Realistic Combat that's Simple(ish)

[Theory of Games]

As someone that used to run GURPS I entirely disagree.

I mean, in say 1999 I would have agreed with you. Although, I was at that time using the GULLIVER house rules out of frustration with the lack of realism of GURPS. I also hadn't at the time realized that the GURPS skill system was the most badly designed system in all of gaming. I was at the time bamboozled by all of exactly what I'm talking about here, with all the complexity of its process simulation where if you read it you go, "Of course. That makes sense." Because some skills are harder than others, and if you train on one skill then that ought to make you better at related ones, and so on and so forth. I too was equating realism with "Has a subsystem to cover the mental steps that occur to you."

But a realistic system wouldn't be like that. It would need to make accurate predictions in the way that F=ma and V=IR and so forth make accurate predictions. And the trouble is that most "realistic" systems cover up that all the numbers are arbitrarily chosen behind all the "We thought of that" steps. All those steps bury the math, increase the complexity of analysis, soothe your suspension of disbelief, but don't necessarily make any more accurate predictions.

You keep writing "realistic systems" when there are none. No system, even the most advanced computer models, can predict the future. All they can do is project possibilities based on data from past events. I notice that very often people get turned around when discussing game theory and confuse what is abstract with what is concrete - which leads to further confusion surrounding simulation vs. emulation. And I get that you don't enjoy GURPS' system. Everything isn't for everybody.

My point here is, again, while we can't get realism from a ttrpg, we can get a fair emulation of reality using systems like GURPS or HERO because those systems give us modifiers for exact range, target velocity, target size, recoil, the effects of different damage types on different types of material, ect. Most ttrpgs don't get into the weeds like that. Do these modifiers make GURPS a perfect model of reality? Of course not, but, do the modifiers provide a better emulation (not simulation) than most ttrpgs?

 

[Mannahnin]

Body armor exists for the inevitability of getting hit, stabbed, shot, whatever. More importantly, it is designed with expectations in mind for what "getting hit" means. That is why armor has changed so much over time.

So if a fantasy world, armor is going to be a complex choice. If you are a knight or mercenary who is going to primarily fight other knights or mercenaries, you are going to go with some standard choices depending on the "era." But what if you are fighting wizards who regularly throw lightning or dragons that breath fire, or squamous monstrosities dripping with acid? From a game design standpoint, "realistic" damage reduction armor in a fantasy world is going to be a complex mess: think AD&Ds AC vs Armor Type dialed up to 11.

This reminds me of the weirdly baroque initiative for OD&D article Gygax had in The Strategic Review where he goes into speed of movement in plate armor vs. unarmored Conan-types, and makes it so that the latter is better at closing the distance to and curb-stomping an M-U trying to Lightning Bolt you.

 

[Celebrim]

You keep writing "realistic systems" when there are none.

Well, that was my first proposition, so we don't have a lot to disagree on.

My point here is, again, while we can't get realism from a ttrpg, we can get a fair emulation of reality using systems like GURPS or HERO because those systems give us modifiers for exact range, target velocity, target size, recoil, the effects of different damage types on different types of material, ect. Most ttrpgs don't get into the weeds like that.

My point is does "getting into the weeds" actually model reality in any way given that the modifiers are typically arbitrarily chosen by the designer and not actually tested or drawn from any real experience or data. We might can say, "Yeah, it's reasonable that longer ranges mean being harder to hit" but we can't really say -3 to hit on the 3D6 bell curve produces the result we would observe in reality. There is no testing to suggest that that 3D6 range of values represents the fine grain we need. Maybe we need 5d6 so that -1 or -2 isn't as gross of a penalty. Why was 3D6 chosen? Because it was convenient, not because it reflected a good approximation of observations.

What I'm trying to say is that taking in all these factors - getting "into the weeds" - soothes your sense of disbelief because it checks off the boxes that you feel are intuitive as the factors that would go into some sort of solution, but that focus on the process kind of obscures the answers its throwing out. It makes it less intuitive where the answer you end up with came from, or what range of answers and with what probabilities you were actually producing. The outcome gets hidden behind the process, and it's the outcome that determines realism, not the process. We're modelling something but plugging all these variables into the process, but it turns out that when we get to the outcome not only by that point do we not know whether the answer is realistic, we know longer care. The process has convinced us that this is close enough because well it was detailed and plugged in reasonable variables. Whether we crunched those reasonable variables into reasonable answers is a wholly different matter.

Pheonix Command is my favorite example here because it's so pretentious about this process and all the variables and factors that go into it. But does it really produce anything like the right number of casualties per bullet fired or the right range of battlefield injuries per bullet that struck the target? Most players don't actually even concern themselves with that. It feels right. It feels like it's taking the right things into consideration. So you accept the answer largely without context. And that's what it's actually producing. Satisfaction rather than correctness.

If you were to start deriving tables for common situations so that you could map a single random D% to all the possible outcomes of all the Pheonix Command results for that situation, I think a lot of weirdness would actually come out and the recognition that the percentages assigned to the likelihood of each outcome first not only didn't match the realistic percentages assignable to each possible outcome, but that the entire complex process for that situation could be approximated by a simpler process (the single die mapping to the final table being a case in point). Ultimately, the question is, "Is Pheonix command actually more realistic than D&D with very large ranges of possible damage for each weapon hit?" is not one I find easy to answer (or that anyone really wants to answer). Like if behind the scenes I was using some simpler system like D20 + Bonus >= DC and then just rolling 1d45 for damage against targets with 10 hit points, and randomly assigning some hit location weighted to the severity of the damage, how long would it take to penetrate that black box and go "wait, those results aren't realistic!"?

 

[Mannahnin]

By realistic, I primarily mean differentiating what D&D calls AC into its two components: dodge/parry and damage reduction.

Or to put it another way, a more realistic version of D&D would have to-hit vs. an opponent's ability to dodge, parry, or otherwise evade being actually hit, and also damage reduction as the result of armor, toughness of skin/hide, and magic.

May be tangential and orthogonal to what you want, but I'm intrigued by the concept of achieving greater realism through greater abstraction, as opposed to greater granularity:

Injury and the Abstract Combat Round

Let me give a natural description of how combat works in D&D. Each person gets to make (at least one) to hit roll against an enemy per rou...

maziriansgarden.blogspot.com

 

[Mannahnin]

My point is does "getting into the weeds" actually model reality in any way given that the modifiers are typically arbitrarily chosen by the designer and not actually tested or drawn from any real experience or data. We might can say, "Yeah, it's reasonable that longer ranges mean being harder to hit" but we can't really say -3 to hit on the 3D6 bell curve produces the result we would observe in reality. There is no testing to suggest that that 3D6 range of values represents the fine grain we need. Maybe we need 5d6 so that -1 or -2 isn't as gross of a penalty. Why was 3D6 chosen? Because it was convenient, not because it reflected a good approximation of observations.

Absolutely.

On a related note, did you ever read Delta's series of articles on archery in D&D, where he tries to apply some good real world math and stats?

Basic D&D: On Archery

Math, history, and design of old-school D&D.

deltasdnd.blogspot.com

 

[aramis erak]

You keep writing "realistic systems" when there are none.

Yes, there are. There are systems where the probabilities are close replicas of the observed odds of outcomes from real life. They tend not to be overly detailed, tho'.

Some of the computer ones are so good that they are used in forensic investigations. And in training. Microsoft Flight Simulator, when used with a flight stick and pedals, is good enough to be used for simulated flight hours towards one's pilot's license. My CFIs had me use it for one hour of training.

No system, even the most advanced computer models, can predict the future. All they can do is project possibilities based on data from past events.

Realism is when those match the probabilities experienced in real life. It's not that it predicts the future, but that it matches the probabilities. But if the probabilities match, the useful predictions can be made... Orbital mechanics are pretty well simulated by NASA, ESA, Роскосмос, JAXA, and ISRO... sufficiently so as to be able to put space probes into orbits around planets to 40 AU (≅3.8e9 miles)... Pluto...

Firefight simulators are used in training US soldiers. Aircraft simulators have been training pilots since WW II. Tabletop wargames have been and still are training US, UK, and other Armies' and Navies' officers in tactics, since the 1890's.

Chess, Go, Shogi, Xiangqi have been used in training officers in the Orient for centuries; they're not even high fidelity tools, but they teach officers to think about how to win, how to use diverse strategies... The IJN prior to world war II actually used a simulation to decide if the Pearl Harbor attack was worthwhile... it was... they didn't predict the weather... CVBG Enterprise was delayed by weather, so escaped the massacre. If the IJN had managed to take out the battlegroup, it would have made for further raids to keep the battleships out of action. CVBG Enterprise was, for a while, the only operating US Capital Ship in the Western Pacific... and she was able to make a huge difference. The biggest was warding off IJN carriers so that they couldn't stop the repairs on most of Battleship Row...

 

[Thomas Shey]

May be tangential and orthogonal to what you want, but I'm intrigued by the concept of achieving greater realism through greater abstraction, as opposed to greater granularity:

Injury and the Abstract Combat Round

Let me give a natural description of how combat works in D&D. Each person gets to make (at least one) to hit roll against an enemy per rou...

maziriansgarden.blogspot.com

The problem with that is that it inevitably produces one of two results: 1. Damage becomes completely disconnected from any effect, its just a counter until people fall over, or (this is probably the more benign of the two but to me that's damning with faint praise) 2. It offloads all the handling on the humans at hand, including, quite likely, people who have no idea how to make the process feel authentic and/or disagree about what does that.

Something like old RuneQuest hit locations was not perfect, but it told you a number of things via resolution, and often you didn't really feel a need for something beyond what it told you; the little bit of extra was mostly going to be color anyway so you could supply it or not to suit yourself.

 

[Celebrim]

Absolutely.

On a related note, did you ever read Delta's series of articles on archery in D&D, where he tries to apply some good real world math and stats?

Basic D&D: On Archery

Math, history, and design of old-school D&D.

deltasdnd.blogspot.com

No I hadn't, but his conclusions are really close to the numbers I'm using in my 3e homebrew, albeit I have more range increments. Of course, the real ranged increments probably vary rather than are linear increments, but I'm not actually trying to be realistic just "casually realistic" in as much as the results are in line with casual expectations. As such, I don't deal with trying to figure out from the data how much velocity a longbow shaft loses over the course of its flight and corresponding losses of penetration (which in D&D would affect both 'to hit' versus AC and expected damage). Things like that would probably require me to renormalize damage to deal with the actual shearing force of a weapon and whether it could cut cloth and skin, and so on and so forth. And that's before we get into the problem that "damage" effects would need to be dealt with as a combination of pushing force (does it knock you down), pain (does it stimulate your nerves), and tissue destruction (how many arteries/capillaries ect. are severed) since the real killing things are shock and blood loss and infection. We'd have to actually deal with the real trauma response of bodies. Whose got time for that?

 

[Theory of Games]

Well, that was my first proposition, so we don't have a lot to disagree on.



My point is does "getting into the weeds" actually model reality in any way given that the modifiers are typically arbitrarily chosen by the designer and not actually tested or drawn from any real experience or data. We might can say, "Yeah, it's reasonable that longer ranges mean being harder to hit" but we can't really say -3 to hit on the 3D6 bell curve produces the result we would observe in reality. There is no testing to suggest that that 3D6 range of values represents the fine grain we need. Maybe we need 5d6 so that -1 or -2 isn't as gross of a penalty. Why was 3D6 chosen? Because it was convenient, not because it reflected a good approximation of observations.

I mean. I look at the modifiers and all I see is abstract fluffiness. They aren't simulating reality, they're emulating it - that is, they're making up numbers and saying "hey these numbers = the target is X yards away and it's raining and your PC's blasted on Novacoke so there's a grouping of penalties". It doesn't make sense but it's all made up so that's fine. You're talking about "reality in ttrpgs" requiring real data and testing and I'm like "Hunh? It 'aint that serious. It's a game, friend." And again even if we wanted to we couldn't simulate reality with a ttrpgs - it's far too complex. GURPS is fun as emulation but it can't mirror The REAL. No game can.

What I'm trying to say is that taking in all these factors - getting "into the weeds" - soothes your sense of disbelief because it checks off the boxes that you feel are intuitive as the factors that would go into some sort of solution, but that focus on the process kind of obscures the answers its throwing out. It makes it less intuitive where the answer you end up with came from, or what range of answers and with what probabilities you were actually producing. The outcome gets hidden behind the process, and it's the outcome that determines realism, not the process. We're modelling something but plugging all these variables into the process, but it turns out that when we get to the outcome not only by that point do we not know whether the answer is realistic, we know longer care. The process has convinced us that this is close enough because well it was detailed and plugged in reasonable variables. Whether we crunched those reasonable variables into reasonable answers is a wholly different matter.

You keep writing that: "Suspension of disbelief". Do you get that from ttrpgs? I never have, but then I can't escape the reality that I'm playing a game. And no, I like GURPS because it has a lot of rules, which allows me to shave off the ones I don't want and keep the rest. But I wouldn't say that makes the system more intuitive - actually IME more rules = less intuitive. Again, you keep bringing up "realism" - there's no "real" in these games. It's funny tho: the way you write about "the process" reminds me of how stage magic works.

Pheonix Command is my favorite example here because it's so pretentious about this process and all the variables and factors that go into it. But does it really produce anything like the right number of casualties per bullet fired or the right range of battlefield injuries per bullet that struck the target? Most players don't actually even concern themselves with that. It feels right. It feels like it's taking the right things into consideration. So you accept the answer largely without context. And that's what it's actually producing. Satisfaction rather than correctness.

I mean we're accepting the results the system kicks out with ANY system, not just the "complex" ones. Could be tic-tac-toe or checkers or GO or Daggerheart - they all have play processes. It's all the same ish. What matters is did we have fun doing it? Like you wrote - "Satisfaction", right?

If you were to start deriving tables for common situations so that you could map a single random D% to all the possible outcomes of all the Pheonix Command results for that situation, I think a lot of weirdness would actually come out and the recognition that the percentages assigned to the likelihood of each outcome first not only didn't match the realistic percentages assignable to each possible outcome, but that the entire complex process for that situation could be approximated by a simpler process (the single die mapping to the final table being a case in point). Ultimately, the question is, "Is Pheonix command actually more realistic than D&D with very large ranges of possible damage for each weapon hit?" is not one I find easy to answer (or that anyone really wants to answer). Like if behind the scenes I was using some simpler system like D20 + Bonus >= DC and then just rolling 1d45 for damage against targets with 10 hit points, and randomly assigning some hit location weighted to the severity of the damage, how long would it take to penetrate that black box and go "wait, those results aren't realistic!"?

Are you spelling Phoenix Command wrong on purpose? I just noticed. Anyway, the best marksmen in the world can't shoot ten perfect bullseyes in a row on Championship courses. Why? Because there's these infinitesimal factors that play into ballistics. Things that even the best instructors and seasoned snipers can't get perfect every time they fire a weapon. It's just too complex and that's why ttrpgs can't simulate it. I mean as creative as the mind is we can only do so much with these games. And I'm fine with that because IME it's fun reading and playing games that try to be "realistic".

 

[Theory of Games]

Yes, there are. There are systems where the probabilities are close replicas of the observed odds of outcomes from real life. They tend not to be overly detailed, tho'.

Name one. Also, "odds" aren't "reality", they're statistical probabilities.