Playing on hexes: Notes from Battle for Wesnoth

[Avs]

Play Heroes of Might and Magic 4 or 5. Both battle systems are based on hexes. The strategic map gameplay is turn based like Civilization, requires resources and gold per turn, and you can build buildings in each town to upgrade it.

Battle system is hexed base like mentioned. Melee units are in front of ranged units. Heroes have spells to augment the battle (like direct damage, stuns, buffs, debuffs, teleport units past obstacles, etc). The entire system is based on protecting strong ranged units with strong melee units at choke points on the hex field as I imagine it would be like for Civ 5. It's a great system and requires you to make decisions such as putting X many units in a 3 man formation infront of the ranged unit to protect it from top, middle, and bottom areas. Check it out if you havent.

 

[Ahriman]

Homm games are great, but don't do a very good job of representing the kind of hex war we're going to see, because the movement speeds of each unit are just too far, there's no terrain differentiation, and each tile can have a stack of unlimited size.
Plus all the various special abilities (flaming breath, etc)

 

[the343danny]

BTW on the AI aspect its unfair to judge hex based AI on a game thats made for free. It would be more convincing if the example was a hex based game made by professional developers paid for their work with a lot more budget than this game, and also had previous non hex games that had AI better in those games.

 

[Ahriman]

BTW on the AI aspect its unfair to judge hex based AI on a game thats made for free. It would be more convincing if the example was a hex based game made by professional developers paid for their work with a lot more budget than this game, and also had previous non hex games that had AI better in those games.

We're not "judging" BfW, saying "oooh, bad AI, bad game". We're saying that the AI there didn't have a huge amount of effort put into it, and the outcome of that highlights the need of putting a lot of effort into AI for a 1upt system for a Civ game.

Its no criticism of BfW.

 

[GreatProfit]

I probably phrased that wrong. What Civ IV doesn't do (anymore) is have, say, a tank get killed by an archer. In BfW, I'm seeing the equivalent -- seeing it a lot. For whatever reason, it seems a lot more random, and frustrating, in BfW.

This is exactly why I stopped playing Wesnoth. There's nothing more frustrating for your mage to have 70% CtH (as always) but missing all three of his attacks. Or your horseman to have 80% CtH and missing both attacks. The Wesnoth combat system is very hit or miss, with sometimes huge consequences for bad luck. The Civ IV combat system feels a lot more deterministic and granular; and IMHO that's the way it should be.

Then there's the ******** trait system (strong mages, ugh) but that's another issue.

 

[Ahriman]

There's nothing more frustrating for your mage to have 70% CtH (as always) but missing all three of his attacks.

You should expect this to happen sometimes.... there is a 2.7% chance that your mage will miss all 3 times in a row. 1 in 37 times this will happen.

Or your horseman to have 80% CtH and missing both attacks

4% of the time, or 1 in 25.

These events are uncommon, but hardly rare.

The BfW probability generator isn't messed up, its just that people are weird about perceiving probability, particularly low probability events - and you only notice the times when something uncommon happens.

 

[Truronian]

Romance of the Three Kingdoms 11 is another hex based warfare game. The movement is 'faster' than it will be in Civ 5, but the game incorporates ZoC so I imagine the defensive lines that build up in RotTK11 will be similar to those in CiV.

 

[EmpireOfCats]

The BfW probability generator isn't messed up, its just that people are weird about perceiving probability, particularly low probability events

That's as maybe -- in the end it isn't fun, and that is what counts. Never had this form of frustration with Civ (except in the bad old days of archer beats tank). This is simply wrong.

 

[civ_king]

hehehe, I had a scout beat a Modern Armor once!

 

[Auncien]

As to the AI in Civ V, if it works they way they say it will, it's going to mop the floor with us for a while until we figure things out. Multi-tier AIs working together toward a single goal? That's some SkyNet stuff right there... If anything the combat system changes work in the AIs favor because it can be made really, really good at supply lines and unit positioning. There aren't as many random factors as in Civ IV. I predict much hurt the first few days of play.

 

[Gelvan]

how would a higher difficulty level make a harder AI, if not by giving it bonus/handicap the human?

there are two ways to get different levels: as in Galciv where the AI get's more subroutines on higher levels - or you program the best AI ever, put it on "noble" and adjust the pace it can produce and research.

do you really believe there will be an AI with more "intelligence" on immortal? I don't think so. this would take too much time to program 6 "different" AI instead of one "perfect" that get's a handicap or bonus.

 

[Ahriman]

A If anything the combat system changes work in the AIs favor because it can be made really, really good at supply lines and unit positioning. There aren't as many random factors as in Civ IV.

How do you figure?
a) There is no evidence of any supply line mechanic
b) Why would you think the AI would be better at positioning? This is the kind of decision that AIs in general are bad at, because it requires looking forward to see what your units are being exposed to on the enemy turn.
Positioning in Civ4 is easy because there is basically no positioning, just big stacks.
c) How can you possibly conclude that there aren't as many random factors as in Civ4? What random factors are there in Civ4 that we know won't be present in 5?

 

[the343danny]

We're not "judging" BfW, saying "oooh, bad AI, bad game". We're saying that the AI there didn't have a huge amount of effort put into it, and the outcome of that highlights the need of putting a lot of effort into AI for a 1upt system for a Civ game.

Its no criticism of BfW.

I know you arent insulting the game. I am saying that its hard to judge an AIs capabilities in a 1upt system based on a free game without professional developers working for money

 

[Ahriman]

I am saying that its hard to judge an AIs capabilities in a 1upt system based on a free game without professional developers working for money

This makes no sense to me.

You can judge a specific game's AI by playing the game. It doesn't matter who developed it. Yes, a free game without professional developers working for money will have weaker AI... but that doesn't mean you can't judge it as it is.

The salary of the coders is completely unrelated to the ease with which you can see how effective that particular AI is or not.

 

[PieceOfMind]

???
Both Civ and BfW have probability-driven battles. And the probabilities are displayed in-game (though BfW displays the distribution of outcomes, which you only get in Civ through BUG).

Just a correction. You're talking about my baby - ACO - from what I can tell. This is not actually included in BUG, but rather in BULL and therefore any other mod that uses BULL, plus some more that don't use BULL. It's also available as a standalone mod to BtS 3.19. Link to ACO is in my signature.

 

[Ahriman]

Sorry, apologies for misrepresentation. I play using so many bundled modcomps that I forget which is which.

I'm a big fan of the various modcomps that improve UI transparency, such as showing the player what the distribution of combat outcomes is. These kind of things should have been in core IMO.

So thanks, great job.

 

[Sirp]

I'm the founder and lead developer of Battle for Wesnoth. Here are my thoughts on this thread:

- I'm very glad to see the Civilization franchise go in the directions of hexes with 1upt. For unit movement it offers a nice feel, and for city radiuses it absolutely makes sense.

- The 1upt will make wars much more strategic. In Civ3 the dominant strategy was to make a 'stack of doom' of (swordsmen|knights|cavalry|tanks|modern armor) and charge it into enemy territory taking city after city. In Civ4 things became more interesting since your stack had to be mixed. In Civ5 you will actually have to plan out the formation of your armies.

- I think that a limited amount of comparison can be made between the Civ5 system and Wesnoth. The big difference in Wesnoth is that units can move many more tiles in a turn than is typical in Civilization.

- Overall I think the AI in both Wesnoth and Civ4 is one which on equal terms can challenge a casual player, but which will be demolished by an expert player. It is true that a 1upt/hex system will be more of a challenge to write a good AI for.

- Wesnoth is an Open Source game, which means we have very different production values from a typical commercial project. However, this doesn't mean that we don't have very capable developers. I started Wesnoth because I wanted to write a game, but didn't want to work for a game company due to poor wages and conditions that most video game companies offer. I wrote most of the AI code for Wesnoth, and it was one of my prime focuses for quite some time. I have a successful career in the software industry, including time at Google as a senior software engineer, so even if I didn't get paid for Wesnoth, I'd really like to think I know what I'm doing.

- I am hoping that the Civ5 developers really do concentrate on the AI heavily, but I am doubtful that more than fairly incremental improvements over the Civ4 or Wesnoth AI's will be achieved.

-Sirp.

 

[Ahriman]

and it was one of my prime focuses for quite some time. I have a successful career in the software industry, including time at Google as a senior software engineer, so even if I didn't get paid for Wesnoth, I'd really like to think I know what I'm doing.

Then apologies for claiming that you didn't put much effort into it.

Thanks for BfW; a truly incredible game. I think the decision to have the probability of being hit be a simple percentage based solely on the terrain there on is one of the most elegant design decisions I've seen.

I agree that the differences in movement rate will make an absolutely massive difference. Hopefully with smaller movement, the differences between human player and AI will be smaller than in BfW.

But the human/AI differences in BfW are very large; if those kinds of disparities occurred in Civ5, I think we would have a problem because of the sheer number of wars that happen in a civ5 game; much more than a single BfW scenario.

 

[Frasco]

I'm the founder and lead developer of Battle for Wesnoth.

 

[Sirp]

But the human/AI differences in BfW are very large; if those kinds of disparities occurred in Civ5, I think we would have a problem because of the sheer number of wars that happen in a civ5 game; much more than a single BfW scenario.

I don't think that human/AI differences in past Civ games are much less than in Wesnoth.

To begin with, I don't think the advantages the AI gets in Wesnoth are as large as some people think. On most campaigns, the AI gets quite a lot of gold in many scenarios, but the player gets to recall higher level units from past scenarios which are typically more powerful than any of the units available to the AI. I used to play Wesnoth matches involving myself and a couple of other developers against the AI. We found any game where the AI had more than a 2:1 advantage quite difficult to win.

In Civ3 I could regularly have defensive wars with other civilizations that had an immense military advantage and defeat them. For instance, I played Always War games on Monarch difficulty against six other civilizations and won. In Civ4 it's harder, but still achievable.

What one perhaps can't do quite as well in Civ as in Wesnoth is prosecute an offensive war. In Civ you largely have to have something approaching parity to actually capture enemy cities, because cities are always strongholds. In Wesnoth you can sometimes achieve a "decapitation attack" of the enemy leader.

But this is easily solvable in Civ5 -- just make it so cities are sufficiently strong that regardless of other factors you will have to have decent strength to overcome them. I can imagine a large number of improvements -- stockades, city walls, castles, etc, that one can build to make a city more difficult and costly to attack.

-Sirp.