[EX3] Battle group adjustments

[Heavy Arms]

That's kind of hard to interpret because it's code and there's plenty missing from it (not that I'm looking for the full code or anything, but I was looking more for how you think the order of operations work rather than the specific coding first).

Like, I can't tell if that accounts for the Size boost to accuracy from what I read, the BG should be have a ~7% advantage in hit rate (and from hit be out performing in raw damage as well) that doesn't seem to be there. The relatively small sample size of the examples provided seem like the system is also rolling a tad high.

I'm still not sure what, "half of the rounds" means.

Why is the BG's initiative always 3? BGs still roll Join Battle and have an initiative score that determines when they go in the round. It seems like you've denied them this, (which should be a straight 50/50 division with the same roll), and made it much harder for BGs to get in two hits in a row by going second one round, and first the next, (since this isn't going to happen until they can get the individual down to 2 initiative) which can swing things pretty fast.

[Aaron S]

I was afraid of obscuring any of my bad understanding through too high-level of a description of what the simulation does. It's great you caught the mistake about battle group Initiative. Thank you very much for taking a look. Presumably I read the section title "Inert Initiative" and translated it as meaning they have only base Initiative.

Fixing Initiative gives quite a few more draw scenarios but doesn’t meaningfully lower the individual win rates. I can completely throw away the Initiative modification I was suggesting, because it effectively did what the rules actually say.

The bolded fields represent what I expect would be the default drill for a battle group composed of those types of characters.

Militia vs. Size 1        Med. Infantry vs. Size 1   Elite Troops vs. Size 1   Buck-Ogre vs. Size 1                  
#   Won%  Draw%  Drill    #   Won%  Draw%   Drill    #   Won%  Draw%  Drill    #   Won%  Draw%  Drill
1   45.8    7.6  Poor     1   41.8    2.1   Poor     1   65.5    2.8  Poor     1   29.4    1.0  Poor
2   77.7   11.8  Poor     2   76.5    5.0   Poor     2   91.6    2.6  Poor     2   77.4    2.6  Poor
3   96.7    2.5  Poor     3   95.7    1.7   Poor     3   99.2    0.4  Poor     3   98.1    0.5  Poor
[B]4   98.7    1.2  Poor[/B]     4   99.0    0.7   Poor     4   99.9    0.1  Poor     [B]4   99.8    0.1  Poor[/B]
[B]5   99.8    0.2  Poor[/B]     5   99.8    0.2   Poor     5  100.0    0.0  Poor     [B]5  100.0    0.0  Poor[/B]
[B]6   99.9    0.1  Poor[/B]     6   99.9    0.1   Poor     6  100.0    0.0  Poor     [B]6  100.0    0.0  Poor[/B]
1   22.5    3.5  Avg.     1   13.8    0.6   Avg.     1   36.0    2.2  Avg.     1    3.7    0.2  Avg.
2   56.1   12.5  Avg.     2   47.0    4.4   Avg.     2   75.8    4.3  Avg.     2   35.1    2.2  Avg.
3   88.9    5.7  Avg.     3   82.4    3.3   Avg.     3   96.1    1.3  Avg.     3   83.2    1.8  Avg.
4   96.0    3.2  Avg.     [B]4   94.6    2.2   Avg.[/B]     4   99.3    0.4  Avg.     4   97.5    0.7  Avg.
5   99.0    0.9  Avg.     [B]5   98.8    0.7   Avg.[/B]     5   99.9    0.1  Avg.     5   99.8    0.1  Avg.
6   99.7    0.3  Avg.     [B]6   99.7    0.3   Avg.[/B]     6  100.0    0.0  Avg.     6  100.0    0.0  Avg.
1    6.8    0.8  Elite    1    2.3    0.1   Elite    1   11.9    0.8  Elite    1    0.1    0.0  Elite
2   27.7    7.3  Elite    2   16.1    1.8   Elite    2   45.6    4.4  Elite    2    4.8    0.4  Elite
3   66.4    8.5  Elite    3   50.9    3.3   Elite    3   83.7    2.9  Elite    3   38.9    1.8  Elite
4   86.1    7.5  Elite    4   79.1    4.2   Elite    [B]4   95.8    1.6  Elite[/B]    4   80.5    2.5  Elite
5   96.1    2.9  Elite    5   93.8    2.3   Elite    [B]5   99.3    0.4  Elite[/B]    5   97.5    0.6  Elite
6   98.3    1.6  Elite    6   98.1    1.2   Elite    [B]6   99.8    0.1  Elite[/B]    6   99.8    0.1  Elite

"Half of the rounds" was my comment from eye-balling a few lines of the detailed fight log. One individual Elite Troop against a battle group actually does land the expected 65% of attacks. My hand-waving summary bears a similarity, in its unrepresentative nature, to the examples I picked. I chose ones looking for short or long fights. I think that selection criterion is why you notice dice rolls at the margins. I just verified that the rolling algorithm for a million throws of 11 dice gives the expected outcomes from the Troll probability calculator.

Is this about the level of description of the algorithm you were looking for, Heavy Arms?
An attack turn resolves this way.
- eliminate any onslaught penalty, because this is a new turn for the character
- attack pool = weapon attack – wound penalty + size
- roll attack pool
- use these same attack successes against each opponent’s Defense
- if (successes >= Defense), hit
- increase target’s onslaught penalty

A hit works in this way.
- raw damage pool = weapon damage + threshold successes + size – target’s soak
- if raw damage pool < weapon overwhelming rating, set it equal
- roll raw damage pool
- apply successes as damage to target
- gain Initiative if an individual

Damage applies this way.
- apply damage to Initiative first, if an individual
- apply remaining damage to health boxes
- adjust wound penalty based on health boxes, if an individual
- note if entered Initiative Crash
- reward Initiative bonus to the attacker

[Heavy Arms]

It's great you caught the mistake about battle group Initiative.

No problem. Even if we don't agree on changing the nature of BGs, I figure it's best you get the most solid simulation to best handle your ideas of house rules.

Fixing Initiative gives quite a few more draw scenarios but doesn’t meaningfully lower the individual win rates.

I had a suspicion this would be the case. The order reversal is a lot more important for winning in individual vs individual fights because individuals want to optimize decisive attacks via crashing the enemy and then attacking on the next turn before they can go (since this is usually the best way to minimize their Defense to compensate for losing accuracy).

In the simulation you're using, it mostly increases the ability for BGs to reduce the individual opponents' initiative twice in a row, meaning they have better odds of doing enough health damage to get a win condition but not incapacitate to the point of avoiding enough damage to be routed in response for the draw.

I am curious if this changes the math your simulation was generating regarding the impact of commanders. The extra dice on attacks seem like they'd increase the impact here. Though I'm not sure you want to complicate the simulation even more with a few ranges of commanders, the Drill bonus to commands, and commanders using the Rally to Numbers option under certain conditions (like automate it if the BG's magnitude drops to 50% or lower).

I just verified that the rolling algorithm for a million throws of 11 dice gives the expected outcomes from the Troll probability calculator.

Cool. Not surprised, just wanted to make sure.

Is this about the level of description of the algorithm you were looking for, Heavy Arms?

Yeah this is what I was going for, thank you. Looking at the code, and this, I don't see any of those silly little bugs that can creep into this sort of thing (I've done it more times than I can count where something like the attack pool in the code is actually "weapon attack - ( wound penalty + size)" instead of, "weapon attack - wound penalty + size"). There's a few things missing here, but they're in the code as far as I can tell (like Defense being lowered by wound penalties).

-----

I think this really leaves the sticking point for what you want out of BGs vs. how they're designed to perform is the Onslaught penalties. It's the only part of the mechanics that is so strongly amplified by 4-6 individuals vs 1 BG.

The problem is that onslaught immunity is really potent and doesn't slot well into the way combat works on the wider scale when you start getting Exalts and Charms involved.

A different option though is to go the other direction: individuals engaged in close combat with BGs don't reset onslaught penalties until they've disengaged for a turn. This makes sense in a descriptive fashion ("you're mobbed by a BG so you can't get the balance to re-establish proper defensive posture"), and doesn't interact as poorly with Charm tech (Exalts that want to be in close combat are going to want to have onslaught negation Charms anyway, and the offensive Charms that key off of targets with high onslaught penalties on them being boosted by BGs doesn't feel like it will be too potent).

I mean, it doesn't take much to see that the Defense bonus from Drill is a big deal in win%. Widening the Defense gap as the fight drags on should make facing a BG more scary.

[BlueWinds]

I haven't run any rigorous tests, but my experience matches your numbers - battle groups are far weaker than I'd hope to see, especially at larger sizes. The house rules I applied were a bit stronger than yours - I doubled all Size related bonuses. So in addition to increasing their attack pool, my rules also increased their damage, health track and soak. I was more interested in making an army a serious threat, rather than balancing size one battle groups.

It's reassuring to hear that I'm not alone in desiring that numbers matter more. "Armies are important" is a core premise of the Exalted setting that hasn't received much love in the mechanics these last two editions (and I never played 1e, so no comment on that). ^^;;

[Heavy Arms]

"Armies are important" is not a core premise of Exalted without a lot of caveats. Large numbers of troops especially are not, and have never been, presented as the trump card in Exalted. If a Dawn has to struggle significantly to take on a Size 5 group of fresh peasant conscripts something is wrong. If the system doesn't let you recreate the Battle of Thermopylae something is wrong. Etc. The Realm isn't the dominant military force in Creation because of it's ability to field as many legions as it can; it's because those legions are lead by Dragonblooded commanders, to support Dragonblooded heroes fighting up close and personal with enemies, and to compliment Dragonblooded sorcerers bringing their own small units of supernatural fighters to add to the potency of the over military might in question.

[HalfTangible]

Individuals won too handily. For an extreme outlier, note one individual Elite Troop won almost two-thirds of matches against a whole battle group.

That's because, under most circumstances, the individual elite troop would be an Exalt, ninja or some other great hero. In Creation, the great hero theory of history is very much in effect: individuals impose the greatest change on the world, not nations or armies. (see Heavy Arms on the strength of the Realm's military)

That said, keep in mind a few things:
1) For battle groups initiative/health also represents morale: A single soldier fighting his way through a size two force and coming out largely unscathed is going to be MUCH more demoralizing than losing the same number of troops to an army of the same size.
2) A size 2 battle group could easily split into two size 1 battle groups with other traits remaining the same. Did you try this? Because having an army of that size remain as one unit strikes me as something similar to sending your mooks to fight the hero one at a time in [insert video game franchise/favorite sentai here]. And if I'm reading these rules right, by doing this you increase your withering dice from 4 total to 6 total

[Heavy Arms]

The books heavily advise against splitting BGs up just to gain a mechanical advantage by having two BGs instead of one. Multiple BGs should either be different types of troops (So you might have a Size 5 infantry BG, a Size 3 archery BG, and a Size 3 light cavalry BG), or be because you're dealing with enough people that even Size 5 can't handle them.

[HalfTangible]

The books heavily advise against splitting BGs up just to gain a mechanical advantage by having two BGs instead of one. Multiple BGs should either be different types of troops (So you might have a Size 5 infantry BG, a Size 3 archery BG, and a Size 3 light cavalry BG), or be because you're dealing with enough people that even Size 5 can't handle them.

That strikes me as odd. It's not uncommon in war to split off a group of your troops to try and flank the enemy, or hold some back in reserve, or similar actions that require the commander split his force.

I get why you wouldn't do so as a matter of course (since BGs are meant to speed things up) but if you specifically want to make the BGs more threatening, why not?

[Heavy Arms]

Well, to the OP's stated goals, the "BGs are meant to speed things up," is still an important consideration.

And really, if you want to be threatening, diversity of troops is better anyway. If you're going to track multiple BGs as antagonists, you might as well get the much stronger benefit of mixed troops. Have an infantry BG to tie up the PCs and their allies, and a cavalry group to charge in. Etc. It's way more of a threat than two identical BGs.

[TheCountAlucard]

That strikes me as odd. It's not uncommon in war to split off a group of your troops to try and flank the enemy, or hold some back in reserve, or similar actions that require the commander split his force.

Those sorts of actions are covered by strategems, using the War Ability, not by playing hob with the action economy.

[HalfTangible]

Well, to the OP's stated goals, the "BGs are meant to speed things up," is still an important consideration.

And really, if you want to be threatening, diversity of troops is better anyway. If you're going to track multiple BGs as antagonists, you might as well get the much stronger benefit of mixed troops. Have an infantry BG to tie up the PCs and their allies, and a cavalry group to charge in. Etc. It's way more of a threat than two identical BGs.

.... and it still does? I'm not suggesting stating out every individual soldier.

Yes, that would be a way to introduce more threat. But the scenario above was for a size 2 battle group against a single troop of the same type.

TheCountAlucard : Implying that trying to squeeze as much mechanical benefit out of as tiny a space as possible is a bad thing :P

[Heavy Arms]

.... and it still does? I'm not suggesting stating out every individual soldier.

For the specifics of one Size 2 BG to two Size 1 BGs? It's not a big deal, but following that logic for larger BGs gets bad really fast. As the book itself points out a single Size 5 BG could be ten Size 3 BGs. At which point you're still getting into the overworked ST issue. The problem even with wanting to make BGs more powerful is that this also scales with workload. The more power you get out of this, the more work you put on the ST. Two Size 2 to one Size 1 isn't a big deal, but it isn't a big boost either. The mentioned one Size 5 to ten Size 3s is a huge amount to track for a system designed to reduce stuff to track, and an insanely bigger jump in threat.

Yes, that would be a way to introduce more threat. But the scenario above was for a size 2 battle group against a single troop of the same type.

The simulation the OP is a simulation of a highly specific scenario. I'm pretty sure they recognize that, and are capable of seeing how multiple BGs of different Sizes stack up against the same number of individual combatants.

My point is going beyond the simulation, you get more out of multiple BGs if they're not identical than if they are. So you might as well used mixed troops, avoiding the against-the-spirit-of-the-concept of splitting a BG into lots of smaller BGs, and getting a more potent tactical threat out of it at the same time.

TheCountAlucard : Implying that trying to squeeze as much mechanical benefit out of as tiny a space as possible is a bad thing :P

No, just pointing out the explicit thing the book says:

"Otherwise,don’t split battle groups into smaller units for greater offensive power." Ex3, p. 211

As well, pointing out things like dividing your troops to attack from different angles is something the Pincer Attack Stratagem is for.

[Aaron S]

Recently I have tried a variety of tweaks to even out the win rates for my given scenario. All failed to be universal. I’m still looking for a solution to achieve what I want. (Those goals have been refined a bit from my initial post, but I won’t state them here.) I need to spend time learning why three-combatant scenarios range from 83% to 39% wins against an elite Drill battle group.

That’s future work.

Following topics raised in some responses, I pondered game balance for enemies, allies, and Merits. I don’t think I hit on anything new here, but I thought I’d share my thoughts anyway. Balance in the rules seems to be stem from world and story more than combat aspects. Something costing five dots seems like the best thing of that type that powerful rulers, sorcerers, or Dynasts have across Creation. A few better instances of such things exist, but they will become the focus of a campaign rather than the trappings of a starting character. That’s the context I read and am working with. Four different Merits all directly provide bodies assisting a character—Allies, Command, Followers, and Retainers. Each brings its own in-world consideration and complication. Two dots in Command gives access to a Size 3 battle group of Medium Infantry. I like using logistics, upkeep, motivation, and coordination to portray the consequences of keeping even just 125 soldiers at the ready. If a player finds commanding a battle group to be an interesting part of the game, I want to support that, challenge that, and also ensure the battle group has the right ‘heft’.

Maybe that ‘heft’ arises in play already. Part of what I understand people who have run battle groups to be saying is that they like the feel. The feel comes from their altered mechanics, which conveys the sense of fighting a massed unit, not a single creature. Battle groups attack every creature on their fronts and handle damage differently. Even strong characters are threatened if other members of their Circle are at risk. I shouldn’t underestimate the psychological effect for players when facing one.

While I’m still stuck on arriving at a comparable force in arms, reminders of the unit size leads me focus on three to six individuals facing one battle group of Size 1. Actual play mostly has a number of protagonists in that range, and it matches the description for what Size 1 means. I’m glad that the basic effect from changing numbers or Drill remains predictable and intuitive. E.g., adding individuals always increases that team’s win rate and lowers the time needed to win and, not coincidentally, the number of casualties. Increasing a battle group’s Drill always makes it harder to defeat. I had wondered if more individuals would give a battle group more opportunities to get in a lucky hit and win, at least more than the extra person helps the individual defeat the battle group. That appears never to be the case for a team of 3-6 Militia, Medium Infantry, Elite Troops, or Buck-Ogres.

The degree of change is hard to predict. Increasing Drill tends to shrink a 95% win rate to 80%, and an 80% win rate to 50%. 9-13% of fights have at least one individual death when the individuals win 80%. This drops to 2-4% of fights when individuals win 95%. Another way of thinking about that point is in terms of the victory condition. The battle group wins the simulation if at least half of the individuals are heavily injured or any are dead. Given that, about 60% of battle group wins happened because of at least one individual dying. (This observation ties back into the gross limitations of my simulation while also spurring thoughts of more tweaks. I need to avoid that path for now.)

I am interested learning about how larger battle groups perform. I want to check different battle group sizes against each other, strong individuals against large but weak battle groups, and the effects of commanders. My goal would be that larger battle groups win in some ‘reasonable way’, and I don’t know what that means yet. I personally want not to have to split a battle group into multiple for it to play out as effectively as I think it should. Splitting for different forces, such as infantry, archers, and cavalry, makes sense. Maybe I can gather these data before the system learning I mentioned at the start.

Additional details to mention for critique: I kept the Buck-Ogre battle group simulated as Might 0. Descriptions from the Bride of Ahlat’s Blessing of Ahlat (4. 498) and the Cataphract’s Marching Fever Dream Nightmare (p. 535) left me thinking that some extra infusion of energy should be around for it to be Might 1. I also don’t simulate use of the Buck-Ogre’s Dual Strikes special attack, or anyone’s special attacks, really, but that one stands out. For other battle group types, I want to use the special chopping, piercing, or slashing attacks enabled by weapon tags, but I don't have a good rationale for when. No combatant uses those techniques at this point.

[Heavy Arms]

One of the reasons the change in percentages are hard to predict is because of the nature of a dice-pool system of rolling.

There isn't a flat change from Drill because the impact of increased Drill (notably the increased Defense) isn't flat. Individuals winning at 95% are likely hitting successes need to overcome the increased Defense from Drill already, so there's a minor change. Individuals winning at an 80% rate is likely hitting at a much lower average value, so the increased Defense from Drill is going to have a much more drastic effect.

A simple way to look at it, is that if your average successes is 6, and you're facing a base Defense of 4, Drill going up isn't going to be a huge difference, because you're still hitting on average. 12 dice vs a boosted Defense of 6 still hits 57% of the time.

If your average successes are only 5 vs. a base Defense of 4, the impact of increasing Defense is more pronounced (esp. Elite Drill's +2). 10 dice vs. a boosted Defense of 6 only hits 39% of the time.

Though for your simulation, the impact is probably more on the onslaught side of things.

[Isator Levi]

The Realm isn't the dominant military force in Creation because of it's ability to field as many legions as it can; it's because those legions are lead by Dragonblooded commanders, to support Dragonblooded heroes fighting up close and personal with enemies, and to compliment Dragonblooded sorcerers bringing their own small units of supernatural fighters to add to the potency of the over military might in question.

Poor Dragon Blooded Sun Tzu, left out in the cold...