hex versus squares for war games

Last non-computerised turn-based war game I played was a Warhammer varient, which didn't use hexes or squares. You can just move units their max move distance in any direction.
Last computer war game I played was Napolean Total War, which is simmilar - in the turn-based section it's pretty much like moving circular tokens across your dining table ;) When resolving a battle, any army tokens within a certain distance from the attacked-unit can count as 'flanking' units (i.e. you draw a circle around the attacked-unit, any armies within the circle participate in the battle).

Do you need to discretise the map into cells?

I feel like the appeal of turn based play is you have certainty about your moves. You know exactly how your units are able to maneuver on the terrain. Unless there's a great game play benefit to doing away with tiles, I don't see why I should?

Both the games I mentioned above are turn-based and do have certainty about how far your units can move without a descritised map. In Napoleon it's visualised the same as in many tile-based games - when you select a unit, the region that it can move to in a single turn is highlighted, it's just that this region is a smoother shape. You can also visualise the "attack region", which is usually a smooth circular shape instead of a group of 9 squares or 7 hexes.
Is there a great gameplay benefit to restricting movement to discrete tiles?

Wouldn't you be spending a lot of time trying to position your pieces, so they're just in-range or just out-of-range? Anyway I'd be interested to check out Napoleon Total War.

As a gameplay decision I would say that dividing the area into cells makes it a lot easier to reason about distances, which is useful in turn-based strategy. It also lets you be clear as to whether certain areas are occupied or not, and makes explicit the case when multiple units are grouped together.

It also makes it much easier to write planning algorithms for if you're developing an AI player.

Of the two, I would prefer hexes for wargames, although they come with their own issues (eg. having to go North-East then South-East to move 2 steps East).

I'm inclined to agree with Kylotan that a free-ranging map is not free from problems. Having a strongly quantized map makes it much easier to tell which units can reach you, which units can't reach you but will be close enough to bombard you, which units you'll be able to reach next turn, etc. I'm not saying that free-ranging maps can't work (they obviously do for the appreciable number of people who play tabletop wargames), but they aren't an obvious choice either.

I tend to prefer hex as well. Squares feel toyish for some reason.

In terms of time and distance scales, I'm thinking a movement speed of 3-5 squares per turn, 4 turns every 24 hours awarded at a randomized time for each player (but the same time each day). In a fighting advance, if the player advances 10 squares per day on average, he will cross half of the board (300 squares) in 30 days of online play. My board will measure 1024x1024. That should be able to comfortably handle half a dozen to a dozen players on each side of the front.

Just my opinion, but hexes feel much more natural than squares in terms of how distances work, what it takes to surround something, what it takes to go around something, etc. I haven't played much of this genre but I was definitely bothered by Civilization III operating on square grid when I tried it.

I agree hexes are better for tactical game play. On a square map, either you can attack diagonally, or you cannot. And they both feel unnatural.

Squares seem to make for nicer looking maps. A rectangular feature just "looks right".

Personally I'm a fan of dumping the hard cells of "One unit fits in one full cell" idea. To me it feels weird, and limits movement.

If you use small enough squares then you can calculate out the distance of a slope and keep your movements even between all directions, and allow movement in any direction.

If you want it to be 'easy to see how far something can move', then high light its range from its current position. Problem solved. (from the game play view.) You can actually use a lot of different highlight modes to show a lot of information.

Quote:

Original post by Talroth Personally I'm a fan of dumping the hard cells of "One unit fits in one full cell" idea. To me it feels weird, and limits movement. If you use small enough squares then you can calculate out the distance of a slope and keep your movements even between all directions, and allow movement in any direction.

Why would you still use squares and not move to a continuous space if you are going to make them so small as to be insignificant?

Quote:

Original post by Stroppy Katamari Quote: Original post by Talroth Personally I'm a fan of dumping the hard cells of "One unit fits in one full cell" idea. To me it feels weird, and limits movement. If you use small enough squares then you can calculate out the distance of a slope and keep your movements even between all directions, and allow movement in any direction. Why would you still use squares and not move to a continuous space if you are going to make them so small as to be insignificant?

Because while a Pixel is not a little square, they're usually a squarish form sooner or later. (Reference to a paper everyone should read, Google "A pixel is not a little square" and you should get it, Alvy Ray Smith.) The "small square" is one of those things that can be there, but not there at the same time, depending on how you choose to handle your data and such.

There are pros and cons to both methods. Locking things to a 'substrate' can allow for different ways of handling your data and collision detection than having it "centered on point Float X,Y". If units "Snap to Grid", but are allowed free movement at any angle within their range, then you can possibly get more even and structured looking placements. If a small infantry unit exists as a "16*16" circle, then it is easier to line them up perfectly in neat columns when moving one by one. If they're in a totally free space, then a line that runs along the X or Y axis is likely going to look wobbly if each was placed by hand. This minor detail can subconsciously annoy a surprising number of people.


Basically my point was to eliminate the movement restriction of "You can only move these 4, 6, 8, or 12 directions!", while preserving the fairly structured "Board game" feel that can be comforting to some people.

Of course you can always ignore that and go for completely free-floating maps where nothing aligns to a grid, which is good as well.

Wargames aren't about hand-placing infantry fighters in a neat line. That's for hands-on abstract RTS games like Starcraft. Even a Napoleonic war game, which involves uniform formations, should do that stuff automatically within the unit.

Quote:

Original post by Stroppy Katamari I haven't played much of this genre but I was definitely bothered by Civilization III operating on square grid when I tried it.

I believe they are switching to a hex based system in Civ V.

Quote:

Original post by Rattrap Quote: Original post by Stroppy Katamari I haven't played much of this genre but I was definitely bothered by Civilization III operating on square grid when I tried it. I believe they are switching to a hex based system in Civ V.

omg...

Honestly, I've never understood why nobody (I have seen) has done an octagonal system. It's got all the four cardinal directions, straight lines, diagonals, discrete steps, etc. You could go all the way to sixteengons and beyond, but I don't know if that would make the game better or just more complicated. Remember that the number of sides has an effect on how LOS and movement is calculated, but also on how players interpret the game environment.

Octagons don't tile... The. They don't really do anything that squares-with-diagonals do; they still suffer from the "one set of directions is faster than the others" problem of diagonal vs orthagonal movement.

Hexagons tile the plane completely, and movement in six directions is uniformly scaled.

Quote:

Original post by Talroth Quote: Original post by Stroppy Katamari Quote: Original post by Talroth Personally I'm a fan of dumping the hard cells of "One unit fits in one full cell" idea. To me it feels weird, and limits movement. If you use small enough squares then you can calculate out the distance of a slope and keep your movements even between all directions, and allow movement in any direction. Why would you still use squares and not move to a continuous space if you are going to make them so small as to be insignificant? Because while a Pixel is not a little square, they're usually a squarish form sooner or later. (Reference to a paper everyone should read, Google "A pixel is not a little square" and you should get it, Alvy Ray Smith.) The "small square" is one of those things that can be there, but not there at the same time, depending on how you choose to handle your data and such. There are pros and cons to both methods. Locking things to a 'substrate' can allow for different ways of handling your data and collision detection than having it "centered on point Float X,Y". If units "Snap to Grid", but are allowed free movement at any angle within their range, then you can possibly get more even and structured looking placements. If a small infantry unit exists as a "16*16" circle, then it is easier to line them up perfectly in neat columns when moving one by one. If they're in a totally free space, then a line that runs along the X or Y axis is likely going to look wobbly if each was placed by hand. This minor detail can subconsciously annoy a surprising number of people. Basically my point was to eliminate the movement restriction of "You can only move these 4, 6, 8, or 12 directions!", while preserving the fairly structured "Board game" feel that can be comforting to some people. Of course you can always ignore that and go for completely free-floating maps where nothing aligns to a grid, which is good as well.

If you have played Command HQ (remember Microprose?), this is essentially the system they used. I'm surprised that system hasn't shown up in more games. It worked well as a board game/pre-RTS hybrid.

Quote:

Original post by Katie Octagons don't tile... The. They don't really do anything that squares-with-diagonals do; they still suffer from the "one set of directions is faster than the others" problem of diagonal vs orthagonal movement. Hexagons tile the plane completely, and movement in six directions is uniformly scaled.

Thanks for this succinct explanation.