13%
Skill 7: -18%

Just dump those numbers into a table and say "Add the percentage of base PV shown in table X-X, rounding normally." In fact, I'd prefer if ANY values we end up liking could be abstracted out to a table of percentages. That is about as easy as the math can get for figuring out skill increases. Beyond that, maybe add a second column for units with no damage at medium or long, as those numbers change enough to matter. My recommendation for units with no medium or long range damage would be as follows:

Skill 0: 4%
Skill 1: 3%
Skill 2: 3%
Skill 3: 1%
Skill 4: 0%
Skill 5: -2%
Skill 6: -4%
Skill 7: -6%

I thought about making a third column for units with no long range damage, but I'm not sure it made enough difference to be worth it. I did the calculations and other than 0, 1, and 7, the rest of the skills only changed by a couple percentage points.

Assuming OneDrive will cooperate with me, here is a link to the spreadsheet I was messing with:

https://1drv.ms/x/s!ApDHaN8p702HgwC9f5HF-fpgbsrT

It is a little cryptic, but generally, the yellow boxes are variables I'm playing with, and the green stuff are the results I'm looking at.

[Xotl]

Defense Mod Spread
(This includes all sources of defense modifiers, not just TMM)

-1: 0.17 %
 0: 10.8%
 1: 29%
 2: 27.6%
 3: 16.5%
 4: 7.1%
 5: 1.8%
 6: 0.14%

(Also, thanks Joel47 for joining the discussion: that data tool was immensely helpful with the last thread)

[Joel47]

Instead of comparing a Hunchback IIC at skill 4 versus many other, why not Hunchback IIC versus IIC at differing Skills?

Because the higher skill would be expected to win. While there might be some small understanding gleaned from the margin of victory, I think we'll learn more from studying supposedly equal units that differ in some ways. That said, while I'm already choosing units without specials that affect PV but aren't simulated (such as ECM & C3), I'd love some targeted collections of units. For instance, if someone could give me the following lists, I'd appreciate it:

• units with the same speed, differing only in firepower & armor, all either having long range firepower or not
• units with identical firepower differing only in speed

Each list would have differing skills to keep the final PVs equal (+/- 1). Also, please keep lists to 6-8 to keep the table size realistic; beyond that and my response will take the form of a CSV spreadsheet.

(Also, thanks Joel47 for joining the discussion: that data tool was immensely helpful with the last thread)

You're welcome. As before, if there's anything I don't support that you'd like to test, just ask -- I'll figure out if it's feasible.

[sadlerbw]

Defense Mod Spread
(This includes all sources of defense modifiers, not just TMM)

-1: 0.17 %
 0: 10.8%
 1: 29%
 2: 27.6%
 3: 16.5%
 4: 7.1%
 5: 1.8%
 6: 0.14%

(Also, thanks Joel47 for joining the discussion: that data tool was immensely helpful with the last thread)

Thanks! I updated my spreadsheet, although I just rolled -1 into 0 and 6 into 5 because they are so small and the numbers so far at the extremes that it doesn't really matter! Of course, that all adds up to 93.1% so I'm debating whether or not to 'scale' it so that the numbers add up to 100%.

EDIT: With the REAL distribution, the final outcomes in the spreadsheet only change by one percentage point, or don't change at all thanks to rounding. I'll still go edit my other post with the new numbers though.

[NeonKnight]

Because the higher skill would be expected to win. While there might be some small understanding gleaned from the margin of victory, I think we'll learn more from studying supposedly equal units that differ in some ways. That said, while I'm already choosing units without specials that affect PV but aren't simulated (such as ECM & C3), I'd love some targeted collections of units. For instance, if someone could give me the following lists, I'd appreciate it:

• units with the same speed, differing only in firepower & armor, all either having long range firepower or not
• units with identical firepower differing only in speed

Each list would have differing skills to keep the final PVs equal (+/- 1). Also, please keep lists to 6-8 to keep the table size realistic; beyond that and my response will take the form of a CSV spreadsheet.You're welcome. As before, if there's anything I don't support that you'd like to test, just ask -- I'll figure out if it's feasible.

Yes...and that is what I want to see.

How much of a difference is there in a Skill 4 vs Skill 3 matchup as opposed to a Skill 1 vs Skill 0. I predict the # of victories or ratio to be much closer and thus negligible on the Skill 1vs Skill 0 while skill 4 vs skill 3 to show a much larger margin of victories to the Skill 3 unit.

because now we are not so much looking at PV values for TMMs, but an All Things being equal what value is SKILL in  match-up.

Or, that is how I thought this was supposed to go.

[Xotl]

I've linked the same spreadsheet used in the main PV testing in the first post, for those who want to play around that way.

[nckestrel]

Or, that is how I thought this was supposed to go.

The end result should be PV 100 = PV 100, whether it comes from skill or number of units or quality of units.
How to get there is negotiable :).

[NeonKnight]

The end result should be PV 100 = PV 100, whether it comes from skill or number of units or quality of units.
How to get there is negotiable :).

True, but sometimes a Unit is just better ;)