Combat actions

Matt Miller 

Hence, the need for combat actions. It makes combat more than two stacks of HP chipping away at each other. And it removes the need to optimize your character to do that. I'm not saying there isn't a place for that style of play, of competitive optimization...just not  a good table-top game. It throws away the primary advantage of playing with a live GM instead of an AI: You have a referee who can decide the outcome of events that transgress the rules, or which are outside the rules.

Damage scaling and danger of death

Matt Miller 

JRPG's had highly variable damage, compensated with a high 'buffer' of HP. As levels rose, both damage and HP scaled. At lv 1, a hit was 9 damage to your 50 life. At lv10, a hit was 99 damage to your 500 life, and at lv100, a hit was 999 damage to your 5000 life.

 Matt Miller 

Part of what informed Mayhem was that high level fighters are piles of HP. I had a fighter with...112 HP? While successful toasted by a dragon, mere human weapons had small effect--even the most successful attack was never going to deal more than 30 damage. 

Keith J Davies

It makes for a steeper survival curve between levels. At any given level you'll see a comparable fight time (as measured in rounds; spell animations might take longer) because damage is a pretty consistent fraction of hit points. However, if you see a significant mismatch in levels it's going to be very unbalanced

For instance, two Ftr1, two Ftr5, two Ftr9, two Ftr13, two Ftr17... at each level, the pair might fight for about 3 rounds before one or both drop, and the outcome is probably unpredictable. However, a Ftr5 vs Ftr1 might be only a round or two and very predictable because the Ftr5 can one shot the Ftr1, while even if the Ftr1 hits the Ftr5 he won't do enough damage to matter... and the Ftr9 vs. Ftr1 might get his armor scratched but is in totally no danger..which is fine. 

Rolemaster, and IIRC later editions of D&D, and several others have made similar design decisions. It encourages a degree of grittiness and caution. Even early editions of D&D had a lot of places where you could die pretty easily, even at high level (save or die effects, really big damage against hit point totals rather low compared to 3e-era D&D, etc.)...There is a place for it. Hit point vs. damage escalation works well for the heroic warrior vs. sword fodder trope. 

Other games aren't looking for that. They want to have deliberate distinction between characters of different grades. Depending on how other things work it might be a matter of hit points and damage scaling by level (which prety quickly moves low-level and high-level into different arenas), or it might be different effects trumping each other (*cough* casters in D&D 3+ vs noncasters), or different ranks being able to affect others or not (in Paranoia, low-grade weapons don't affect higher-grade characters), or defensive powers that come online and render higher-level characters functionally invulnerable (many supers games).

Matt Miller

You make good points. I think I'm less and less jazzed on benchmarking fighters. A fight should always be dangerous, even from a position of overwhelming superiority.

Bad Defaults

Matt Miller 

When making Mayhem, we mentally defaulted to 20 as a top number, after so many years of DnD. Bonuses we re scaled accordingly. It's one of the things I regret the most. The end result was that all numbers were scaled by 5x, which did nothing but make the mental math harder.The higher numbers made margins larger, and since margins affect total damage, higher HP was required to compensate. So HP gets measured in dozens, not singles. It also made skill mismatches, for different levels of enemies, much much more deadly.

 Keith J Davies 

Heh, as it happen Echelon keeps hit point totals pretty low. The attached (mocked up, lots of text is _wrong_) character sheet shows a 28th-level character (D&D 20th-level equivalient in principle) with 35 hit points. One hit point per level, one more at the top of each tier (so one every four levels, by default). If this characer built up her Fortitude more she'd have more hit points, potentially a total of _49_.(For reference: dice pool mechanism, every die that rolls 5+ is a success, every 1 is a complication, you can spend a success to cancel a complication. She's 7th-tier and thus rolls 7 dice for everything, but you don't have to add them up... adding them up would suck,)

Fate Style Conditions

Matt Miller

Wondering if it might be possible to use fate style conditions, where you do a set up that adds to your next attack.

Keith J Davies 

I really like the fate aspect mechanism, and being able to 'cause' aspects (I forget the exact term, it's not invoke, you're making a new aspect exist temporarily) is a nifty mechanism.

Arkham Horror Mechanics

Matt Miller 

I've been playing a great deal of Arkham Horror, where a 5 or 6 on a d6 count as a success. Rolling 8 or 9 dice, it takes a bit to pick out successes. But estimating how likely something is to succeed is easy; every dice you roll gives you 1/3 of a success. Which makes it tempting to adapt as a base mechanic for an RPG. But I'd like more than two ways (1) Extra dice, 2) Extra success) to affect a players roll.

Matt Miller 

Arkham has a mechanic which asks you to roll a success. Mentally, it's hard to imagine how that affects the outcome, as it is deeply unintuitive. In effect, it squares your chances. If you roll 1 dice, your chance is 1 in 9 (1/3*1/3). But if you roll 2 dice...I become confused. The probabilities get wonky. Chance of getting 1 success, on X dice, squared, is the match. X*(1/3)^2...I think. If I have 5 dice, the chances of getting 1 success, twice, is not the same as the chance of getting 2 successes on 10 dice; the first is harder.

Timing and Initiative

Matt Miller

In Mayhem, in devising round-timing we started with a DnD based 10-count round, then expanded to a 12-count 'combat clock' (with 10+ combatants, players were 'tied' too often. But our actions take 5,6,7 or 8 ticks on the combat clock. Somewhat with the idea that people would not overlap too often. But it happened often, as initial position on the clock was set so that everyone started at a different number. But after several hours staring at number systems, I begin to wonder if was a good idea.

If starting position+action is the formula, 1+8,2+7,3+6 or 4+5 all wind up on the same tick on the combat clock. Yet starting all at 1 would not help: All 'fast' actions would wind up at the same tick-mark. Increasing the magnitude of numbers makes for more mental math, not using single digits, but the biggest problem seems to be the change-over, when 11+7=6

Rob Hicks 

Actually, we started with a 100-point combat clock first. Then went down to 50, then 30, then 10, then 12, assuming my memory serves. 

Matt Miller

That seems...not unreasonable. Why did we ditch 30? It was such a reasonable number. Divisible by 15,10,6,5,3,2,1...60 might actually have been even better.

Matt Miller

Given that the virtues of the Fibonacci sequence are that two numbers in it add up to a third number, I am thinking we'd have been better to use non-Fibbonacci numbers, viz: 4,6,7,9. Rounds take a long time, so I recall estimating HP on the basis that no battle should average longer than 10 rounds.  I tried a number of combinations, and they all overlap to some degree. 56789 just happens to overlap a great deal (20%) in the first two moves, which always bothered me. But alternate numbers are as big a head-ache: Either ratio between the highest and the lowest becomes too big, or the numbers become too big to do mental arithmetic with. 6-10 or 7-11 both have better ratios, but both still have the same 'overlap' problem. (7+11=8+10=9+9). Adding gaps between the numbers (1,3,5,7,9) fixes the overlap problem, but either generates speed ratio problems (9:1) or (using 5,7,9,11,13) mental math problems. The speed ratio was a problem because...players get bored when they take a 9 speed action, and someone else takes 9 1-speed actions. 

This was particularly true because there were a couple of effects in the game that adjusted speed slightly. Even a speed adjustment of 1 was HUGE at the lower ranges. An action at speed 3 becoming a speed of 2 is a massive boost across the combat wheel, where a 5 becoming a 4 was a little more reasonable and easier to plan effects around. Gave a boost without being a relative game-changer. 

Keith J. Davies(?)

...given 7, 9, 11, 13, your fastest is not quite twice as fast as your slowest. I'd probably consider a 1-20 scale for this (which happily aligns with a commonly-available die... not that that's really important)(and is still about twice as big as you're looking for) 

 Matt Miller 

Why is 12 better than 10, or 30?  Even fractions of APR, I guess. 60,30, 20,15,10, 6,5,4,3,2,1. So why not 24 or 18? Both are still 'hex'.

Rob Hicks 

Heuristics. Convenient Mental representations. People who do statistics for fun and profit tend to have no problem with number crunching, and don't understand why people who don't can't wrap their heads around them. 30 is as good as 24 or 18 for them. 10, however, is iconic in human math. Decimal counting system, fingers and toes, very easy. 12 only slightly less easy, but was a little kinder than 10 on the backend, and had the added parallel of the clock, both represented in the fact that it is a time-tracking system, and is something that people could reasonably count around. 18 and 24 have neither advantage.

Keith J. Davies

10 or 12 are good, 24 might be too big (but D&D 3.x-style initiative is d20+Dex mod+stuff and can easily reach that high).Perhaps use prime numbers for your speeds. Not 5, 6, 7, 8. 12-step scale, speeds 2 (really really super fast, probably not seen often), 3, 5, 7, 11 (really really slow, probably not seen often). If you start on phase 12 (*cough* HERO) you end up with everyone using the same speed actions hitting the same points and staying together. If you randomize, even slightly, then you might find that people at the same 'speeds' never overlap (if they didn't in the first place), while those acting at different speeds who overlap do so _once_ before they need to lap the field. That is, speed 3 and speed 11, once they sync, will need to go 33 phases before they sync again (11 actions and 3 actions respectively). These might be too far apart; being able to act so many more times than an opponent is pretty brutal. 30-phase rounds with actions 5*, 7, 9**, 11, 13, 17*** might be a better fit. 

* aligns with round, will always be on the same phases each round. Perhaps not desirable. 

** not prime, but _relatively_ prime 

*** maybe too big, less than two actions per round

Rob Hicks

Makes sense. Didn't think of primes, which is very interesting mathematically. Part of the reason actions varied from 5-9, (very fast, fast, average, slow, very slow,) was because the lower numeric end of the scale was trickier to balance for weapon damage. A weapon with a speed of 3 would be 3x faster than one with a speed of 9, where 5:9, was a little less than half as effective. Made the numbers easier to balance on weapon damage and such, when we were trying to get relative efficacy.

Rob Hicks

In the final release, we left it at 5-9, because ultimately the "tied" locations didn't affect things as much as they were initially thought to do. Everyone started on the same spot, and we broke the initial tie by going in order of Agility scores. When two characters tied, we broke the ties in the same way. When those ties were allies, it didn't usually matter. If one wanted to go first strategically, they lost nothing by letting one go before the other.

Matt Miller

We were also trying to keep it to keep the actions speeds to 'counting numbers' under ten. 

Andy Klosky

What is the purpose of having such a convoluted initiative system?  Does it make the game more fun? More realistic? Simpler for player or GM? Emphasize a type of combat, such as ranged or melee? Balance other game mechanics? 

Matt Miller

Mayhem was designed pre-4E DnD, so we made a lot of the same mistakes as 4E. So it has a very crunchy war-game style of mass combat, focused on melee. It also had a very extensive weapons list, with something like 60+ weapons, all of which had unique stats, yet all of which had to be roughly comparable in the amount of damage they did. 

Rob Hicks

The combat clock allowed an initiative system that was actually very easy to track and intuitive for players to run, but allowed for faster and slower actions in the game that had tactical meaning.  That in turn let us adjust the damage for various weapons, so that daggers fought like daggers. Less damage per hit, but faster attacks and more strategic maneuverability. A maul, by contrast, took ages to swing, but hit like a truck. It really, really improves the cinematics and style of combat, giving different play styles that highlighted different character concepts. In addition, working on a timing-based system, rather than a bonus-based system, meant that where you were on the map and when you were going to act next was strategically important. It gave a great deal of depth in combat for those that like the wargamey tactics style of play, but kept the mechanics still relatively simple and smooth for the people that were more interested in the style and cinematics of the game.

 Rob Hicks 

Ultimately, people were very aware of who had the better agility as combat went by, so ties on the combat clock didn't end up slowing things down at all, and it gave Agility a particularly nice perk for stat balance. Additionally, characters acting at the same spot on the combat clock often had other advantages. Team attacks, for example, getting an extra bonus for simultaneous whelm attacks, or for big cinematic clashes where two characters struck simultaneously, like a classic samurai duel.

Armor and Shields

Matt Miller

Players tend to forget penalties, so all numbers were bonuses, with no "maluses", so rather than -3 to +3, we had 0-7. Given that players could always default to 'bare hands' at +0, all weapons had to provide a better bonus than that, so our range was only +1..+7. But the 'big numbers' made weapon very difficult to balance, so in the final list, the numbers were much compressed, even as it made weapons much more similar to one another than I would have preferred.

Rob Hicks 0 to +3 is what we ended up with for skill bonuses. +3 ended up being rare.Ultimately, for simulation, we ended up doing Jake's method of multiplying the damage of a weapon at a specific die result by the probability that such a result would happen and summing them for average damage. That let us calculate relative bonuses for two opposing dice rolls and two opposing bonuses. (What if your opponent has a shield? For example.)

Keith J Davies Ah, shields are easy. I think. "Shields must be splintered!" rules: sacrifice your shield to prevent a hit. Done.

Or, since armor is a function of coverage (die size: d8 means it applies half the time, d12 means it applies 2/3 the time) and quality (leather is 1 die, steel plate is 3 dice), a shield might simply count as another die of armor (or, more likely, give you new options in the fight; as far as I'm concerned a shield is actually a weapon, specialized to defensive purposes).

Rob Hicks  Matt Miller always pushed for shields as a piece of armor, while I always pushed for shields to be treated as a weapon. My logic to justify my stance was to buy a padded shield and attempt to beat him with it. tongue emoticon

Matt Miller My reply was to build a viking shield, and then Rob attempt to beat me around it....which left only my head and legs uncovered, demonstrating that the Greek hoplite gear (Helmet, shield, greaves) is (1v1) astonishingly effective.

Matt Miller I had forgotten the wrinkle of different defense numbers.

Keith J Davies In previous efforts, I've found that it often is beneficial to treat a shield as weapon specialized for defense...it let me collapse a surprising amount of mechanics.

Matt Miller Yeah, the combination of armor+shield can form a near-wall. Which, while realistic...isn't that enjoyable for gameplay. Armor mechanics are a head-ache all their own.