What's it like playing a diceless wargame?

[SteveBurt]

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.

The real world is inherently not deterministic. Further, all the factors which affect the outcome of a combat are not just unknown, they are unknowable.
If two units fight twice, the outcome will differ, maybe by only a few percentage points, but maybe not. Sometimes a random event during a combat can have a huge effect; for instance, you might say that a well formed square in a Napoleonic game can never be broken, but then you get instances like a dead horse falling across it and the square being broken. Very unlikely, but it did happen. A deterministic system cannot model such outliers; it assumes that things will be nicely in the middle of the bell curve at all times. That is misleading. Effects do not always average out.

[Belligerentparrot]

I feel like there is maybe some talking past each other going on here? Sorry if that sounds condescending, it really isn't meant to be. I'm not trying to over-explain anything to anyone, just trying to keep up with the debate!

Bolingar - and apologies if I am misunderstanding you - you seem to be very focussed on how to model "fog of war".

Steve and others - and again apologies if I am misunderstanding you - seem to be more focused on a different point, which is that there is something unrealistic about combat resolution with no randomness.

I agree (with Bolingar) that much of the fog of war is down to imperfect information, which this thread has convinced me can be modelled quite well in a deterministic set of rules.

But agreeing on that doesn't take away from what I take Steve's (and Boneio's, unless I've got it wrong) main point to be: that every violent encounter on a battlefield goes the way it goes in part due to pure luck, so a ruleset that aspires to realism (insofar as any truly can) should leave room for some degree of randomisation affecting combat resolution.

For me, that point is where I really part company with the idea of wholly deterministic combat resolution, though it was only through reading the thread that I came to realise this. Leaving aside the extreme cases (e.g. trying to shoot down a Stuka with a pistol; or exceptional elite forces coming up against very poor rabble), I do not know of any reason to think that when large numbers of people are involved violent encounters become less subject to the effects of pure luck.

I guess it might also depend on the relevant victory conditions. E.g. no sane person gaming Iwo Jima is going to stipulate the Japanese need to drive the Marines back into the sea. If *that* is what victory looks like for Japan, then yeah, as Bolingar said of the Spartans vs Athenians, every single time one side is going to win. But as soon as you stipulate more interesting victory conditions, it seems realistic (to me at least) to think that very careful planning and some good luck is going to be needed to carry the day.

[Pattus Magnus]

SteveBurt, that?s not quite accurate about the real world- it definitely is deterministic. If it was not deterministic, chains of cause and effect wouldn?t be a thing, but they are. The problem is that the precise cause and effect chain can only be traced accurately after the events have occurred. The reason that is the case is that nearly all real-world phenomena that happen on scales human beings are interested in during day to day life (including military combat) are outcomes of complex adaptive systems - systems with large numbers of mutually influencing parts. Interactions between parts are completely deterministic (something causes something else) but the final outcome is usually the result of multiple sets of interactions, so knowing how a small part works does not tell us how the whole thing behaves. Statistics help for prediction, because over a large number of repetitions (or a large number of events happening at the same time) most outcomes fall within a limited range of variation. But statistical prediction only takes us so far because not all outcomes do, and a complex phenomenon will occasionally produce unexpected outcomes. Those unexpected outcomes look random (and may as well be, for the purposes of predicting when they will occur) but they are generated by 100% deterministic processes.

For the purposes of wargaming, the game designer?s task is to define what phenomenon they are trying to depict, at what scale of interaction, figure out how the parts at that scale usually interact, then come up with game mechanics that generate plausible outcomes in an elegant way (which to me means relatively few modifiers and exceptions). Using dice or cards usually achieves that (it proxies for the deterministic interactions at lower levels that influence the outcome - a shot horse disrupting a charge, or a subordinate officer making a really bad choice, etc). But, as I said a few posts earlier in this thread, a fully deterministic set of rules could also do the job. It just might not be worth the additional effort?

BelligerentParrot, I missed your points, must have been typing at the same time! I think your summary is fair (possibly because you have landed on pretty well on similar conclusions    ). It seems like the main point of divergence for game design is how much randomization is a reasonable representation of the impact of unknowable variables in a battle. There?s a pretty wide range!

[SteveBurt]

SteveBurt, that?s not quite accurate about the real world- it definitely is deterministic. If it was not deterministic, chains of cause and effect wouldn?t be a thing, but they are. The problem is that the precise cause and effect chain can only be traced accurately after the events have occurred. The reason that is the case is that nearly all real-world phenomena that happen on scales human beings are interested in during day to day life (including military combat) are outcomes of complex adaptive systems - systems with large numbers of mutually influencing parts. Interactions between parts are completely deterministic (something causes something else) but the final outcome is usually the result of multiple sets of interactions, so knowing how a small part works does not tell us how the whole thing behaves. Statistics help for prediction, because over a large number of repetitions (or a large number of events happening at the same time) most outcomes fall within a limited range of variation. But statistical prediction only takes us so far because not all outcomes do, and a complex phenomenon will occasionally produce unexpected outcomes. Those unexpected outcomes look random (and may as well be, for the purposes of predicting when they will occur) but they are generated by 100% deterministic processes.

For the purposes of wargaming, the game designer?s task is to define what phenomenon they are trying to depict, at what scale of interaction, figure out how the parts at that scale usually interact, then come up with game mechanics that generate plausible outcomes in an elegant way (which to me means relatively few modifiers and exceptions). Using dice or cards usually achieves that (it proxies for the deterministic interactions at lower levels that influence the outcome - a shot horse disrupting a charge, or a subordinate officer making a really bad choice, etc). But, as I said a few posts earlier in this thread, a fully deterministic set of rules could also do the job. It just might not be worth the additional effort?

At the lowest level, down in the quantum realm, reality is probabilistic, not deterministic. You can say that 50% of the atoms in your radioactive sample will have decayed after a certain time. But you can never say, even in principle, when a particular atom will decay. There is no 'cause' which makes it decay at a particular moment, and there is no system of hidden known variables which can reproduce quantum behaviour.
However, that's beside the point. The point is the one you bring up next. Complex systems exhibit chaotic behaviour. The three body problem is a good example, so is the weather, and combat is another. Tiny variations in initial conditions, or events which you can never model, influence the outcomes. So while it might be deterministic in principle, in practice it exhibits chaotic behaviour. 95% (or whatever) of the time the legionaries will withstand the charge of the Gauls, but the other 5% something happens (maybe a centurion dies, an order is misinterpreted, the Gauls are a bit bigger and have more impressive moustaches.
The job of the game designer is to decide what the outcomes should be, and what the probability of each one should be. A lot of that is educated guesswork, plus playtesting. But if the rules say the Romans always withstand the charge, they are not doing their job properly.
Spartans beat other hoplites pretty consistently, until they didn't. What changed? Maybe the Spartans were less good, maybe Epaminondas came up with novel tactics. But the point is that nobody, certainly not the Spartans, and probably not the Thebans either, knew who was going to win. Once the Spartans had been beaten, their mystique was lost. It would certainly not be realistic to let either commander go into battle knowing that his tactics would work for sure. I'm not talking about the evolution of a battle here; hoplite battles were very static affairs until one side broke. In a deterministic set of rules you know exactly what will happen in that first clash. That might be correct for many battles featuring Spartans, but not for all (some of their initial struggles against Argos also seem to have been close run).
Claiming the a deterministic set can reproduce fog of war may have some truth, but clearly a set with chance involved can also do so. Taking chance out doesn't make the rules more realistic.

[TheDaR]

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'd argue that most wargame rule sets, even if otherwise entirely free of randomizers, would not be algorithmically solvable.  The instant you put any sort of movement or terrain placement into the game which does not rely on a strict units of measure on a fixed board (grid, hex, territories, or any other analogue of an undirected graph), the choice of each action becomes potentially infinite and in turn unsolvable.

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.

Agreed.  Again, I think the general issue in this thread is the mistaking of any sort of physical randomization as the only possible source for unpredictability and that it needs to be applied to most or even every possible action to get "realistic" uncertain results.   There are lots of non-mechanical ways to introduce unpredictable results.  Hidden information, bidding, resource management, rock-paper-scissors effects, etc.  There's also lots of ways you can engage with randomness (or lack thereof) in the rules.

Let's do a mini game design exercise showing how big this space can potentially be.


Imagine a game where all combat resolution is deterministic.   It doesn't even have to be complex, simple attack and defense values like a board game with some number of wounds/hit points/morale to reflect durability, just 3 simple stats and a unit type.  We keep the idea of the standard WRG/DBx type tactical factors, such that flanking and herding your opponent into unfavorable ground and how various unit types interact matters, affecting how much "damage" occurs as a result of combat.   But each player also has stack of tokens equal to the number of units with a pair of small modifiers to attack and defense (such as +1/-1, or -2/+1, or 0/+2, etc), with the totals of attack and defense across all tokens summing to 0.  The numerical effect of tokens should be to change how much a combat is won or lost by or nudge very close combats one way or the other, rather than being the primary factor.  Maybe a recoil becomes a rout, a unit that should have broken hangs on by the skin of its teeth in the face of a charge, or a bloody and expensive victory becomes a steamroll of negligible loss.


In our first set of game variations, we'll go with allocating tokens repeatedly, either per combat or at least per turn.  As a result, like in many wargame rule sets, units may vary in capability over the course of the game.  They get a token when needed (either at combat time or the beginning of a turn) and only keep it until their next combat is resolved.  Despite this, one version is still completely randomless, and only one is completely random each time in the style of traditional wargames.

First we play the most traditional wargame version of the game where on every combat resolution, both sides blindly draw one token from their own bag, treating the tokens as pure mechanical randomizers, until the bag is emptied and refilled with the discarded tokens.  The tokens here might as well be dice, and have all the same strengths and weaknesses any such game design would have.  Every combat has some degree of uncertainty and unpredictability due to the tokens, but playing to the fixed tactical factors can mitigate that to a large degree, since the modifiers are fairly small. The use of a set of fixed tokens that are exhausted does at least mean that no one has to suffer the effect of an unrelieved string of very good or very bad draws, every token is seen once per bag refill.  There's a large degree of mechanical randomization here, every combat on every turn sees a new randomization.

Second variation we flip the script completely.  For each combat you can choose any one of your undiscarded tokens to give to a unit until the end of the turn.  You can't reuse any token until you've discarded all of them. Now there's no actual randomness involved at all, but still a great degree of uncertainty; you have no idea which tokens your opponent will chose to allocate.  Maybe he plays his aggressive tokens on one flank and his defensive ones in the center.  Maybe the reverse.  Maybe you have to not use an aggressive token in one combat to shore up the line in order to guarantee an effective unit kill elsewhere.  Every action you have to play the same guessing game of which token will most effectively counter or overcome your opponent's.  Plenty of fog of war, you never know each action if your opponent's units will be especially effective or not. 

Third we do the same as second, but at the turn level.  Each turn allocate one token to each unit until the end of turn.  Same general guessing game, but now you have to anticipate all the potential combats you could potentially get involved in, not just making the best choice as they happen.  Still no mechanical randomness, but high unpredictability because both sides have to guess what the other will do.

Next, we give a similar turn level treatment to random allocation.  At the beginning of each turn we blindly distribute one token to each remaining unit without a token, and only flip it face up if that unit engages in some sort of combat, and then discarding all face up tokens at the end of the turn.  Some tokens will have no effect in a turn, if that unit doesn't get into combat.  There's no knowledge of which token is where until combat occurs, so you can't do too much planning around it.  Every combat is still uncertain, maybe even more so, since some tokens are effectively out of play each turn.  But if a unit does engage in multiple actions (perhaps because they are shot at and then charged, then perform a combat during pursuit), their performance remains consistent for the whole turn, so the total randomization is probably slightly smaller than our first version.

Last is a mixing of the previous two, to reduce but still keep some amount of mechanical randomness.   At the start of the turn each player blind draws tokens one at a time, then looks at and places them face down on a remaining unit without a token, until all units have one.  At the end of the turn any revealed tokens are returned to the unused pile.  In the beginning of the match there's no real randomness as every tokens will end up on a unit, but as units are removed from play, the pool of available tokens is larger than the number of units and some tokens will remain unrevealed on the field, so you can't count on drawing the exact tokens you want for each unit every turn.  You still have fog of war in that you don't even know in advance what tokens are on enemy units, but have knowledge (if not exact control) of your own forces.  This nicely feeds into the idea of commanders slowly losing surety of control as the battlefield grows chaotic.


Now a set of variations where instead of doing token allocations every turn, we do them at deployment time, placing the tokens face down beside the unit.  They stay face down until that unit engages in combat of some sort, at which point they're flipped up and revealed, but stay the same value until the end of the game.

First variation here is to blindly allocate tokens at deployment.  Neither side knows the exact performance of their units until they engage, and only thereafter can they accurately plan around unit capabilities, but said capabilities are consistent.  Here you've got a good representation of many aspects of fog of war in a large mass combat scenario.  Some units are just having good luck or a good day, and others might not be (or if the modifiers aren't symmetrical, maybe they're feeling aggressive but not very disciplined, or stubborn but worn out), but no one knows for sure until it's go time.  It's a fairly small amount of mechanical randomization compared to per turn or per combat allocations, just once at the beginning of the game and then never thereafter, but still a high degree of uncertainty, especially in the early game.

Next variation, each player gets to select a small number of tokens based on the rating of their commander/army and place them where they like, and then every other token is shuffled and placed blindly.  Each general knows how a few of their units are going to do, but a majority of their units are still somewhat uncertain.   Same fog of war applies, you know nothing of your opponent's forces (except maybe by reading on their few units that move with more purpose because the opponent knows what they're doing), and only a bit more about yours.  There's even less mechanical randomization overall, but still more or less the exact same level of uncertainty and unpredictability.

Final variation, each side gets to select which token to place on each unit during deployment.  From the other side of the table the lack of information is exactly the same, you still have no idea to start which of your opponents units have which bonuses or penalties until first engagement, except maybe by divining what they're doing with specific units before they're engaged.  So you've still got the same fog of war effect to nearly the exact same degree, because the lack of information on opponent capabilities is still equal and symmetrical.   But there's now exactly zero mechanical randomization anywhere in the rules again, despite the conditions of combat being highly uncertain, especially until most units have engaged at least once.


So more than a half dozen variation of the rules.  The exact same physical components, the exact same simple deterministic resolution mechanics for combat, with the only difference being how a unit gets their modifier token.   Multiple versions where the randomness/choice is made every turn, multiple where it's done strictly at deployment.  All use hidden information in some form, except the token draw on every combat variation. All of these variations and only really 1 of them is even remotely equivalent to "just roll dice for every action".  Some have a lot of mechanical randomness, some have very little, and some have none.  How, and even if, we choose mechanical randomization ends up mattering quite a bit in terms of game feel, but relatively little in terms of actual uncertainty and unpredictability of play, thanks to the hidden information.

You could mix and match aspects of several of these variations.   Here's just a few random examples I thought of while typing this up.

• Select some number of units to get specific tokens at deployment face down, then for every other unit you have to draw from the bag each turn or even each combat (representing some units which are very consistent, while others are more fickle).
• When placing random/blind tokens, but the player gets to look at them upon placing them (but maybe not after that) so they have knowledge but not control.
• Revealed tokens are turned face down at turn end, as part of player skill testing (did that unit there have +2/-1 or +1/0... it was 3 turns ago, now I don't remember....).
• Tokens are allocated face up so both sides can see.  Maybe only some are face down based on various factors.
• Tokens are only discarded when a unit takes damage to their durability/morale score, otherwise they remain in play (face up or flipped down, depending on other variation rules choices).
• Instead of one per every unit every combat/turn, each side has a limited number of tokens to allocate, or limits to how many can be deployed per turn, leading to some combats with no modifier tokens involved

And this is just the tip of the iceberg in the design space.  All these are variations on one basic core mechanic, hidden information, and the possibility of chosen or random allocation of a fixed set of resources.  I've workshopped similar ideas around using Rock-Paper-Scissors type mechanics, including variations with 4, 5, or 6 "elements", or using cards where you can't replay the same card the next turn, but your hand might be larger than 3 and thus have more than 1 of the same element in it.   Where success of a specific type allows you to activate unit or army level special abilities (and each faction could purchase extra cards with their own faction-specific special abilities as part of their roster).  Or each card gives different numerical bonuses depending on win/loss/draw, rather than being just an absolute win/loss/draw.

You could easily go in a dozen different directions for less-randomized more-deterministic rules without giving up any degree of uncertainty or fog-of-war unpredictability.  Chris Engle's Ritter series of games, mentioned earlier, uses a deterministic combat matrix list, relying on positioning and very limited active actions per turn relative to force sizes.   The Crystal Clans board game has a shared track for "initiative points", and various actions cost different points, and once you've spent into your opponent's side of the track, they take over the turn until they've spent back to your side, in combination with a Rock-Paper-Scissor effects on various unit cards.   Imagine extending this to a wargame where you could buy varying degrees of success if you won the RPS aspect or the reverse, spend points on actions and RPS determines the degree of success.  Either way you have a combination of guessing game and resource management to give you unpredictable results, as you'll never know for sure what your opponent will choose to spend on or which RPS element they'll decide to use for a given action, but neither is truly random, mechanically speaking.

[Bolingar]

@TheDaR: very thorough and very interesting, thank you!

[Bolingar]

The real world is inherently not deterministic. Further, all the factors which affect the outcome of a combat are not just unknown, they are unknowable.
If two units fight twice, the outcome will differ, maybe by only a few percentage points, but maybe not. Sometimes a random event during a combat can have a huge effect; for instance, you might say that a well formed square in a Napoleonic game can never be broken, but then you get instances like a dead horse falling across it and the square being broken. Very unlikely, but it did happen. A deterministic system cannot model such outliers; it assumes that things will be nicely in the middle of the bell curve at all times. That is misleading. Effects do not always average out.

It's true that there are individual events that can substantially affect the outcome of a battle or even determine it: a small force appears over a crest on the enemy's flank, the enemy mistakes it for the arrival of an entire new army and routs in consequence; a keg of gunpowder unexpectedly explodes during a battle, nearby raw recruits panic and run, triggering a general rout, and so on. But - as you say - these events are rare and wargamers don't actually model them: Throw a D100. If you get 99 something spooks your army and it automatically routs. Meh. Dice certainly don't model them, since unexpected events don't substantially affect every single combat of every single unit in every single turn. Wargamers affirm that a typical unit will substantially vary in combat capabilities every 10 or 15 minutes and dice perfectly replicate that. I say bollocks: dice are an integral part of wargaming purely and simply because they are fun: they supply the gambler's thrill: the odds are good but who knows? Will I make it? Won't I make it?....Yes!

BTW very interesting thread chaps, and I appreciate the thought you've put into the posts. 

[Bolingar]

At the lowest level, down in the quantum realm, reality is probabilistic, not deterministic. You can say that 50% of the atoms in your radioactive sample will have decayed after a certain time. But you can never say, even in principle, when a particular atom will decay. There is no 'cause' which makes it decay at a particular moment, and there is no system of hidden known variables which can reproduce quantum behaviour.

That sounds a bit odd: I don't think purely spontaneous behaviour is a thing with matter, though it can be with free will. Does quantum theory allow for a possible causality in decay?

However, that's beside the point. The point is the one you bring up next. Complex systems exhibit chaotic behaviour. The three body problem is a good example, so is the weather, and combat is another. Tiny variations in initial conditions, or events which you can never model, influence the outcomes. So while it might be deterministic in principle, in practice it exhibits chaotic behaviour. 95% (or whatever) of the time the legionaries will withstand the charge of the Gauls, but the other 5% something happens (maybe a centurion dies, an order is misinterpreted, the Gauls are a bit bigger and have more impressive moustaches.

The three body problem is a good example of the incalculability of deterministic causes. Orbital mechanics is deterministic just impossible to predict in the case of three objects orbiting each other. I wish you'd give Optio a try - it has the same unpredictability, surprises all the time. Point is that that unpredictability doesn't require bigger Gauls or bigger moustaches. My contention is that a large enough body of men will behave in a predictable way within narrow parameters, enough to assign determinism to it. It excludes wild and rare events that would make it behave differently, but I think wargamers exclude such events anyway.

The job of the game designer is to decide what the outcomes should be, and what the probability of each one should be. A lot of that is educated guesswork, plus playtesting. But if the rules say the Romans always withstand the charge, they are not doing their job properly.

Why not? Here's the question: wargamers accept that one unit that is seriously superior to another must beat it, every single time. So knights charging skirmisher foot in open terrain must overrun them, every single time. If the skirmisher foot rout the knights then the player will conclude, quite naturally, that there is something wrong with his ruleset. But if decent quality legionaries are notably superior to Gallic warband, why shouldn't they beat them in open terrain, every single time, if no external factor affects their performance? At what point must the inferior unit be near enough in quality to the superior unit to have a chance of beating it? When things are balanced on a knife, with both units virtually equal, then sure, you need something that can tip the balance. In my system commanders affect combat performance and a die is thrown for a commander after each combat. If he gets a 6 twice he is dead and his unit is in trouble. Dice for individuals is fine and that's your balance-tipper.

Spartans beat other hoplites pretty consistently, until they didn't. What changed? Maybe the Spartans were less good, maybe Epaminondas came up with novel tactics.

Exactly. You need to change the numbers for the Spartans to lose. The Theban column drives them back irresistably, not because Thebans are better hoplites, but because the column is a tactic the Spartans hadn't catered for. Or you swop out real Spartan hoplites for inferior hoplites that are Spartan in name only. You have to change something.

But the point is that nobody, certainly not the Spartans, and probably not the Thebans either, knew who was going to win.

Epaminondas knew he was going to win because he had a massive modifier the Spartans had no answer for. The player knows this too: Theban column +2 in melee combat. Nothing random about that.

Once the Spartans had been beaten, their mystique was lost. It would certainly not be realistic to let either commander go into battle knowing that his tactics would work for sure. I'm not talking about the evolution of a battle here; hoplite battles were very static affairs until one side broke. In a deterministic set of rules you know exactly what will happen in that first clash. That might be correct for many battles featuring Spartans, but not for all (some of their initial struggles against Argos also seem to have been close run).
Claiming the a deterministic set can reproduce fog of war may have some truth, but clearly a set with chance involved can also do so. Taking chance out doesn't make the rules more realistic.

Again, to alter the sure and certain outcome of a superior vs inferior unit combat, you have to either convert the superior unit into an inferior one, or impose an external modifier - which is better represented as a fixed combat modifier, not a random dice throw.

[SteveBurt]

That sounds a bit odd: I don't think purely spontaneous behaviour is a thing with matter, though it can be with free will. Does quantum theory allow for a possible causality in decay?The three body problem is a good example of the incalculability of deterministic causes. Orbital mechanics is deterministic just impossible to predict in the case of three objects orbiting each other. I wish you'd give Optio a try - it has the same unpredictability, surprises all the time. Point is that that unpredictability doesn't require bigger Gauls or bigger moustaches. My contention is that a large enough body of men will behave in a predictable way within narrow parameters, enough to assign determinism to it. It excludes wild and rare events that would make it behave differently, but I think wargamers exclude such events anyway.Why not? Here's the question: wargamers accept that one unit that is seriously superior to another must beat it, every single time. So knights charging skirmisher foot in open terrain must overrun them, every single time. If the skirmisher foot rout the knights then the player will conclude, quite naturally, that there is something wrong with his ruleset. But if decent quality legionaries are notably superior to Gallic warband, why shouldn't they beat them in open terrain, every single time, if no external factor affects their performance? At what point must the inferior unit be near enough in quality to the superior unit to have a chance of beating it? When things are balanced on a knife, with both units virtually equal, then sure, you need something that can tip the balance. In my system commanders affect combat performance and a die is thrown for a commander after each combat. If he gets a 6 twice he is dead and his unit is in trouble. Dice for individuals is fine and that's your balance-tipper.Exactly. You need to change the numbers for the Spartans to lose. The Theban column drives them back irresistably, not because Thebans are better hoplites, but because the column is a tactic the Spartans hadn't catered for. Or you swop out real Spartan hoplites for inferior hoplites that are Spartan in name only. You have to change something.Epaminondas knew he was going to win because he had a massive modifier the Spartans had no answer for. The player knows this too: Theban column +2 in melee combat. Nothing random about that.Again, to alter the sure and certain outcome of a superior vs inferior unit combat, you have to either convert the superior unit into an inferior one, or impose an external modifier - which is better represented as a fixed combat modifier, not a random dice throw.

This has been a very interesting discussion.
I don't agree that wargames don't allow for unlikely events. I can think of plenty of rules where things with a very low probability do occur occasionally, even if most of the time the results are predictable. In DBA, if I throw a 6 and you throw a 1 (only a 1 in 36 chance) we may well see an outlier result. In a system using d10, a 1/10 split is a 1% chance. Likewise, 'buckets of dice' systems will mostly give results in the middle of the curve, but every once in a while you'll throw 4 6s or whatever. Very unlikely things will happen from time to time. Not too often; that would be neither realistic nor fun to play.

My problem with the Theban/Spartan fight I think goes to the heart of our disagreement. You say that the Theban commander knows he has a +2. Of course, the Spartan commander knows that too. Because there is no chance involved, we know what the outcome will be.
But in the real battle, neither side knew what the outcome would be. You have given the commanders too much information.
My contention is that some randomness better models that lack of information than a fixed system. The Theban commander may know he has an edge, but he can't be sure of what will happen until the fight has started. If the Spartan knows that his unit may survive the initial impact of the Sacred Band for long enough, he can hope to win the fight with the rest of his line, but if he knows what will happen, he will not engage, and the historical battle will never happen.

I'm unlikely to try Optio not because it is diceless, but because in your shots of battles I see each unit carrying around a little tray of markers. Way too much book-keeping for my taste, I'm afraid. To the Strongest is more my thing.

[boneio]

I think we're perhaps in danger of going in circles and certainly, as can easily happen in debates about complex topics, some of us are arguing slightly different points to one another, I agree with whoever pointed that out 

Also echoing both above that it's an interesting conversation

Just to attempt pr?cis again:

• Modelling fog of war can be more realistic with a diceless system than one with some manner of randomisation - my position has shifted a little following the examples given recently, I feel it is reasonable to say that a complex deterministic decision space can be in practical terms just as (or more) 'unpredictable' feeling to the players, but I still think it doesn't meet the definition of being a more realistic simulation in strict terms. I can agree that it certainly can reasonably simulate the experience of being a commanding general as long as hidden information is a key mechanism.
• Combat outcomes can be statistically predictable enough that removing randomness of outcome is OK - in a purely game sense, of course that's down to preference and many games play well in this manner; in terms of real-life engagement modelling I am never going to agree, the variables are too complex and nobody repeats the same or similar enough battles so we'll never really know, other than looking to general statistical modelling etc.
• Trying to not incorrectly paraphrase Bolingar - dice are only used because gamers like the risk/reward thrill - strongly disagree, dice are used in many different ways to model unknown outcomes and to make the game be not just a battle of will vs will (as chess is, really)
• Again trying to not incorrectly paraphrase Bolingar - the use of randomness can result in uncomfortably jarring outcomes - yes, but that's down to the system not the principle of randomness. As someone else pointed out, the probablity curves are often a lot flatter than people intuitively think, GW games (especially current ones) being a great example. Although they are still often hinging on some key rolls, so the point here isn't to prove/disprove the assertion but to highlight it's a scale from very random to only slightly random

As an aside it seems that we don't actually know if randomness is real - quantum theory says yes, but some sub-field of quantum theory so no, we just don't know all the variables yet  I discourage thinking too deeply about it as you may wind up convincing yourself that free will isn't real and that you as a person will always take exactly the same actions under exactly the same circumstances etc  More spiritual people may have a way out of that thought trap regardless 

[Belligerentparrot]

Very nicely put, Boneio!  I don't have anything to add other than this:

I discourage thinking too deeply about it as you may wind up convincing yourself that free will isn't real and that you as a person will always take exactly the same actions under exactly the same circumstances etc 

Honestly, if there is no free will then the idea of convincing yourself of anything also goes out the window, doesn't it? (Convincing yourself is a judgement, and so presupposes the spontaneity of reason, which determinism seems to foreclose. Sorry, this is a rare moment where LAF converges with my day job - philosopher - so I am legitimately working in replying to this  )

[Storm Wolf]

I think we're perhaps in danger of going in circles and certainly, as can easily happen in debates about complex topics, some of us are arguing slightly different points to one another, I agree with whoever pointed that out 

Quite

https://en.wikipedia.org/wiki/Oozlum_bird

[boneio]

Very nicely put, Boneio!  I don't have anything to add other than this:
Honestly, if there is no free will then the idea of convincing yourself of anything also goes out the window, doesn't it? (Convincing yourself is a judgement, and so presupposes the spontaneity of reason, which determinism seems to foreclose. Sorry, this is a rare moment where LAF converges with my day job - philosopher - so I am legitimately working in replying to this  )

I think you may have won my award for 'most awesome job on LAF' 

For what it's worth (which won't be much to a professional such as yourself, I'm sure!), my current belief is that consciousness is simply one becoming aware of the decision, the decision actually having happened by either a minute random event which then cascades into being perceived as a choice (why did I put the glass down at that exact moment rather than one just before or just after, those moments being equal insofar as a conscious choice goes...?), or being effectively predetermined and inevitable.
I am quite concerned that the growth of 'AI', LLMs, neural networks etc will lead us to understand that the brain and consciousness is simply an incredibly complicated but deterministic machine that only appears to possess free will because the variables are so immense and rapidly processed.
But... I haven't thought too much about it because it's a bit depressing and I'm distantly aware there are some quite nice contra-positions.   

[boneio]

Quite

https://en.wikipedia.org/wiki/Oozlum_bird

This is new to me... I resemble this remark 

[Storm Wolf]

This is new to me... I resemble this remark 

LOL sorry I couldn't resist, if you are old enough to remember "Carry on up the Jungle" then you may remember the expedition to capture the Oozlum bird