This is a mass combat resolution scheme - so we should all know right off that it's silly and fun and shouldn't be taken too seriously. It might, however, actually work; I have done zero testing, however. That should not be any reason not to scratch out the idea in words, nyet?
Fine. Here is what we will need. A number of dice equal in number to the total number of battle figures that can fit inside a 45-foot diameter hex, times two. Well, we can use any hex size we want, but since my combat system is based on 5-foot hexes, a 45-foot diameter makes a nice hex circle that will fit 61 well-spaced combatants - an approximate company-sized unit. Plus most long-time players can mass around 122 total dice.
Yes, 122 because we will need dice for both sides. There are some fun little details in which dice we pick, but for the moment let's just assume both sides have piled up a bunch of dice that accommodate the number of combatants.
We will need two buckets and an empty table. Flat-sided buckets are better than round and if at all possible, the sort used for digging in sand (or for office supplies) is better than something you'd use to clean your kitchen floor. The best sort of buckets are six inches in diameter - these can be gotten cheap from most box stores.
Fill the buckets with the dice. Take opposite sides of the table. Now pour your dice onto the table in the direction of your opponent, at the same time. That's very important. I recommend that the square edges of the buckets be rested on the table (after a good shake) and that the bucket edge NOT be lifted from the table as the pouring is done (it can be called a mis-battle, like a mis-trial, if someone breaks the rule or there could be a -1 modifier placed on every die rolled by that opponent).
At this point, obviously, it is a good idea to be able to tell whose dice are whose, based on color! So you want all green and blue dice versus all white and black dice, if possible - the greater the differentiation, the better.
If my imagination sees this right, the temptation to pour your dice out too quickly will cause them to scatter on the table, spreading out your forces and making them more vulnerable when the combat is resolved (getting to that). On the other hand, pouring them out too slowly will result in too many of the combatants being rendered "behind the lines" and useless to you. What's wanted is a nice, even line that gets as many combatants into the resolution as possible without there being too many gaps.
Any dice rolled off the table are considered either a) deserters; b) captured by the enemy; or c) routed and available to the opponent for later rolls.
Okay, we should have two masses of dice on the table, having bounced against one another to varying degrees.
Now, if we have the means, we can pick individual dice to apply to specific types of combatants. We can d6 for pressed recruits, d8 for combat-trained men-at-arms, d12 for levels and d20 for spellcasters (we can also apply adjustments for certain colors, such as saying all the orange d12s on the board are +2 while all the green d12s are read as is, comparing that to levels). This gives us a nice balance for facing different kinds of trained troops against one another and for comparing one race to another (all the dice from the dwarves are +1, for instance).
Remove all dice that are more than 1 or 2 inches from opponent dice (I say one or two because I haven't tried this, so perhaps 1 inch leaves too few combatants or perhaps 2 inches leaves too many). This should hopefully leave a line of combatants that have either 'damaged' or 'killed' the other - it depends on how gritty we want our paper work. We can either resolve the battle to give damage to a long list of numbers we have that make up the companies or we can simply say the result eliminates combatants permanently.
Start with the highest dice on the table; each high die 'kills' an enemy die that the player chooses to kill (the whole battle must be observed carefully, because a good die might potentially kill two or three enemy of different ranks/power or in slightly different positions! Picking the wrong die might result in losing one of your own combatants from the high die you've left in the enemy's hands.
If two dice match in size, then either side may declare a 'duel,' taking both high dice out of the competition (whether or not the other side agrees to the duel). On the other hand, both opponents with the high matching dice may opt to agree to take out an enemy each, if that seems to suit them both better.
Conversely, if we prefer the damage variant (rolling damage as it's indicated by the dice) then duels both do damage and multiple higher rolls could all do damage to an adjacent lower roll (potentially having four recruits beating a leveled character hard-core.
Steadily, as we resolve the 'combats,' both the high dice and their 'hits' (plus any duels) should be removed from the table until there are no remaining matches (none are close enough to 'engage'). In all cases, the number of remaining combatants should be noted in number and rank - since both will represent a part of the larger battle. Suppose we are trying to resolve this in D&D:
|Picture Carpenter, Edith, Sir Conrad and Fletcher, et al,|
each leading a party of 40-60 combatants.
Breaks between combat would represent both sides getting tired after a set amount of time, agreeing to retreat momentarily before charging again. We might employ morale rules to see if one side has to retreat a whole hex. And if we have the dice, we can have three- or four-way attacks with dice pouring from every side of the table. Or we can keep with usual D&D and presume each unit faces each opponent one at a time. Of course, a beleaguered unit fighting multiple fresh units would roll less and less dice each time - though a set of freaky rolls coupled with a higher percentage of d8s or d10s might prevail against a lot of d6s.
The way in which the rolls were performed physically by the players could make a big difference, too, since a scattered roll would mean most combatants would 'stand alone' against three or four tightly rolled dice while there might be many dice that skip over and into the midst of the enemy or bounce right off the table.
Anyway, this is about 60 minutes of thought put towards the idea. I think it might be fun and it might actually work.