Please use hexagons instead of squares!

[Truronian]

This is a 4x triangalized icosahedron and a 18x triangulization of an icosahedron.
I've made the pentagons visible, as you can see, there are 12 in the same place as they are on a soccer ball (or football, whatever)
I think you can see the potential of a highel level of triangulation.
A fitting icosahedron is possible with any triangilization of a pair number (2, 4, 6, ... 142, ...)

I didn't make the hexagons in the 18x one visible though, I ain't got THAT much time. But I think you can see what happens if you take 50x ... the triangles almost flatten out and hexagons appear.

Thanks, I see how this could work. I still don't like the idea of multiple tile shapes, although if you want to go down the spherical route this would seem to be the way to go.

If you can't explain why, then you're just some guy making a wild assertion. Besides, I doubt you have the data to reach that conclusion with any degree of confidence.

Its not an assertion, its an opinion. An opinion that tends to be shared by many (at least those to whom I talk)

That's not completely wrong, just mostly wrong. Straight lines are the exception. Before the industrial age, there were few straight lines in cities or borders. What mattered was easy of transport, and rivers, roads, and topography shaped that. Buildings are irrelevant because they operate on a completely different scale.

Read post above yours

 

[Atrebates]

I don't see any particular reason to favour squares or hexagons, hexagons look cool and represent movement well (but the game is not a simulation, so this is just a nice thing to have) but take perpendicular tangents to any movement and you can't go in a straight line, making your lines of communication a bit odd.

Squares work well enough, they're just a bit boring.

Any system that introduces different tiles (e.g.Octagons + squares) would open a nasty can of strategic issues for cities etc...

PS. I discount arguments as to certain 2D nets convert better into 3D worlds, you only use a 2D net in the game dammit, even if a 'globe' would be very classy. More importantly any 2D net converts to 3D (the angles distort so that your squares no longer have 4 right angles and the sum of their angles is in fact more than 360) the problem is that the 2D object as a whole is a rectangle.

 

[frekk]

4)Tactically, I think hex based games are weeker. I cant really explain why, but squares tend to lead to more strategy

Really. That's interesting, considering all the meticulously realistic games of the wargame genre (from Avalon Hill boardgames to things like Steel Panthers) are, almost without exception, hex-based.

It's only the abstract, simple games that don't aim for high realism that still use squares.

I take it you've never played anything that looks like this:

Hex format has been universally favoured for tactical wargames for, oh, about 30 years now ...

Even the military preferred it until computers got advanced enough they could do away with grids altogether ...

The Difference Between Military & Civilian Wargames
by Michael K. Robel
June 18, 2004

... The U.S. Army recognizes three types of simulations. The first are Constructive Simulations. These are much like typical civilian wargames with a helicopter like view of the battlefield and the older ones actually used hexes to control movement. ...

http://www.strategypage.com/dls/articles/2004618.asp

 

[apatheist]

Its not an assertion, its an opinion. An opinion that tends to be shared by many (at least those to whom I talk)

Which is irrelevant. Piling more people onto a subjective opinion doesn't make it objective. You're asserting one thing is better than the other and expecting people to be convinced.

Read post above yours

I read it. Several of the ones above mine, in fact, and saw nothing any less arbitrary or baseless..

I don't see any particular reason to favour squares or hexagons, hexagons look cool and represent movement well (but the game is not a simulation, so this is just a nice thing to have) but take perpendicular tangents to any movement and you can't go in a straight line, making your lines of communication a bit odd.

...

PS. I discount arguments as to certain 2D nets convert better into 3D worlds, you only use a 2D net in the game dammit, even if a 'globe' would be very classy. More importantly any 2D net converts to 3D (the angles distort so that your squares no longer have 4 right angles and the sum of their angles is in fact more than 360) the problem is that the 2D object as a whole is a rectangle.

I'm not sure of your meaning. If you think a spherical world wouldn't add anything, well, that's what you think. I like the idea myself, because it avoids the ugliness of map edges and it's more realistic, but I don't expect everyone to value those things the same way. However, if you think all tiling systems can be converted to a sphere, you are mistaken. Even if you use irregular shapes, any tiling of only quadrilateral shapes cannot be converted into a spheroid (assuming a cube is not considered a spheroid, of course). Forget about physical shapes and think of it simply in terms of graph theory. You cannot construct an analogous graph if all nodes are connected to exactly 4 other nodes.

 

[warpstorm]

Even the military preferred it until computers got advanced enough they could do away with grids altogether ...

The military did away with hexes for computer sims about the time computers became readily available. I don't remember one sim that used hexes in the last 20 years (not to say there weren't any, but none I've been involved with). A fair number gridded the terrain up into squares, though (but did not use the squares themselves for movement). Like you said, most have done away with grids for movement.

One thing I will say on this, there hasn't been a hex-based game that broke the 100,000 sales mark in ages (the Civ franchise would end if it only sold 100,000 copies in an installation). Not to say that it can't happen, just that hex-based wargames do not sell well based on historical sales data. A computer wargame that sells 30,000 copies these days is considered a hit. (Even Age of Wonders did not sell well).

Having said that, maybe Shattered Union will remove the myth that hexes are the kiss of death that publishers seem to think they are. They imagine (maybe rightly) that the people who want hexes are the same guys who play ASL in their moms's basement (a small and shrinking demographic).

Wow, I realized that I didn't say anything about the technical merits. They are a total wash and neither is intrinsically better in a computer format. The reason hexes were used in paper wargames is two-fold (and neither apply to computers). Hexes can simulate movement ranges without additional computation (on a computer, the computer can do this easily so it no longer is a relevant argument). The second reason was that 40 years ago, when the first commercial paper wargames came out, most dice were six-sided and if you needed a random direction (for, say, a grenade bounce) you rolled a die and it mapped to a direction. The computer can hadle the die rolling.

A third reason (yes, I said two-fold), is historical, "It's always been done with hexes". Yup. It has and some of my friends and associates who are still making paper wargames, do use hexes (although the "new" (as in Mark Herman's "We The People" from 1994) point-to-point nodal systems are gaining much ground and I think that they will surpass hex-based paper wargames soon) and see no good reason to change. One could argue that for the Civ series "It's always been done with squares", and make an equally valid argument.

Incidentally, Sid did make hex-based games before Civ and chose to move away from them for Civ1.

 

[Akka]

One thing I will say on this, there hasn't been a hex-based game that broke the 100,000 sales mark in ages.

Well, there was still Fallout 2

 

[warpstorm]

True, I had forgotten about the RPG segment and was mainly focused on the TBS and wargame market. They sold ~300,000 copies.

But that was over 7 years ago, before the publisher consolidation.

 

[SonicX]

Warpstorm, Settlers I and II used hexagons ... (III and IV too, but those games weren't tilebased anymore)

 

[warpstorm]

...and Settlers came out when?

Settlers 1 was in '93 and Settlers 2 came out in '96, IIRC. This is even longer ago than Fallout 2. It was a DOS game. (Settlers 2 was the best of the series, IMHO). Even though it was a big seller in Germany (the series is about as big as Civ there for some reason), it sold next to no copies in the USA (which an American company like Firaxis has to worry about).

 

[SonicX]

It was also big here, but the lack of sales in the US and the success in Europe is probably due to nothing more than marketing and advertising strategies (and less competition in the native language)
But I just meant to say that Settlers II worked great with the hexagon tiling...

Now, almost no games at all use tiles so it's hard to compare. Settlers and other RTS games all became 3d instead and don't really have tiles anymore, only grids and that usually is only to simulate the American square way of city building (like the Command and Conquer series, Sim City, ...) and squares are easier on the eye to build these buildings.

In Civ, where everything involves radii, I would prefer hexagons, but then again, it would be a loss because you can't simply move west or north anymore.

 

[warpstorm]

It was also big here, but the lack of sales in the US and the success in Europe is probably due to nothing more than marketing and advertising strategies (and less competition in the native language)
But I just meant to say that Settlers II worked great with the hexagon tiling...

Many years ago, Settlers 2 was my favorite game. It is a shame that they went away from what made the series unique and tried to be yet another RTS like all the rest of the non-descript RTSs.

Like I said, technically it's a wash. They both can be made to work fairly easily. It has to do with marketing.

The game marketers I know (most of whom would not deign to actually play a computer game, but that is a different gripe) when they think of hexes they think of the uber-geeky wargamer demographic (and certain people around the office who fall smack into the core of that demographic) and think to themselves "Is this where the money is?".

 

[Truronian]

Which is irrelevant. Piling more people onto a subjective opinion doesn't make it objective. You're asserting one thing is better than the other and expecting people to be convinced.

I'm not asserting that its better, I'm asserting its more popular, but then popularity is not something that game manufacturers are looking for, is it

I read it. Several of the ones above mine, in fact, and saw nothing any less arbitrary or baseless..

what a convincing rebutle...

 

[Atrebates]

I'm not sure of your meaning. If you think a spherical world wouldn't add anything, well, that's what you think. I like the idea myself, because it avoids the ugliness of map edges and it's more realistic, but I don't expect everyone to value those things the same way. However, if you think all tiling systems can be converted to a sphere, you are mistaken. Even if you use irregular shapes, any tiling of only quadrilateral shapes cannot be converted into a spheroid (assuming a cube is not considered a spheroid, of course). Forget about physical shapes and think of it simply in terms of graph theory. You cannot construct an analogous graph if all nodes are connected to exactly 4 other nodes.

No 2D net (ie. the current way the civ world is mapped) that is rectangular can become a sphere, put simply the top+bottom of the map must be narrower than the middle. The fact that the individual tiles are a certain shape makes no difference, hexagons form a sphere as well as squares.

What do you mean by spherical? A spherical gameworld like populous? or having the map represented as a sphere? After all there are no map edges in the game world, I'd be interested in seeing a spherical representation of the map but its not essential (IMO).

Whilst a sphere is a 3D object its surface (what we're interested in) is 2D and any point can be defined in relation to 2 points (distance+direction) and from 3 points only by distance. All maps of of our planet use grid squares (slight trapezoids obviously) but that is because lines of latitude converge on two points, squares can be formed into a sphere with only slight distortion (they would retain equal sides and angles).

I feels like you make a valid (geometric) point but I cannot see it, other mathmatic principles seem to go against it, can you provide the maths behind your point please.

 

[frekk]

One thing I will say on this, there hasn't been a hex-based game that broke the 100,000 sales mark in ages (the Civ franchise would end if it only sold 100,000 copies in an installation). Not to say that it can't happen, just that hex-based wargames do not sell well based on historical sales data. A computer wargame that sells 30,000 copies these days is considered a hit.

Sure, but, take what I was saying in context. As far as tactically realistic games go, the hex was generally preferred. Tactical wargames don't sell now, but that isn't because they use hexes.

Incidentally, Sid did make hex-based games before Civ and chose to move away from them for Civ1.

Sure, but he was also moving to a simpler, more general format altogether. (Also, civ is a strategic game, not tactical, so it's easier to be generalistic about things).

 

[warpstorm]

Why "simpler"? Hexes aren't really any harder to implement than squares (especially if you've got the source code for your last four hex-based games lying around).

To put it into context, why would a system that you claim is better for tactical representation (I disagree, but bear with me) matter for a game that isn't modeled at that level of abstraction?

 

[apatheist]

I'm not asserting that its better, I'm asserting its more popular, but then popularity is not something that game manufacturers are looking for, is it

What's ultimately popular is quality. If you choose to make an inferior game because a superficial element is more popular, you will likely suffer in the long run.

what a convincing rebutle...

Rebuttals only apply to actual arguments. All you have done is repeatedly state your opinion. You don't have a leg to stand on so there's nothing to rebut.

What do you mean by spherical? A spherical gameworld like populous? or having the map represented as a sphere? After all there are no map edges in the game world, I'd be interested in seeing a spherical representation of the map but its not essential (IMO).

I want the actual map to have spherical shape. They already have a spherical representation of a cylindrical world for Civ4, but that's just UI prettiness.

Whilst a sphere is a 3D object its surface (what we're interested in) is 2D and any point can be defined in relation to 2 points (distance+direction) and from 3 points only by distance. All maps of of our planet use grid squares (slight trapezoids obviously) but that is because lines of latitude converge on two points, squares can be formed into a sphere with only slight distortion (they would retain equal sides and angles).

The shape (trapezoid vs. square vs. whatever) doesn't matter if the number of sizes is always 4. Pretend that the Earth is divided into quadrilaterals one degree of longitude tall and one degree of latitude wide. At every latitude, then, there are 360 squares. The squares narrow in absolute terms as you approach the poles. However, when you get to 89 degrees North (or South), you end up with a conundrum. Every tile borders one tile each to the east, west, and south, but what is to the north? You could have the north be the directly opposite tile, but that's kind of ugly. After all, if you quantize your map such that every tile is the exact same size, that would be equivalent to the current cylindrical civ world, but with the ability to go from 90 degrees W to 90 degrees E by crossing the pole. Alternately, you could have each tile around the pole be accessible to/from every other one, but then that tile is no longer a quadrilateral because it borders 3 tiles to the south, east, and west, and 356 tiles to the north. Since we're assuming the game retains the "all tiles are of equal size (for game purposes)" concept, then to model a spherical world, as you move from the poles to the equators, you have to be able to drop a tile. That's why spheroids composed of hexagons have to have pentagons in them, as that allows them to reduce the number of tiles at latitudes further from the equator. I don't have the mathematical knowledge to express this in such terms, but hopefully the description above suffices.

 

[Milan's Warrior]

You cudn't use both hexs and pents on a map, it'd be so annoying to play on!

I think that actually it would be fun to have the occasional different tile

 

[frekk]

Why "simpler"? Hexes aren't really any harder to implement than squares (especially if you've got the source code for your last four hex-based games lying around).

Huh? I meant a simpler game from the user's perspective ... not a simpler game to create. Not that the choice between hexes and squares directly impacts on that necessarily, but the general idea of civ was something more publicly accessible and generalized, which squares probably suit better - at least, from a marketing perspective.

 

[Atrebates]

You could have the north be the directly opposite tile, but that's kind of ugly. After all, if you quantize your map such that every tile is the exact same size, that would be equivalent to the current cylindrical civ world, but with the ability to go from 90 degrees W to 90 degrees E by crossing the pole. Alternately, you could have each tile around the pole be accessible to/from every other one, but then that tile is no longer a quadrilateral because it borders 3 tiles to the south, east, and west, and 356 tiles to the north. Since we're assuming the game retains the "all tiles are of equal size (for game purposes)" concept, then to model a spherical world, as you move from the poles to the equators, you have to be able to drop a tile. That's why spheroids composed of hexagons have to have pentagons in them, as that allows them to reduce the number of tiles at latitudes further from the equator. I don't have the mathematical knowledge to express this in such terms, but hopefully the description above suffices.

Ah yes, of course (I presumed to drop squares as the map moved N/S without thinking how they fitted). I believe you would have a similar crunch with any shaped tiles unless a mix of shapes is used (?). The best solution is to designate 1 tile North, This means only 2 unusual shaped tiles are needed (and any tactical advantage of several *EDIT* linked tiles is offset by crappy polar terrain, obviously movement would need the mouse off of this 1 tile) but results in a spherical gameworld.

 

[apatheist]

I believe you would have a similar crunch with any shaped tiles unless a mix of shapes is used (?).

Not necessarily. You can do it with all pentagons. I suspect, though I cannot begin to prove, that you can't model a spheroid with more than m faces using only polygons with exactly n sides where m is greater than, say, 12, and n is an even number. There is a proof that there are only 5 Platonic Solids, but I don't require that the polygonal faces be regular (or even convex), only that they have a fixed number of edges (the spheroid must be convex).

The best solution is to designate 1 tile North, This means only 2 unusual shaped tiles are needed (and any tactical advantage of 360 linked tiles is offset by crappy polar terrain, obviously movement would need the mouse off of this 1 tile) but results in a spherical gameworld.

I wouldn't call that "best" at all. That would basically amount to the same world that exists now, except with magic transportation rules at the poles. Remember, all tiles are the same size. In the model you propose, you would have a world that is 25,000 miles in circumference at 0 degrees latitude and 89 degrees latitude both. It's not just a matter of solving the poles; to have a truly spherical world, you need to have the circumference of the world (in number of tiles) be progressively less at progressively higher latitudes.

There is no unambiguously best solution. Every one of them requires compromises. I have settled on the hexagons with impassable pentagons as my preferred option. Hexagons have been proven to be an effective and playable choice for board games, and making the pentagons impassable is to me a better option than having 12 accessible tiles with special rules.