What's it like playing a diceless wargame?

[Elbows]

That's a much more polite way of putting what I had written earlier.

Simply put, Bollingar, I disagree with you immensely on almost everything you've stated.  I found it insulting and childish to pull out the "Well I did a study and wrote a book on it" nonsense.  If that's true, it gives me even more concern...

But you're on your sales pitch, so I won't disrupt it further by opposing your opinion - and let's be very clear, what you're stating is opinion and not fact.

[Bolingar]

That's a much more polite way of putting what I had written earlier.

Simply put, Bollingar, I disagree with you immensely on almost everything you've stated.  I found it insulting and childish to pull out the "Well I did a study and wrote a book on it" nonsense.  If that's true, it gives me even more concern...

But you're on your sales pitch, so I won't disrupt it further by opposing your opinion - and let's be very clear, what you're stating is opinion and not fact.

No, not a sales pitch, but a reply to your post:

Curious how you know what happens on a real battlefield, particularly historical ones?  Even well-read historians admit we know very little about how the battles looked or were fought.  We don't even have a solid idea how Roman legions fought (really), and they're one of the most well-researched periods with the most evidence on hand.

I disagree with this. We can have a pretty good idea of what happened at least with some battles, and the more recent the battle the better our idea. We know what happened at Waterloo. But even for battles in Antiquity we know what happened for the better documented ones like Cannae. The problem with a large part of contemporary academia is that it is overly skeptical of the primary sources, far too ready to doubt them when there is no objectively good reason for doing so. This is compounded by the problem of translations when the source material isn't abundant, for example in the case you give of how the legion worked. I made a study of the primary sources in the original languages, notably Livy and Polybius, and realised that the quincunx is a complete fabrication, right down to the word itself. The legion could not and did not work that way. The quincunx theory developed from a mistranslation of Livy's description of line relief in History 8.8:

"The first line, or hastati, comprised fifteen maniples, with short distances between them." - Prima acies hastati erant, manipului quindecim, distantes inter se modicum spatium.

Every standard translation makes the "between" apply to the maniples, and has maniple-wide gaps between one maniple and the next. In the Latin however, the subject of the sentence is the hastati, not the maniples, so the small gaps logically apply to them. Furthermore the Latin term for "between" - inter - has the primary meaning of "in the midst of". It's the root of the word "internal". So it is the hastati who are standing apart from each other. This immediately suggests the open order disposition described by the military tacticians of Antiquity: Aelian, Arrian and Asklepiodotus, where the files in a formation each occupy a width of about 2 yards, with about 4 feet between the shoulders of the men of one file and the next. This permits other men to occupy the spaces between the files and move along them, i.e. one body of men can pass through another. This could be light infantry (insertion - parentaxis) or heavy infantry (interjection - parembole). Livy, using non-technical language, was describing something familiar to military men who had technical terms for it.

But we can happily agree to disagree. We're talking about games after all, not about politics or religion. 😱  Pax?

[Bolingar]

I think you have missed my point - there is either randomness in life, or there is not. If there is, then no battle, no matter how huge the numbers involved and how much historical data there is to draw on, will always have exactly the same result. Nothing in life is truly deterministic and therefore a deterministic wargame can't be more 'accurate' (in and of that being the definition), than one with some random elements.

Sure. Battles are fought by human beings and human beings have free will. Which is why it's impossible to use deterministic mechanisms to model skirmish-level warfare where individuals are the units. But units composed of hundreds or thousands of individuals tend to behave in more predictable ways, never 100% predicable of course, but sufficient that deterministic mechanisms give at least as plausible a simulation of them as dice do.

The general orders the cavalry reserve to engage; a gust of wind blows the flag the wrong way; the cavalry takes to the wrong flank. Stuff like that can't be simulated by anything deterministic.

You have a point. The charge of the Light Brigade happened because Lord Cardigan misread his orders. Bit hard to simulate though - the entire battle is decided by a single throw of the die: 1,2,3 and the British charge the wrong guns and are wiped out; 4,5,6 and they charge the right guns and are victorious. Then you have the impetuous Ney at Waterloo who thinks the English are retreating and orders an immediate cavalry charge whilst Napoleon is out of action from stomach trouble - and the French cavalry are decimated by the English squares. How would you game that?

I think though that in earlier battles this wouldn't happen, or at least happen rarely: commanders got their orders before the battle took place and there were several ways for the general to time the execution of those orders in a way that is clearly understood by his commanders. Battleplans in any case were much simpler then - generally wait until you hear the sound of trumpets. Then advance. Much less chance of a commander getting a message in mid-battle that he misunderstands.

[Pattus Magnus]

Battleplans in any case were much simpler then - generally wait until you hear the sound of trumpets. Then advance. Much less chance of a commander getting a message in mid-battle that he misunderstands.

Bolignar, you may be underestimating the universal human capacity to bugger up even the simplest plan!

I?m kidding, of course, but there really is a wide range of ways that people land on sub-optimal choices.

For gaming, I think one of the most important factors is building into the game mechanics the need to make decisions with imperfect information. In a dice or card driven game, the randomization mechanic serves that function.

There are ways to achieve imperfect information in purely procedure driven games that would give a complex game experience. First, have both players initially deploy blinds rather than units. The player knows the strengths of their own force, but not exactly where the enemy will place his strongest units, so has to make some initial guesses. Also, both players deploy without measuring, so the exact timing of contact is difficult to gauge - you want to be the one to have momentum in the charge to contact, but can?t effectively min-max by measuring. Just those limitations on player information will force some  decisions that are analogous to what ancient generals faced. You know what you have, but not necessarily where it will do the most good, so strategies that hedge risks and hold a reserve make sense.

Once the deployment is complete (or maybe when enemy units get close enough to identify - deploying with a scout screen would be useful, but redeploying based on the new information would be risky), the then the figures go on the table. There are still some procedural ways to generate incomplete info, though. Each unit can have a couple of stats (offense, defense, etc), so once contact is made the offense reduces defense at a given rate - the problem is that you may not know the exact numbers the enemy unit has until they make contact and it can take a few turns to defeat the enemy. So, the Spartans might have higher offense than typical hoplites, but similar defense - they will probably beat the enemy unit in front of them, but will it be before the perioikoi or allied hoplites in the line next to the Spartans collapse and leave them isolated? (Always worth consider with Spartans that a lot of their field forces weren?t Spartiates, and might not be especially motivated?).

I can see a purely procedural game giving just as many ?oh, shit!? surprises as a randomized game system, but there are some different game design challenges. I don?t see replayability being a problem- even two experienced players will be able to give each other surprises if they have imperfect information to work from!

[boneio]

Very nice Pattus, that's a great illustration of how a more deterministic game can be both fun and simulate the general not being omnipotent and omniscient (as regards their own forces and the battlefield at least).

If a GM sets up various hidden information then I'd have to agree that's as good a way of simulating unknown conditions as having, say, a card draw which triggers one of a set of possible events.

I still entirely disagree with Bolingar's assertion that a deterministic game is a more 'accurate' simulation than one with some randomness built in.

Even if we agree to disagree on the performance of given units being entirely repeatable and predictable, for a 'realistic' battle simulation there still needs to be some degree of uncertainty for the general in how their commands will work out. That doesn't have to be via randomness, on that I concur - a GM could arbitrate that a unit gets unexpectedly bogged down or confused and so arrives late, for example, simply to make the game more interesting (or because the GM knows that route is via marshy ground but the general doesn't).

The charge of the Light Brigade happened because Lord Cardigan misread his orders. Bit hard to simulate though - the entire battle is decided by a single throw of the die: 1,2,3 and the British charge the wrong guns and are wiped out; 4,5,6 and they charge the right guns and are victorious. Then you have the impetuous Ney at Waterloo who thinks the English are retreating and orders an immediate cavalry charge whilst Napoleon is out of action from stomach trouble - and the French cavalry are decimated by the English squares. How would you game that?

I think you're being reductive in order to try to strengthen your argument. This is the false dichotomy another poster referred to. There simply needs to be some representation of the possibility of the charge going awry, it absolutely doesn't need to be 50/50 - it might only be a 1 in 100 chance.
As I've pointed out repeatedly, the general is not omniscient or omipotent, that's what is in contention - it doesn't have to be simulated via randomness, Pattus (and I'm sure your good self too!) have given some examples of how it can be done differently, but there has to be uncertainty to make the battle (a) worth gaming and (b) 'feel' right.
I understand that you're contending that the actions of your opposing player will provide the uncertainty, which is clearly agreeable to a point, but that doesn't overcome the hurdle that units and battlefield always behaving in a way which is entirely predictable to the player, is terribly unrealistic and therefore no, a deterministic approach is not a 'better' way to simulate a battle.
Too much randomness, I agree, is the opposite side of the same coin as too much determinism.

I think though that in earlier battles this wouldn't happen, or at least happen rarely: commanders got their orders before the battle took place and there were several ways for the general to time the execution of those orders in a way that is clearly understood by his commanders. Battleplans in any case were much simpler then - generally wait until you hear the sound of trumpets. Then advance. Much less chance of a commander getting a message in mid-battle that he misunderstands.

I still don't quite agree it's that simple, but let's say I do. You said yourself, there is a chance - which needs modelling.

[Bolingar]

I entirely agree that fog of war is best simulated by fog of war: a player has imperfect knowledge of the capabilities and whereabouts of his opponent's units. Columba Games' blocks do that nicely. But I suspect it doesn't go down well with most players which is why it isn't widely practised. Players want to be God: seeing from heavens their army and their opponent's army deployed in all their glory. Sure, some rulesets allow for hidden units secretly deployed for ambush, but how popular is that, really?

Wargames invariably give the player far more control than an historical general actually possessed. For a typical pre-gunpowder battle, the general usually made his plans the night before, gave his commanders their orders, then took up position next morning with his personal unit and prayed his plan was a good one because there was precious little he could do about it now. The advance is sounded, he leads his unit into battle (or stays back in reserve) and watches his army attempt to carry out his prearranged battleplan. If his opponent has outguessed him or he has forgotten something then it's curtains. Not much fun as a wargame.

It's better for a horse and musket army with its system of mid-battle orders but even then the general was not always in a position to see the entire course of the battle himself. Often he had to rely on reports and they could be vague, distorted or out of date.

Anyhow, point is that most players IMHO want more control than that, including perfect information of their opponent's dispositions.

[Pattus Magnus]

Bolignar, that?s a fair point about the preferences of players. I was thinking more in terms of the feasibility of procedure driven games to generate fog of war than in terms of whether or not the game would be widely popular. Even then, how a procedure driven game is received has a lot to do with individual gamer ?type? (as discussed in other threads, and magazine articles). I am solidly a scenario or narrative gamer- I play mainly for the process of finding out what happens, so a procedural game appeals to me as long as it tells a good (interesting) story. Folks that are more into tournament gaming will probably have more trouble getting behind the assumptions built into a  procedural game engine.

[Bolingar]

I think you're being reductive in order to try to strengthen your argument. This is the false dichotomy another poster referred to. There simply needs to be some representation of the possibility of the charge going awry, it absolutely doesn't need to be 50/50 - it might only be a 1 in 100 chance.

Sigh...I wrote that in answer to your post, agreeing with it. A single commander can by a bad decision virtually give a battle away. But I don't think we attempt to wargame that. Other decisively fortuitous events like the appearance of a small force over the hill on the enemy flank that the enemy takes as an entire new army and routs in panic we also don't wargame.

As I've pointed out repeatedly, the general is not omniscient or omipotent, that's what is in contention - it doesn't have to be simulated via randomness, Pattus (and I'm sure your good self too!) have given some examples of how it can be done differently, but there has to be uncertainty to make the battle (a) worth gaming and (b) 'feel' right.

There's dollops of uncertainty in a determinist game, but I think it depends on what kind of uncertainty you're looking for.

I understand that you're contending that the actions of your opposing player will provide the uncertainty, which is clearly agreeable to a point, but that doesn't overcome the hurdle that units and battlefield always behaving in a way which is entirely predictable to the player, is terribly unrealistic and therefore no, a deterministic approach is not a 'better' way to simulate a battle.

Not quite. I'm contending that the impossibility of predicting outcomes is what constitutes the uncertainty in a deterministic game. Battlefield units behaving in a predictable way without any arbitrary random variability is not terribly unrealistic, just no less unrealistic than using dice to impose randomness.

Too much randomness, I agree, is the opposite side of the same coin as too much determinism.
I still don't quite agree it's that simple, but let's say I do. You said yourself, there is a chance - which needs modelling.

I'm not on a crusade to abolish dice in wargaming - might as well try to fly by flapping my arms. However I do affirm that chanceless wargaming is fun in its own way and just as capable of satisfying a player's suspension of disbelief as a dice-driven game, perhaps even more so. But let players decide that for themselves.

[SteveBurt]

Bolingar wrote:
Not quite. I'm contending that the impossibility of predicting outcomes is what constitutes the uncertainty in a deterministic game.

Except that clearly you can predict the outcomes; if the rules are simple enough to work out combat outcomes, then they are simple enough to work them out in advance. I get that what you are saying is that you may know what will happen in one round of combat, but what might happen after another when another unit intervenes is harder to know. But it is still possible, and clearly much easier than in a combat system with an element of chance. I have certainly in the past worked out the odds of something several moves in advance even with chance; I can easily figure out rough percentages of different outcomes over several different scenarios. In rules without any chance, that is much simpler.

You say that predictability is no less unrealistsic than chance, but I think real combats, even between large numbers, are not nearly so quantifiable as you do.
Don?t get me wrong. I admire that you are trying something different. I just feel that the assumption you have made about mass combat being predictable is not supported by evidence.

[jon_1066]

The key to the rules to me appears to be that a small change on the input to a combat can have a marked effect on the output of the combat.  So trying to quantify all of those inputs in advance and how they will cascade through the battle line will be beyond most players.  Sure a computer could spit out the answer but we are not computers.  It also does have user input in ordering stuff about, especially the light troops.  If you combine that with hidden info on leaders then it would be very difficult to predict the final outcome at the start.  You also aren't going to be fighting the same battle all the time so even if you "solve" one just move on to the next.

For a wargame you need to be able to suspend your disbelief (does this give a realistic feel within its own setting?) and have interesting decisions to make.  Trying to do that in a deterministic way is really quite difficult so I take my hat off to the OP for trying.

[TheDaR]

Bolingar wrote:
Not quite. I'm contending that the impossibility of predicting outcomes is what constitutes the uncertainty in a deterministic game.

Except that clearly you can predict the outcomes; if the rules are simple enough to work out combat outcomes, then they are simple enough to work them out in advance. I get that what you are saying is that you may know what will happen in one round of combat, but what might happen after another when another unit intervenes is harder to know. But it is still possible, and clearly much easier than in a combat system with an element of chance. I have certainly in the past worked out the odds of something several moves in advance even with chance; I can easily figure out rough percentages of different outcomes over several different scenarios. In rules without any chance, that is much simpler.

Without weighing in on the rest of the argument, I will disagree with this specific assertion.

Reducto ad absurdum, you have chess.   No variability at all, extremely limited board actions allowed, fixed grid, etc.  Yet very few people will assert that chess is too predictable.  Predicting more than 2-3 moves in advance puts you in the very highest echelons of play.  It doesn't matter that you can calculate that a pawn takes a bishop an infinite number of turns in advance; every pieces always takes any other piece if you can force the position such that the attack can happen at all.  The trick is putting that pawn where it will be able to take the bishop without giving up other more tactical sound positions.

Now compare that to a wargame where you have factors like larger variability of unit capabilities, terrain's effect on movement and combat factors, movement that isn't necessarily on a grid, tactical modifiers to combat like flanking, etc.   Yes, you can calculate the factors of a single combat many moves in advance, be it diced or diceless.  The question is can you actually force that combat to occur as you want without those other factors changing the results along the way.   

I came to historical games via a fairly typical route of passing through Games Workshop mass market games first; games that were largely throwing large buckets of dice and the strategy and tactics being much more about army building, fairly trivial tactics (because you don't get to move much in a 6 bound game) and getting occasionally lucky on a roll.  When I passed into games like DBA, I struggled quite a bit before realizing that most historical rulesets were not about winning the dice rolls, but forcing the match ups on the table you needed.  Artillery vs Infantry vs Cavalry as a classic example.  Or getting multiple troops into position where the opponent could not evade an attack with flanking and rear support.

Zooming out to the larger conversation, I reiterate a point I made earlier in pointing out that how you apply randomness can be much more important than a lot of people tend to consider.  It is interesting to note (without judging), that the more dice you throw, the closer to statistical averages you are likely to see (Law of Large Numbers).  Thus, in some ways, huge buckets of dice games like Warhammer are actually closer to diceless resolution than games which use only use 1 or 2 dice per combat.  When you start a round throwing 40 dice, then re-throwing 30 of them that registered as hits, and the opponent effectively rerolls 20 of those to "save" the wounds...  Then the opponent goes and rolls his 30 attack dice, etc, etc.   150 total throws in a single combat out of a half dozen on the board each turn starts making significant deviations from averages quite uncommon.   Compare with a WRG DBx type game where a single combat resolution is 1 roll on each side, and thus a much higher variability, and a half dozen dice rolled for an entire turn of combat might be typical, and suddenly that randomness of the classic 1-6 roll feels a lot more impactful.

Another consideration is what are you making measurements of.  The standard deviation of human performance isn't that high for bodies of trained troops.  Are Spartans really that much better than Athenians because they were individually stronger, faster, more enduring, or so forth and thus won more hand vs hand combats across the length of a battle line?  Or did Sparta win because their officers trained enough to be able to follow directions in chaotic tactical battle conditions, and were thus were more often able to perform maneuvers that broke them out of slogging battles of direct attrition and end up with "tactical factors" like getting bodies of troops into flanking positions?  Or did they win because at the highest levels their kings and generals strategically they ensured that they didn't take the field in losing conditions and instead refused to give battle until they had superior battlefield and morale conditions?   Each of these have different ways they can be represented.  Maybe it is fundamentally okay to assume that a Spartan body of troops is within standard deviation of Athenians in close combat score and thus a die roll is fine, but that instead Spartans should have better Command points or scores or whatever allows them to maneuver and rally better, or that they should be the ones to dictate terrain and deployment.  In the end, none of these things necessitate the lack or presence of physical randomization; you can construct rules that use any method of unpredictable behaviors, so long as you're putting the unpredictability in the right places.

[SteveBurt]

TheDaR wrote:
Without weighing in on the rest of the argument, I will disagree with this specific assertion.

Reducto ad absurdum, you have chess.   No variability at all, extremely limited board actions allowed, fixed grid, etc.  Yet very few people will assert that chess is too predictable.  Predicting more than 2-3 moves in advance puts you in the very highest echelons of play.  It doesn't matter that you can calculate that a pawn takes a bishop an infinite number of turns in advance; every pieces always takes any other piece if you can force the position such that the attack can happen at all.  The trick is putting that pawn where it will be able to take the bishop without giving up other more tactical sound positions.

------------

But that is my point; if both players in a game of chess, or a diceless wargame, make the same moves, the outcome will be the same.
I contend that is manifestly *not* the case in warfare. Even if the commanders make the same decisions and commit the same units at the same time, the outcome will not be the same. There is randomness in real warfare; even at the level of mass combats. It doesn't even out into some kind of nice statistical smoothness. Outliers can and do occur. Imagine a game of chess where there is a certain percentage chance that a capture fails. Suddenly many things which used to work do not, but conversely some things that were impossible become possible.

[Bolingar]

But that is my point; if both players in a game of chess, or a diceless wargame, make the same moves, the outcome will be the same.

But who does that? Chess players aren't trying to replicate a previously played game; they trying to beat their opponent, which means catching them out with a different strategic development and different tactical combinations. Ditto for wargamers, but in their case it's compounded by the fact that no two games start out the same: with hundreds of different army lists to choose from and free deployment on terrain that is different each time, the odds of an identical setup are vanishingly small. Deterministic games in consequence are as variable and unpredictable as dice-driven ones.

And in the real world the same thing can have absolutely the same result if done the same way. When the Romans besieged the hillforts in southern Britain after the initial fighting had been settled, the commander would write a report on the victorious outcome of the engagement before it had even taken place. They knew exactly how to assault a hillfort and got the same routine result each time they did it.

[boneio]

But that is my point; if both players in a game of chess, or a diceless wargame, make the same moves, the outcome will be the same.
I contend that is manifestly *not* the case in warfare. Even if the commanders make the same decisions and commit the same units at the same time, the outcome will not be the same. There is randomness in real warfare; even at the level of mass combats. It doesn't even out into some kind of nice statistical smoothness. Outliers can and do occur. Imagine a game of chess where there is a certain percentage chance that a capture fails. Suddenly many things which used to work do not, but conversely some things that were impossible become possible.

This ^ This is the point in contention.

Chess is not a solved game. It's beyond current human capability. However there is soft proof that it *can be* solved/.

Someone mentioned that wargames, even diceless, are more complex and therefore even harder to solve. But they can *be* solved, should sufficient computing power become available (decades away unless quantum computing pans out quickly).

I do think there is a distinction between some of the assertions being made:
Compex diceless is more realistic due to being harder to predict - disagree, consensus appears to disagree, logic appears to disagree! It's in practical terms hard to predict but it's deterministic so it's only a matter of capability and the underlying premise that for the same exact set of conditions, the outcome for the same set of actions is identical, is literally the opposite of realistic. If that were true in the real world, free will doesn't exist. One can argue over just how much variance a given army might have in performance if the other parameters on the day were somehow miraculously identical, but there absolutely would be variance.
Complex diceless wargames in practice can be varied enough that the lack of randomness doesn't make the play experience worse or make it 'feel' pre-determined - yes, I think that's fair, as someone referenced for chess a couple of posts ago.

[Bolingar]

Compex diceless is more realistic due to being harder to predict - disagree, consensus appears to disagree, logic appears to disagree!

I'm not saying that. My point is that a deterministic wargame better simulates fog of war in that fog of war consists of lack of knowledge of factors that themselves are not random in their operation. Whereas fog of war is best simulated by fog of war, deterministic mechanisms - which are impossible for a human player to calculate in advance right to the conclusion of a game - provide that imperfect knowledge of effects that are not random. Dice provide lack of knowledge but only of random effects which correspond to nothing in the real world.

As for the human factor - not knowing how an opponent's forces will act - that is supplied by the opponent himself. Two human beings playing a deterministic game ensure that gameplay is anything but deterministic. Free will remains intact.

With the caveat that this works only for games where units represent large bodies of men whose individual behaviours average out. It doesn't work for a skirmish-level game where the individual is the unit and the complex web of cause and effect in a human being is impossible to simulate with deterministic mechanisms.