Hexes vs. Squares

[Galdred]

The Codex already had this debate. The arguments in favor of hexes were found to be wanting.

http://www.rpgcodex.net/forums/index.php?threads/turn-based-square-grid-vs-hexagonal-grid.34078/

It's funny because I had the same discussion concerning my current gaming prototype. The main reason why I prefer hex is not only because of move cost (after all, 1.41 is not that far away from 3/2), but because diagonals movement poses a lot more problems :

AX
XB
X : monsters
What is supposed to happen when a character goes from A to B?
Either he can, and blocking anyone becomes infuriatingly hard, because you'd need a third guy to block the move, either he cannot, and it allows to interdict movement way too easily.
Attacking and LoS blocking poses the same problem. You end up with lots of non intuitive rules about what you are able to see and attack, and what blocks movement. It has a very big effect on gameplay in games about melee, where you are supposed to shield your frail casters or archers from harm (or get to the opposing casters), without using MMORPG boring skills like taunt and whatever : it makes you need to use 50% more troops to block a direction compared to another, which is not a trivial change at all .

The only problem with hexagons (and it is a big one indeed), is that our architecture does not map well into hex. hex tiles definitely look horrible most of the time (although they are beautiful in Age of Wonder).

I'm glad Chaos Chronicle is going hex based.

I'm curious : How did you work Line of Sights? are they vertex based(only some parts of the hexagon block LoS, which makes more sense when the hexagon represents a wall), or hexagon based (all the parts of the hexagon block sight, it makes more sense, but makes it harder for the player to know beforehand where he will have LoS to a given point)?

 

[Davaris]

HobGoblin42
Hexagons are beautiful, end of story. If they talk you out of using them, then please give me that code. I'll find a good home for it.

 

[J1M]

Well @Galred, I think it is a little silly to try and create a physical wall with a party of 6. You really only have three pieces to build the wall with usually. Rules for opportunity attacks and engagements make more sense. There's not a lot an evenly matched opponent can do to stop you from going through them if you are willing to take a hit that doesn't end up in a pileup.

Also, hexes don't solve any of the line of sight issues you have raised.

 

[tuluse]

Rectangular room mapping right into hexes (green Xs are hexes that can be occupied). If you have such severe OCD that you can't handle the character pieces being shifted a little bit when next to each other, you should be asking for no grid and just using circles to represent characters, and pixels to represent size of areas. Essentially what ToEE did.

 

[Jim the Dinosaur]

Rectangular room mapping right into hexes (green Xs are hexes that can be occupied). If you have such severe OCD that you can't handle the character pieces being shifted a little bit when next to each other, you should be asking for no grid and just using circles to represent characters, and pixels to represent size of areas. Essentially what ToEE did.

That looks insane; like someone else said, Fallout handled rectangular rooms with hexes just fine:

 

[HobGoblin42]

From one of our blog entries:

Since modern levels are not constructed with small tiles but with free 3d objects, the difference between hexagons and squares regarding the topology isn't that big.

 

[villain of the story]

Will elevation and angle of slope matter in CC? Eg. attacking from a higher vantage point or from an unstable position up or down a slope?

 

[HobGoblin42]

Will elevation and angle of slope matter in CC? Eg. attacking from a higher vantage point or from an unstable position up or down a slope?

yes, because it influences the line of sight, especially for ranged attacks. An archer on higher position has better chances to hit enemies. For melee attacks, we haven't included an influence of height differences yet since both, attacker and defender, use to share the same height level in almost any case.

Fighting on tables or chairs could be a neat element though.

 

[AbounI]

Can't you apply the same kind of rule for melee attacks as the sizes of creatures which can grant bonus/malus depending of this ruleset?Is it too late & hard (specially if it need balance tests) now to implement this?

 

[Davaris]

Another reason hexagons are nicer, is because no matter which adjacent hex you move to, the distance is the same. With tiles the distance changes if you move on the diagonal. Yuck.

http://www-cs-students.stanford.edu/~amitp/gameprog.html#hex

 

[HobGoblin42]

Can't you apply the same kind of rule for melee attacks as the sizes of creatures which can grant bonus/malus depending of this ruleset?Is it too late & hard (specially if it need balance tests) now to implement this?

In theory yes, that could work. The size modifier applies in D&D to both: for the attacker as attack bonus and for the defender as AC bonus: http://www.dandwiki.com/wiki/SRD:Attack_Roll

A height step (e.g. of 2 feet each) above the enemy could add an additional attack bonus of 1, each height step below -1. But as I already said, usually the involved characters are standing on the same height when in melee.

 

[villain of the story]

Indeed. Would it be too much trouble? Extra animations to jump up and down furniture which would otherwise be blocked hexes and maybe extra animations for and against targets at higher and lower points too.

 

[Galdred]

yes, because it influences the line of sight, especially for ranged attacks. An archer on higher position has better chances to hit enemies. For melee attacks, we haven't included an influence of height differences yet since both, attacker and defender, use to share the same height level in almost any case.

Fighting on tables or chairs could be a neat element though.

Does it mean that LoS is not only tile based? Is it possible that some tiles only partially obstruct los?

Well @Galred, I think it is a little silly to try and create a physical wall with a party of 6. You really only have three pieces to build the wall with usually. Rules for opportunity attacks and engagements make more sense. There's not a lot an evenly matched opponent can do to stop you from going through them if you are willing to take a hit that doesn't end up in a pileup.

Also, hexes don't solve any of the line of sight issues you have raised.

Even with Attacks of opportunity, you don't really solve the problem, because with square tiles, your "area of control" radius will be either a square or a cross, which make defending much easier if the ennemy is lining up diagonally, for square shaped ZOC or horizontally, for cross shaped ZOC, while the area of control of an hexagon is hexagonal shaped too.

Concerning the LoS problem, it is true that hexagons are faced with the same partial obstruction problem, but square poses another problem :
You can have 2 tiles that only have 1 point (or zero depending on rounding) on the ray, and thus miss them easily, although they completely obstruct the ray :

AX
XB
The two X obstacles both have only a single point of contact with the Los, it is still possible to see A from B if you miss both "points of contact" (which is easy to do if using something like the Bresenham algorithm) even though it is physically impossible (and many games allow to shoot from A to B or to melee Attack). In this case, it is either too easy (if the two obstacles block LoS) or too hard (if they don't) to protect someone from ranged attack diagonally.

With hexagons, you cannot run into a situation where a line passes right between 2 hexagons with only 1 point of contact : you could obviously draw a line joining 2 "hexagonal tips", one obscuring LoS on the right, and the other on the left, and end up with a similar problem indeed, but it is not a frequent as going for a straight diagonal.

Note that none of these problems matter much in a game which focus on emptying your gun from long range with a spread out squad, as the range calculations obviously count diagonals as 1.5 cost of the "standard" cardinal directions, and you don't rely on your guys to shield each other from harm, but for tactical dungeon crawlers (like Descent or Warhammer Quest), it makes a huge difference.

 

[HobGoblin42]

Does it mean that LoS is not only tile based? Is it possible that some tiles only partially obstruct los?

A pure tile-based LoS system doesn't work for us since we have various kinds of cover like fences, trees, tables, etc.
That's why we need to include the actual physical topology of the environment as well.
We don't want to surprise the player with total different LoS results than the combat battlefield obviously looks like, that's why our LoS computation is much more complex than the tile-base tabletop D&D ruleset.

 

[tuluse]

That looks insane; like someone else said, Fallout handled rectangular rooms with hexes just fine:

There's no real difference here other scale and the fact that I drew mine in roughly 15 seconds and didn't bother giving the walls width.

 

[Jim the Dinosaur]

That looks insane; like someone else said, Fallout handled rectangular rooms with hexes just fine:

There's no real difference here other scale and the fact that I drew mine in roughly 15 seconds and didn't bother giving the walls width.

Yours looked like it had "split" hexes (more than half inside room = occupyable, otherwise not), that has nothing to do with wall width and is pretty insane.

 

[tuluse]

That looks insane; like someone else said, Fallout handled rectangular rooms with hexes just fine:

There's no real difference here other scale and the fact that I drew mine in roughly 15 seconds and didn't bother giving the walls width.

What? Yours had "split" hexes (more than half inside room = occupyable, otherwise not), that has nothing to do with wall width and is, as I said, insane.

Fallout had straight walls. It had split hexes too, it just hid them well.

 

[Jim the Dinosaur]

That looks insane; like someone else said, Fallout handled rectangular rooms with hexes just fine:

There's no real difference here other scale and the fact that I drew mine in roughly 15 seconds and didn't bother giving the walls width.

What? Yours had "split" hexes (more than half inside room = occupyable, otherwise not), that has nothing to do with wall width and is, as I said, insane.

Fallout had straight walls. It had split hexes too, it just hid them well.

No it didn't, look at the screenshot, a hex is either blocking or it isn't, which occasionally (south and north walls in this case) creates thick walls of blocking hexes. That's all I meant, a rectangular rooms can be done easily with hexes, but only at the cost of a greater degree of blocking than with squares.

 

[tuluse]

Oh yeah I you're right. Fair enough.

 

[J1M]

Rectangular room mapping right into hexes (green Xs are hexes that can be occupied). If you have such severe OCD that you can't handle the character pieces being shifted a little bit when next to each other, you should be asking for no grid and just using circles to represent characters, and pixels to represent size of areas. Essentially what ToEE did.

A disingenuous solution at best. The room you have presented is not rounded to the width and height of a hexagon, and thus will not tile at all with an adjacent room or a hallway. One might say you have not drawn a room at all, but a game world with only 12 grid locations. o_O

 

[J1M]

Another reason hexagons are nicer, is because no matter which adjacent hex you move to, the distance is the same. With tiles the distance changes if you move on the diagonal. Yuck.

http://www-cs-students.stanford.edu/~amitp/gameprog.html#hex

Sure, in theory.

Theory ends when you try to move north and then west, or forward and then sideways.

Instead of a diagonal distance calculation, your character ends up walking as if drunk... but only in certain arbitrary directions. With squares, movement is consistent in all directions.

 

[J1M]

Well @Galred, I think it is a little silly to try and create a physical wall with a party of 6. You really only have three pieces to build the wall with usually. Rules for opportunity attacks and engagements make more sense. There's not a lot an evenly matched opponent can do to stop you from going through them if you are willing to take a hit that doesn't end up in a pileup.

Also, hexes don't solve any of the line of sight issues you have raised.

Even with Attacks of opportunity, you don't really solve the problem, because with square tiles, your "area of control" radius will be either a square or a cross, which make defending much easier if the ennemy is lining up diagonally, for square shaped ZOC or horizontally, for cross shaped ZOC, while the area of control of an hexagon is hexagonal shaped too.

Concerning the LoS problem, it is true that hexagons are faced with the same partial obstruction problem, but square poses another problem :
You can have 2 tiles that only have 1 point (or zero depending on rounding) on the ray, and thus miss them easily, although they completely obstruct the ray :

AX
XB
The two X obstacles both have only a single point of contact with the Los, it is still possible to see A from B if you miss both "points of contact" (which is easy to do if using something like the Bresenham algorithm) even though it is physically impossible (and many games allow to shoot from A to B or to melee Attack). In this case, it is either too easy (if the two obstacles block LoS) or too hard (if they don't) to protect someone from ranged attack diagonally.

With hexagons, you cannot run into a situation where a line passes right between 2 hexagons with only 1 point of contact : you could obviously draw a line joining 2 "hexagonal tips", one obscuring LoS on the right, and the other on the left, and end up with a similar problem indeed, but it is not a frequent as going for a straight diagonal.

Note that none of these problems matter much in a game which focus on emptying your gun from long range with a spread out squad, as the range calculations obviously count diagonals as 1.5 cost of the "standard" cardinal directions, and you don't rely on your guys to shield each other from harm, but for tactical dungeon crawlers (like Descent or Warhammer Quest), it makes a huge difference.

RE: Attacks of opportunity + squares = square or cross shaped.
I don't see this as a problem, it is easy to read where enemies will threaten. Additionally, it opens up opportunities for different weapon types to threaten in different ways. Finally, I will remind you that the issue you describe is only really a concern if a character takes up exactly one square. Place that same character so that it takes up a 2x2 or 3x3 square grid location and suddenly the threaten area is essentially round.

RE: Partial obstruction
I am glad you can admit that squares and hexes share the same issues of partial obstruction. I find it unfortunate that you do not see the same applies to your edge case of a line of sight passing through the corner of a hex. Deciding the point of a square is special reveals a bias. The very same issues apply to a line of sight passing along the edge between two hexes. Both situations form tangents with their respective grid shapes.

If a binary "this blocks" or "this does not block" line of sight rule is deemed too punishing in either direction, the obvious solution is some sort of "partial cover" rule, which most systems have in place anyway.

I don't want to just rehash what is in that excellent thread I linked to. Preferring hexes is fine, but most of the advantages people list about them are merely informal fallacies.

 

[baturinsky]

That looks insane; like someone else said, Fallout handled rectangular rooms with hexes just fine:

Except that infamous zig-zag verical movement...

 

[Davaris]

That looks insane; like someone else said, Fallout handled rectangular rooms with hexes just fine:

Except that infamous zig-zag verical movement...

As opposed to jaggy based movement with tiles on every diagonal? The only way you can fix that kind of movement in either system, is if you abandon the idea of tiles/hexes and average to smooth the path, which is kind of hacky in a TB game IMO.

 

[Jim the Dinosaur]

Sure, in theory.

Theory ends when you try to move north and then west, or forward and then sideways.

Instead of a diagonal distance calculation, your character ends up walking as if drunk... but only in certain arbitrary directions. With squares, movement is consistent in all directions.

But what would you want diagonal movements to cost then? You can't have it the same as non-diagonal, so I assume, like Monblut in the thread you posted, that you'd want 1.5 the cost for diagonal movement; that means hex movement is "consistent" in 6 directions, while square movement is consistent in 4 directions at a time. That just makes hexes nicer in my opinion.