1 unit per hex: failed experiment

[tsoky]

The "1 unit per hex" thing would work if a hex actually represented space that 1 regiment (say, division) occupies. But that's not the case, a megapolis irl can hold several hundred divisions on its territory. And if you think of 1 unit as an army group, then you should be able to combine all kinds of soldier types within a single unit - have you seen a 100.000 army consisting only of archers? So if a city occupied 1 hex when just founded and 100 hexes (10x10) when developed into a big city, and if it could work a radius of 50 hexes in each direction, and cities would have to be placed 80-100 hexes from each other, then the 1 unit per hex system would work. But that would be a tactical game, while civilization is and has always been a strategical game.

Also if we want to stick to realism, battles in ancient times and in modern times are different. Battles in ancient times were fought just like in civ 4 - two countries build a stack of units (two 100.000 armies), those armies meet on a field of battle (a single tile) and one army wins, the other loses. The war doesn't cover a huge territory, battle is fought on small clusters oа land. Modern warfare, on the contrary, is about front lines stretched across whole continents, which is what they tried to achieve in civ V. But still those front lines do not work with 1 unit per hex system. They need to put a limitation of 3-5 units per tile/hex, as in reality a single part of front is always protected by several types of troops - say, infantry, some tanks, some anti-tank inf, some paratroopers, etc. No general would operate a front line where first 100 km are defended by infantry, next 100 km by cavalry etc.

 

[Deep_Blue]

Stack of Dooms don't make sense because they don't represent combat as it occurred in the history of civilization. And I'm not referring to grognardy details like attrition either but basic stuff like a smaller army beating a larger one when both are at technological parity, the power of defense, blocking off territory, evasive delaying tactics (withdraw sort of does this but unless you kill the stack in the same turn your troops are dead since they can't withdraw in defense and not all troops get it) Under the Civ IV BTS model Persia would beat the Greek City States every time. There would never be a Thermopylae.

actually SODs represents history accurately, USA stacked 300'000 soldiers in a small part in Kuwait before they invaded Iraq.. there are countless examples of this.

now if you are talking about combat mechanisim and realisim then this has nothing to do with the basic idea of stacks, the broken mechanisim is not because of stacks but because of how each unit attack and defend in relation to all variables included, in short you can change the mechanisim to something more realistic and keep stacks concept in the same time.

 

[Doctor Phibes]

This thread has gotten rather long, though I have read most of it. I just wondered if anyone had made the point that the movement hex matrix is currently firmly coupled to the build hex matrix. I've taken on board the point that several posters have made that increasing the granularity of the map would help with the traffic problems and help fix the broken AI, so seems to me one solution would be to decouple things - keep the build matrix just as it is, but subdivide each build hexagon into 7 movement hexagons. In effect this looks like a 7upt stack at the build level, but is 1upt at the movement level.

All sorts of changes have to be made to game rules for this, including how roads work (since roads are 1:1 with build hexagons right now, a close up of a build hex with a road would have to show 6 radiating roads, which is a bit inelegant).

I can see how battles might be fought on a tactical screen on this, since the maximum range of a ranged unit is only 3 hexes (I'm assuming this range will now be in movement hexes) an adequate tactical zoom-in should be one build hex and the 6 build hexes surrounding it.

I don't know if this is remotely workable, just thought I'd throw it in. (Should have worked through some diagrams, really.)

 

[WarKirby]

I was always a fan of the total realism mod for Civ IV, wherein stacking too many units on a tile gave penalties to everything on that tile. Imo, it was far from harsh enough, I recall it started at >20 units and gave a 5% penalty for each after that. But the concept was great

 

[Kruelgor]

I do agree that 1 unit per hex is not good. 3 units per hex sounds better.

 

[niall78]

SOD rule on the strategic level. They fairly represent how battles are fought at this scale, not saying they are perfect but they fit the scale of the game. 1 unit per hex is great in a lot of tactical level games, I loved and still play a lot of John Tiller games where it works great. Civ game hexes/squares represent 100's of square Km, I find 1upt at this scale to be a complete break in immersion, I can't re-enact D-Day, Hastings, Waterloo any more, I'm stuck playing.....I'm not sure what I'm playing.

SOD represented army/army groups in earlier Civs. I have no clue what the modern 1upt units represent. I can't get my head around how an invasion like D-Day now has to hit the whole coast of northern Europe instead of a very limited space like Normandy. The new system completely disregards basic military principles like concentration of force at the strategic level. Is this an attempt to move the game into a Total War Light niche?

 

[Vertico]

Not sure I understood you correctly, but if you're talking about optimizing AI routines for pre-drawn maps, then this would probably be at odds with modding. Soren Johnson once explained that the Civ4 AI could have been better in playing the base game if it had been optimized for it, but that would have come at the cost of AI adaptablity for modded content, which they didn't want to do. Imho it's been one of strengths of Civ4 that the AI is not hardwired to work into the rules environment of the base game, and therefore can achieve decent results even when confronted with content that the developers never thought of.

(Side note: Speaking of Titan, there's also a free computer version named Colossus. The AI actually plays pretty decent tactically, though it mostly fails strategically.)

This is a good argument Psyringe. I was thinking about pre-drawn tactical maps as well. For such maps you could imagine a genius AI. Somebody played chess recently on computer? You should probably noticed that there is not "AI" in chess anymore, no algortims, just pure statistics for every move collected from real game experiences of human players and stored with big databases. Then playing against the computer you have no chance, because for of every your move, computer will find thousands of its moves which will lead to your final defeat.

Pre-drawn maps would work it the same way as chess - as the enviroment won't change everytime like on strategic maps. All possible battle moves could be easily quantified and chosen the best for the final win - based on statistics. They use Steam, so they always could improve AI even more adding billions of new potential moves used by players. So finally we will have the situation when a human player will need to have bonuses to win, if forces would be equal But of course battle would depend of engaged forces on strategic map, so it would never be so dermined, not talking abour randomness of fights.

But comming back to what Psyringe mentioned, AI will potentially have big problems of new maps prepared by modders. The option to solve this problem, could be either to prepare many enough pre-drawn tactical maps, so modders won't need to prepare even more, or to prepare some kind of tools to teach AI deal with new maps.
Modders could play new maps as many times as they will need to colect enough experience for AI to let it learn how to play new maps when mod will be released.

 

[FatCatAttack]

Under Civ4 BTS Thermopylae would be a hill (perhaps fortified), on which several highly promoted units dug themselves in (Spartan battle proficiency), and against which Persia would throw a huge stack of unpromoted units in a failed attempt to win by strength of numbers. Add in some better experienced mounted units on the Persian side which get ripped by Spartan spearmen. Perfectly feasible in Civ4 imho. Of course, players would complain about the stupidity of the Persian AI.

No not feasible at all. Firstly for a hill to be bothered to be taken you'd have to get specifically lucky with a mountain range with ONE opening covered by a hill. Otherwise as I said in my first post: simply walk past it. Secondly you're trying to directly translate the battle instead of interpreting the war in terms of Civ. Persia at that point would probably have catapults. Once siege enters in play defensive strats especially ones involving a small stack take a nose dive. A few suicide cats and the Persian army, which wouldn't be unpromoted either, takes the hill possibly with no standard military casualties. In a one unit per tile system your Thermopylae's wouldn't be restricted to such a specific quirk of the map. You get to make more natural use of wider chokepoints because each unit is a wall. You can have a Thermopylae at the neck of a peninsula for example.

On the contrary, if we really dig deep into details we'd find that the example you used to refute my point only proves it. Under the 1upt paradigm, Xerxes would have been hard put to mass his army and drive them into a single choke point. Units cannot co-mingle and share a tile, a fortiori, cannot even maneuver about a reasonable space without exploding or disintegrating first, like matter and anti-matter meeting at a point, because of the 1upt rule. Unless of course theres a combat overseer who would see to it that archers do not step over a certain line or that other units do not violate the 1upt rule.

But I digress. Back on point, to simulate the wave of melee attacks of the Persian army against the Spartans, turn for turn, would also necessarily mean that each defeated regiment would have to pull back after an attack to make way for the next wave. Remember that each turn in civilization spans several hundreds of years. With each movement point used to maneuver the troops under a 1upt rubric the battle of thermopylae would necessarily have dragged on for a millennium. Under the SOD rubric, the conflict would have been resolved in a single turn which can be taken to mean a period of a month, a week or even a day.

Next, like what the above posters mentioned, a highly promoted Phalanx (incidentally, a greek unique unit---how cnvenient), with combat I, guerrilla I, and cover I (assuming barracks and a positive random event for the last two) on top of a hill and fortified for a few turns, can effectively hold its own against an SOD composition of archers, spearmen, axemen and a number of chariots before dying. If we attach a general (Leonidas), we get additional promotions that would make such scenario reminiscent of the 300 movie. As if these examples arent enough to prove my point, Legends of Revolutions actually have a world unit wonder, the 300, which can do all these things in simple fashion.

So you see, in terms of historical accuracy or at least verisimilitude, the SOD comes close to simulate your example and not the 1upt.

The time scale is always going to be wonky. Even if you abstract it in your mind as the battle taking place in a specific time in a series of years covered by one turn it doesn't cover the rest war or the lead up to war which takes years and years and the big Persian stack continues to move. Thermopylae is not Thermopylae if it doesn't delay anything. Also the sheer weight of the Persian stacks would win the rest of the war. Plataea you see wouldn't be possible either.

To be true to history you have to make combat more than two economies spraying hammers at each other which is what a SoD is essentially. You have to be able to control the battlefield somehow to leverage your force in a variety of ways. SoD's can't exert that control because they can't make use of the terrain to block anything.

actually SODs represents history accurately, USA stacked 300'000 soldiers in a small part in Kuwait before they invaded Iraq.. there are countless examples of this.

now if you are talking about combat mechanisim and realisim then this has nothing to do with the basic idea of stacks, the broken mechanisim is not because of stacks but because of how each unit attack and defend in relation to all variables included, in short you can change the mechanisim to something more realistic and keep stacks concept in the same time.

A single Mech Infantry on a hex could represent 300,000 soldiers. The amount of soldiers is irrelevant. The flaw of Stack of Doom's in representing history has nothing to do with issues of Scale but to do with how military power is used.

 

[Doctor Phibes]

Somebody played chess recently on computer? You should probably noticed that there is not "AI" in chess anymore, no algortims, just pure statistics for every move collected from real game experiences of human players and stored with big databases. Then playing against the computer you have no chance, because for of every your move, computer will find thousands of its moves which will lead to your final defeat.

Nope. This seems to get trotted out a lot but it's completely untrue. There are estimated to be approx 10^50^50 possible chess games, so it would not be exactly practical for a start. There are two areas where databases are used, just to optimise matters - in the opening game and the end game. The 'opening book' contains all the stock responses for the classic chess openings, but once the play departs from the opening line (which could be one move in), the computer player must resort to look-ahead calculations. Because of the combinatorial explosion involved in chess, there is necessarily a limit to the distance into the game that an opening book extends - no different in its way to chess textbooks, that simply tail off once the tree gets too wide.

End game databases just contain lots of set pieces that tell the computer what to do with, say, K & N vs K & B - or how to quickly finish with K & Q vs K, setpieces - stops the play getting embarrassing and boring the human.

In between those, there's no database, it's all down to how good the algorithms are. 'AI' is a vexed term, but this is certainly more of an AI than anything in any Civ, which is really just a collection of heuristics.

 

[MeteorPunch]

This thread has gotten rather long, though I have read most of it. I just wondered if anyone had made the point that the movement hex matrix is currently firmly coupled to the build hex matrix. I've taken on board the point that several posters have made that increasing the granularity of the map would help with the traffic problems and help fix the broken AI, so seems to me one solution would be to decouple things - keep the build matrix just as it is, but subdivide each build hexagon into 7 movement hexagons. In effect this looks like a 7upt stack at the build level, but is 1upt at the movement level.

All sorts of changes have to be made to game rules for this, including how roads work (since roads are 1:1 with build hexagons right now, a close up of a build hex with a road would have to show 6 radiating roads, which is a bit inelegant).

I can see how battles might be fought on a tactical screen on this, since the maximum range of a ranged unit is only 3 hexes (I'm assuming this range will now be in movement hexes) an adequate tactical zoom-in should be one build hex and the 6 build hexes surrounding it.

I don't know if this is remotely workable, just thought I'd throw it in. (Should have worked through some diagrams, really.)

I've thought of this solution as well, and I think it's good one, but small stacks (3-5) is just easier to do. Also if Firaxis had designed the game from the ground up with smaller hexes, or the option for smaller hexes, that would've been nice as well.

 

[katipunero]

No not feasible at all. Firstly for a hill to be bothered to be taken you'd have to get specifically lucky with a mountain range with ONE opening covered by a hill. Otherwise as I said in my first post: simply walk past it. Secondly you're trying to directly translate the battle instead of interpreting the war in terms of Civ. Persia at that point would probably have catapults. Once siege enters in play defensive strats especially ones involving a small stack take a nose dive. A few suicide cats and the Persian army, which wouldn't be unpromoted either, takes the hill possibly with no standard military casualties. In a one unit per tile system your Thermopylae's wouldn't be restricted to such a specific quirk of the map. You get to make more natural use of wider chokepoints because each unit is a wall. You can have a Thermopylae at the neck of a peninsula for example.



The time scale is always going to be wonky. Even if you abstract it in your mind as the battle taking place in a specific time in a series of years covered by one turn it doesn't cover the rest war or the lead up to war which takes years and years and the big Persian stack continues to move. Thermopylae is not Thermopylae if it doesn't delay anything. Also the sheer weight of the Persian stacks would win the rest of the war. Plataea you see wouldn't be possible either.

To be true to history you have to make combat more than two economies spraying hammers at each other which is what a SoD is essentially. You have to be able to control the battlefield somehow to leverage your force in a variety of ways. SoD's can't exert that control because they can't make use of the terrain to block anything.



A single Mech Infantry on a hex could represent 300,000 soldiers. The amount of soldiers is irrelevant. The flaw of Stack of Doom's in representing history has nothing to do with issues of Scale but to do with how military power is used.

Why do i get this feeling that you've not actually answered any of the issues I or anyone you quoted raised?

Mech Infantry? Since when, (oh when!), was it a good idea to have a military invading force composed of a single type of unit? He talked about Kuwait, surely there must have been a mix of choppers, ground troups, tanks, personnel carriers, artillery massed at a single point?

and your repy to Psyringe, shoots everyways like a bag of . hitting the fan. You brought up the power of catapults and why it would do short work of any resisting army and how. this is true for all civ games you know.

ugh. why don't you start up a game in V, and show me how your version of "one last stand" battles in history works and looks in V??

please. civ 5 is trying too hard to pretend what it is not, and in the process upended some core game concepts to accommodate this childish, shallow and pointless hankering for 1upt and tactics by JS. I call BS.

 

[FatCatAttack]

Why do i get this feeling that you've not actually answered any of the issues I or anyone you quoted raised?

Actually both you and the other guy did not address the other things I brought up in my first post choosing instead to cherry pick.

Mech Infantry? Since when, (oh when!), was it a good idea to have a military invading force composed of a single type of unit? He talked about Kuwait, surely there must have been a mix of choppers, ground troups, tanks, personnel carriers, artillery massed at a single point?

Civ is highly abstract. A Mech Infantry can represent all those things. It's abilities however are that of it's symbolical representation. You don't think an artillery unit in Civ IV is literally a horde of ghost cannons trundling down the landscape do you?

and your repy to Psyringe, shoots everyways like a bag of . hitting the fan. You brought up the power of catapults and why it would do short work of any resisting army and how. this is true for all civ games you know.

It was handled more elegantly in 2 however (Never played Civ III). The Stack of Doom in IV makes this a huge problem. Since the advantage goes to whoever vomits their siege onto the other stack first combined with the turn based nature of game, you get a situation where if the computer wasn't dumb as a box of rocks two equal stacks would be permanently deadlocked with one another as no one wants to make the first move towards the other stack. Because then on the next turn the stationary stack would vomit it's siege first and crush the other stack.

ugh. why don't you start up a game in V, and show me how your version of "one last stand" battles in history works and looks in V??

please. civ 5 is trying too hard to pretend what it is not, and in the process upended some core game concepts to accommodate this childish, shallow and pointless hankering for 1upt and tactics by JS. I call BS.

You did a good job of describing basically what it would look like earlier. Bulky armies should have difficulty squeezing into places like that.

Civ V has a much more solid design base than IV. In a lot of ways Civ IV fits the mold of "failed experiment" since everything the devs did to try to mitigate the Stack of Doom just make the stack of doom more important. It was pretty clear from the beginning with the adding of Unit XP abilities that they wanted Civ IV to play like V. It's only that way in the early game where you nurture your woodsman warrior bopping animals and whatnot. Units lose their individual character the further you get in the game.

 

[Psyringe]

No not feasible at all. Firstly for a hill to be bothered to be taken you'd have to get specifically lucky with a mountain range with ONE opening covered by a hill. Otherwise as I said in my first post: simply walk past it. Secondly you're trying to directly translate the battle instead of interpreting the war in terms of Civ. Persia at that point would probably have catapults.

You made an analogy that worked to the opposite of what you wanted to prove, and now try to save it by redefining the context. I don't think this is a worthwhile venue of discussion, since obviously every failed analogy in every discussion can be "saved" this way. If you feel that pulling Persian catapults out of your hat adds something worthwhile to your original argument, fine, I won't debate it.

In a lot of ways Civ IV fits the mold of "failed experiment" since everything the devs did to try to mitigate the Stack of Doom just make the stack of doom more important. It was pretty clear from the beginning with the adding of Unit XP abilities that they wanted Civ IV to play like V.

Who told you that Civ4 tried to "mitigate" SoDs? If the Civ4 design team had wanted to do that, then there'd be lots of possibilities (stronger stack counters like collateral damage, give units in a stack negative modifiers, implement higher logistics costs for large stacks, implement a chance that a non-optimal defender is chosen which increases with stack size, etc.). Mitigating stacks is easy. The problem is whether the AI can grasp the rules.

Civ4 chose the route to make large stacks the best option most of the time, and Soren wrote an AI that coped with them pretty well (assemble stacks, move and attack with them in a coordinated fashion, transport stacks across oceans, etc.). It's a design decision that paid off with an AI that actually poses a challenge for new players. Calling it a "failed experiment" by supposing that the devs wanted to mitigate a feature that you don't like is a bit of a stretch, imho.

Conversely, the design decisions for Civ5 were apparently to have a more tactical combat system and write the AI as an afterthought. This only way the current state of the implementation couldn't be seen as "failed" is when we assume that the devs didn't even care about making the AI competitive. But that's in contradiction to everything they said before.

It's interesting that Civ4's lead designer was also its AI programmer, which means that Soren could design the game's rules so that the AI could have a decent grasp on them. In Civ5, the AI designer is just one of many guys working under the lead designer, and apparently either he didn't object enough to a rules system that would be extremely difficult to teach to an AI, or wasn't listen to, or overestimated his abilities and agreed to it.

 

[Jediron]

Civ is highly abstract. A Mech Infantry can represent all those things. It's abilities however are that of it's symbolical representation. You don't think an artillery unit in Civ IV is literally a horde of ghost cannons trundling down the landscape do you?

what people do to defend their opinion.

Oke, i go along your way. I "imagine" that that cannon, does not represent 100% cannon regiment, but also some additional forces. Cool, nothing wrong with that.

But wait; lets take a Knight, for example. Following your path, i should imagine that that Knight is not made of 100% Knights, but also some additional forces. How come then, that those additional forces, can keep up with the Knight, who, apperently, travel much faster.

You see, i try hard. But i smell some inconsistanties in your reasoning. Now, whatever you answer will be; i bet that i have to imagine alot more

hoe about those great "exp upgrades" you can give your units. You have to imagine alot here also.
Now remember, the dev's have implemented "a PG style of making war" here. And may i add; poorly executed.
Oke, the horsemen and it's "exp upgrades". One of them makes your Horsemen runs faster (=more hex). Why ?
Did the horses get some new diet? Macnoodles ? And there are more "unrealist" upgrades. Very Gamey.

And you know what, it all would not matter that much; if only people didn't make so much fuzz about politics!
Cause that's where people DO complain about realism alot. If only they were so critical with the units they wage war with; i would be a happier man.

PG have updates too, but they were much closer to reality (until PG 3 came out, with indeed, also over the top goodies).

 

[falconne]

I made thread on the FFH forums about the plausibility of using the info screens to emulate a separate 2D isometric tactical layer for battles, making it the similar to AoW and MoM games: http://forums.civfanatics.com/showthread.php?t=388746 but I don't think much of the Civ community is in favour of tactical battles. Which is kind of ironic because implementing tactical battles on a strategy layer is much worse micromanagement and off scale.

 

[Jediron]

I made thread on the FFH forums about the plausibility of using the info screens to emulate a separate tactical layer for battles, making it the similar to AoW and MoM games: http://forums.civfanatics.com/showthread.php?t=388746 but I don't think much of the Civ community is in favour of tactical battles. Which is kind of ironic because implementing tactical battles on a strategy layer is much worse micromanagement and off scale.

I am all ears

 

[jpinard]

The biggest thing people don't think about is scale. # of hex's vs. # of units. This game currently scales terribly when you up the difficulty or the # of Civs playing. 1 unit per hex works fine for a great strategic AI with a decent sized map. But if you took some of the scenarios in Panzer General, added 10 more sides to the battle and cut the space in half you end up with a big problem like we see now.

This needs a major amount of critical thinking to fix and make right. If the AI can't be up to the job in its current format, then more units per hex should be considered, or some other stricture that keeps the game balanced/fun. Arbitrary limits on units isn't always the answer either.

 

[Jediron]

Yes i know and agree with you. But what does that for the playability. Already the game is /gets a hog while playing.

 

[Jolly Rogerer]

Actually both you and the other guy did not address the other things I brought up in my first post choosing instead to cherry pick.



Civ is highly abstract. A Mech Infantry can represent all those things. It's abilities however are that of it's symbolical representation. You don't think an artillery unit in Civ IV is literally a horde of ghost cannons trundling down the landscape do you?



It was handled more elegantly in 2 however (Never played Civ III). The Stack of Doom in IV makes this a huge problem. Since the advantage goes to whoever vomits their siege onto the other stack first combined with the turn based nature of game, you get a situation where if the computer wasn't dumb as a box of rocks two equal stacks would be permanently deadlocked with one another as no one wants to make the first move towards the other stack. Because then on the next turn the stationary stack would vomit it's siege first and crush the other stack.



You did a good job of describing basically what it would look like earlier. Bulky armies should have difficulty squeezing into places like that.

Civ V has a much more solid design base than IV. In a lot of ways Civ IV fits the mold of "failed experiment" since everything the devs did to try to mitigate the Stack of Doom just make the stack of doom more important. It was pretty clear from the beginning with the adding of Unit XP abilities that they wanted Civ IV to play like V. It's only that way in the early game where you nurture your woodsman warrior bopping animals and whatnot. Units lose their individual character the further you get in the game.

Mech infantry has organic artillery and armor (as well as infantry) at the brigade level (@ 5k troops). Yet for some reason it was decided back in 1990 to have independent groups of X or Y unit running around and occupying huge areas as if concentration of force and combined arms don't apply. It's a flaw in the entire Civ series that a strategic game tries to artificially shoehorn tactical units on to the strategic map when they never operated like that in reality.

Sure we can pretend that that a mech infantry unit is an army or army group of hundreds of thousands. But why then does it have the strengths and weaknesses of only mech infantry? Artillery units are even more ridiculous in this regard.

Civ 4 was the only game in the series that got the strategic map part of the equation at least partly right. Its downfall was that combat was still a series of individual matchups fought sequentially (with a massive defender advantage) rather than one larger battle where the terrain, numbers and the composition of the forces could be more realistically simulated. The catapult "fix" was a reaction to this initial poor design decision, and completely unrealistic in addition to not being fun.

Civ 5 just doubles down on the flaw in the designs of Civ 1-3. The result is a terrible simulation that sees the AI eager to be defeated in detail over and over again. It will not be a simple matter to fix the AI. At least in Civ 4 you had to react on the strategic level to AI invasions by building troops etc. Even when you won there was a cost in terms of units lost and opportunity cost as you built reinforcements and replacements. In Civ 5 your ever victorious standing army just has to move the afflicted area and kill. There is no impact on your larger economy.

 

[Jediron]

Finally, someone who can put it better to words then i do.
Good job

And i like to add:
http://forums.civfanatics.com/showpost.php?p=9770151&postcount=9
"this is not the way i intended to play it"....nuff said.