Something that's also worth thinking about when talking about defensive advantages is that the size of a city can limit how large a force is brought to bear against it. Even if you fully surround your target city, only so many troops can be up against (or on or whatever) the walls at once - perhaps add an additional layer around that if you have sufficient ranged (archer/mage) troops. This might be a good way to help out young cities, as a large army might be able to provide reinforcements for fallen troops for a prolonged period of time, but they wouldn't be able to simply sweep over the city as they can't bring their full forces to bear.

Can I also just throw out there that I love Sins of a Solar Empire to death? Great game.

I offer a hypothetical solution to both the space problem and the difficulty for new players joining late:

First, a few key assumptions:
1) You've made it difficult for players to attack other players who are far away on the map. This can be accomplished by a combination of the time commitment to march slow army units a great distance over the map and the supply line concept (which would be costly, and susceptible to being cut off leaving your army to die of starvation in enemy territory).
2) A player can not expand his controlled territory to non-adjacent regions of the map (ie: Player A can't march through Player B's territory to capture Player C's territory).
3) New players start with either a protection timer or extra resources that are scaled based on the size of their nearest neighbor's controlled territory/resources.

Now, for the rest:
Let's say you start the map as a 4x4 grid. When x% of the cells in the grid become controlled, the next new player to join causes the map to grow to 5x5 in size. The new player's starting territory would be spawned on this newly created periphery. This causes two things to happen. First, it creates a hierarchy, such that the players who have controlled territory the longest time (and presumably more territory) will be closer to the center of the map, and players who have controlled territory the shortest time (and presumably less territory) will be closer to the periphery of the map. Second, new players (thus with smaller, more vulnerable territory and fewer resources) will be farthest away from the strongest players (meaning the strongest players have to pay the highest price to attack the weakest players, per Assumptions 1-2). Anyway, the map would just keep scaling outward in size as new players joined. New players would tend to be placed next to other players who are only marginally stronger than they are, which is compensated for by the scaled starting bonus (Assumption 3). Finally, if players get wiped out completely, they can respawn with a new territory on the periphery, perhaps salvaging some of their resources (but at the expense of the starting bonus). This makes the game a constant struggle of the center trying to fight its way outward and everyone else trying to fight toward the center.

The only gaping hole I see in this is the possibility that the center could grow outward so much that it becomes too large to ever conquer. For this, you might want to factor in the resource decay/upkeep to put some kind of upper limit on the size that a territory controlled by a single player can grow to.

Anyway, I think this definitely solves the space issue with the game map - unless you were referring to computational resources for storing/representing it, but I highly doubt there'd be any issue there (nor do I really think that's what you were asking about).

Just to chime in with the Travian approach once more (because I feel it handles this part of the game pretty well): The world is a wraparound grid from -400/-400 to +400/+400, with each coordinate position holding either an oasis (which can be captured and controlled for a resource bonus) or the potential for a village (different spots have different resource distributions - eg some have more iron, others more clay, etc).

The first players appear near 0/0, but as the game progresses the newer villages spiral outwards from the centre. However you can settle/conquer new villages in any location - and indeed its quite common for players to spread out, although villages usually tend to grow together in clusters because it makes them easier to defend and support (sending resources to another village requires merchants, so it's slow work if the village is a long way away).

There are no explicit troops on the map - indeed, doing so would become utter chaos due to the large number of attacks that are constantly taking place. Instead, when you launch an attack the travel time is calculated based on the speed of the slowest unit and the distance between your village and the target village (completely ignoring the terrain type and who controls it). This is then (I presume) added to some sort of event list.

This is a very good point. One of the things you can observe in the Civ games is that several mechanics are simplified as time goes on, in favor of more interesting complexities. The notion of production and upkeep is relatively straightforward in Civ4, but had more aspects in Civ3 (and earlier). Furthermore, several small improvements were made to make micromanagement far less useful, so you could spend time on more interesting things like designing the empire, waging war, or whatever.

So yes, I'm a big fan of simplicity where possible, and complexity when interesting: in other words, the game aspects should be sufficiently complex to be interesting – and this could still be with quite simple mechanics –, but complexity should not be introduced as the equivalent of paper pushing.


That's an interesting idea. Indeed, having a whole bunch of people doesn't necessarily help you when piling up against a wall you can't knock down, except insofar as you can keep reinforcing your position.


I'm not sure of the motivations for this assumption. Why not allow non-contiguous territory in general? I understand that it can help for some things, but it seems somewhat artificial. That said, I think that you can sufficiently discourage non-contiguous empires that people would not want to march through B to capture C: for example, you can dramatically increase upkeep costs if you have to cross a road that's not in your control.

This would let you have little islands of control, but each island would have to be basically self-sufficient. This could be interesting if somebody cuts off your territory, for example, or if you want to capture a strategic cell to cut off somebody's supply chain. You would capture the territory, but not really be able to do anything with it unless you could build it up self-sufficiently.


I like this in principle, but it means that you need to randomly generate maps, which is something I was hoping to avoid. Still, it might not be so bad: the Civ games generate relatively interesting terrains, most of the time at least. (I was hoping to use terrain as a tactical factor, for instance giving importance to location near rivers, mountain ranges, etc.)

I do like the idea though of having a "core world" versus an "outer rim" of sorts, with established empires being close to the center of the world and upstarts being further away. You could maybe hammer this in by having a dynamic center, such that when trying to place new players, you choose the cell furthest away from the strength-weighted average of cells.


I'm not really sure this is a problem, though. If it were all one player, it could be a problem. But then, as you say, you can use other mechanics to make it too difficult to defend this whole territory. Besides, if you're in the middle of a circle of foes, you would essentially be waging a war on all possible fronts if people decided they wanted to take you out.


I'll have to think about this one. I kind of like representing armies as they go places, not because it's pretty but because it lets you intercept them in places more or less advantageous. But if enough armies are moving about, it could get messy pretty quickly, yes.

Could incorporate the ability to spot troops multiple ways. For example, by examining a particular location to see what troops are currently in/near it and the direction in which they are headed, or a more tactical maneuver of sending a spy to a player's capital to see where (all?) that player's troops are currently heading and in what numbers.

This can definitely still work if you treat the non-contiguous territories as being separate. As in, there would be no pooling of resources or armies, and each would be responsible for its own production and defense. This keeps the basic premise working, while allowing more skilled players to multi-task more, keeping watch on even more borders.


Well, you could either go with random generation, procedural generation or a static option. The static option would be just to sculpt a REALLY big map the way you want it, then only display the scope of it that is needed per the current number of players/controlled territory.


I meant with it just being a single player. It would work fine under your interpretation.

When someone launches an attack against you in Travian, you see a little red crossed-swords icon with the number of incoming attacks beside it, and a countdown timer until the first one hits. You can click the icon to see a breakdown of each attack, which lists the player and village, the time until the attack hits and the exact arrival time (but not what troops they've sent) as well as whether it's a full attack or just a raid.

Thus if you wanted some sort of intercept feature, you could allow players to click on one of the incoming attacks and choose an "ambush" option. However this wouldn't work very well in a Travian-style game, for two reasons.

The first reason is because troops have both an attack and defence rating, and the former is only used when they attack while the latter is only used when they defend - thus intercepting an offensive army would be brutally effective, so much so that few people would bother with a traditional defence.

The second reason is to do with the way most attacks consist of multiple waves. To inflict structural damage requires catapults, but because each group of catapults can only hit one or two targets that means you need to split them into multiple waves for maximum effect. This results in people trying to dodge the first wave and kill the catapults, and has led to various counter-tactics such as leading waves of fakes (so the victim doesn't know where to insert their defences), multiple attacks landing on the same second (which is impossible to insert defences into), etc. Combat resolution is also instantaneous - as soon as the troops have hit, they start marching back.

Now personally I'm not too fond of either of the above two features, but changing them would obviously have a big impact on the rest of the game. For example, I'd like to see troops spend some time fighting when they reach their destination - but then you need to cater for other armies hitting while a battle is already in progress, etc. Obviously this can be done, and obviously Travian's approach is far from the only workable solution, but such changes raise a lot of "what if"s that would require some careful consideration.

Another way of handling interceptions would be to require armies to set up camp after a game-day of travel, with the camp functioning much like a village (perhaps even with limited building options, allowing the army to dig defensive ditches). The enemy could then "intercept" the army by launching an attack against the camp. From an interface perspective it would simply be impossible to attack a target more than one game-day away - instead, you'd have to click a 'create camp' option on an empty square within range. However this could become a problem if villages are so tightly packed together that you can't actually get within range…

The notion of attack and defense ratings is fine with me; in fact, Civ4 has this very notion. It means that you have all kinds of interesting decisions to make when forming armies. Do you focus on faster-moving strong attackers, but potentially with weak defense? Or do you mix it up and have versatile units, at the cost of being a little slower? And of course, siege weapons – crucial in attacking a fortified city – are relatively slow, and highly vulnerable to cavalry in particular, and so must be defended and can slow down the entire army.

So, I actually view this as a feature, because it lets you make more interesting tactical decisions. If you know that your opponent tends to build mostly a specific type of unit (say, rock) you can build the counter (paper) at risk of exposing yourself to your own counter (scissors) – and now, you must decide if you want to be versatile or strongly counter what you think the enemy has.

I don't really understand how the waves work in conjunction with placing defenses: do you have to actively choose which units defend which wave, without knowing what is in the wave? Or do your defenders simply hang out at the village, dealing with each wave as it comes?

The LotR strategy game (Battle for Middle Earth) handles this by just having armies move from cell to cell as they're closing in on a target, and interception simply means walking your armies to the same cell: when they meet, you engage in combat. (That said, interception isn't a big deal in that game – it's a lot more about taking territory.) But, you can use the same mechanism: simply have armies be in cells, and when two enemy armies are in the same cell, let them have at it. You could still have camps if you wanted – the idea of defensive ditches etc. is very interesting – but you wouldn't necessarily have to.

You raise a good point about tightly packed villages, though. Perhaps it would make sense to force one settlement per cell, but then have that cell divided into sub-cells that armies walk through.

I'm actually quite fond of cavalry -> archers -> pikemen -> cavalry as the basis of a troop combat system, although of course I'd mix in more options as well.

But Travian's approach is to give each unit an attack rating and two defence ratings (one vs infantry, one vs cavalry), and combat is a pure number-crunching calculation. In a few rare cases it can be worth making a pure infantry or cavalry attack, but in most cases you simply hit them with every offensive troop you've got.

Because the attack rating is only used when attacking, and the defence rating only when defending, that means your defensive troops always stay at home (unless reinforcing an ally) while your offensive troops should be sent away before an incoming attack hits you.

To give an extreme example: A spearman has an attack of 10 and a cavalry defence of 60, while a paladin has an attack of 55 and an infantry defence of 100. That means if you attack with 10 spearmen, it takes 1 paladin to kill them (although he'd also die in the process). But a single spearman would kill that 1 paladin if the paladin were the one attacking.

The Travian approach obviously works, and people enjoy it, but personally I don't think they've introduced the sort of tactical decisions you mentioned. Combat is really just about hitting them with the biggest offensive army you can build, while leaving your defensive troops at home.


Your defensive troops sit in the village, dealing with each wave, one after the other. But you can send troops to reinforce other villages, and reinforcement/attack orders can be cancelled within 90 seconds.

Thus (for example) if you have 5 incoming attacks spaced 10 seconds apart (very sloppy!), then 5 seconds before the first wave hits you could send all your troops to another village - then 5 seconds later, cancel the order. The troops will have spent 5 seconds walking, and therefore require another 5 seconds to return, meaning they'll get home half way between the first and second wave.

So the clearing wave (the main army, designed to clear out the defensive troops) slams into your village and maybe does some damage, but there's nobody there to kill. Then your defensive army arrives home. Then the next 4 waves (mostly containing catapults, perhaps with a small escort of troops each) hit one after the other, and your defensive army easily wipes them out.

Also note that Travian has no graphical representation of troops moving around like Battle for Middle Earth does. You click on your target, select how many troops you want to attack with, and then launch the attack - which is added to an event queue. In fact the game has no animation at all.


Perhaps they could also camp in occupied cells - that would certainly make sense for sieges.

You have no control over this process? Can't you just keep the offensive troops away from the battle without sending them away? That seems like it could be an eventually tiring micro-management process.


Well, as you alluded to earlier, there is definitely something to be said for simplicity, and adding "interesting" tactical decisions can make your game complex enough that you weed out overly casual players. I think that one would have to be very careful when adding tactics.

One of the things I'm gravitating toward is making the AI that helps you manage your empire much smarter – perhaps you unlock smarter AIs with purchases (be it with real money as a business model, or game money like when researching tech trees). I think that a lot of stuff can be pushed to an AI, where you give high-level directives, such as allowing the cavalry to rush ahead of the pikemen or not, or whether the defenders should sally forth, etc.


Oh… ok. Thanks for the explanation. I can see how this simplicity is nice, but (again without having actually played this…) it feels kind of off-putting to me. If I show up at an empty village with a big army, I would expect to be able to simply burn it to the ground, with or without catapults.

In Civ4, siege engines play a very interesting role. First some basic rules about city defense: every settlement has a defensive bonus, which you can think of as a bonus to the defender's defense rating or a penalty to the attacker's attack rating (it doesn't really matter). This defensive bonus can be obtained various ways:
- city culture. Cities that your empire "cares about" (as strong cultural centers) get various bonuses.
- city buildings. You can build things like walls, which might give you a (say) 50% bonus (except against gunpowder).
- time. As time goes by, natural entrenchment occurs (up to a point).
- unit fortification. On a per-unit basis, you can instruct them to "fortify", and they will build up a defensive bonus (up to a point).
(In fact, every tile has mechanisms similar to this, but cities have more components.)

Siege engines can attack a city, but suffer the normal defensive penalty. The difference is that a siege engine deals collateral damage, so you can injure several units at the same time even if your siege engine is destroyed. If the city has, say, 6 defenders, then with two or three siege engines you can injure every single unit. In particular, this is useful because you can injure units that are strong against your non-siege units: when a unit attacks, the defending tile automatically picks the best unit to counter. This means that your cavalry are likely to lose against a full-strength defending pikeman. But, a siege engine – best countered by other units – can deal collateral damage to the pikeman, so that your cavalry won't be threatened as much.

But what is particularly interesting about siege engines is the ability to 'bombard' the city. This doesn't actually do any damage, but it reduces the defense bonus. For example, a city might have a defense rating of 60%, meaning that each defender will be 60% stronger. Your catapult might reduce this by 4% with each bombardment (you can upgrade catapults, and better siege weapons reduce the amount more). What this means is that your catapult is, basically, making every single defender "weaker" by reducing the overall defensive bonus of the city. A fully fortified city is basically impregnable due the defensive rating unless you have oodles of troops to throw against it, or are technologically superior.

For example, if your swordsman has an attack rating of 4, and the defender has a defense rating of 3, but with a 60% bonus, the combat is 4 vs. 4.8, meaning that the swordsman is less likely to win. But chances are that the defender will have more powerful defense, such as archers – who get their own defensive bonus in cities, that you cannot remove – with a "base" defense (after unit bonuses) of, say, 4 or 5. Add the 60% bonus, and you end up with 6.4 or 8 as a defensive rating, meaning that the swordsman is very likely to fail the attack.

Anyhow, I just think this is a really interesting use of siege weapons. You can't do damage without losing your siege engine (which I think is somewhat odd, but makes sense for gameplay), but you can reduce defensive bonuses which makes taking the city feasible.


Ah, that would work, and would be a good way to explain how a besieging army could cut off a city's resource flow – which is indeed part of what sieges are really about. (Most games represent sieges as lobbing lots of rocks at walls, but a siege could just as well have been about starving the city to death, assuming you have the time to do so.)

No, there's no way to automate it. It's not uncommon for players to deliberately destroy each others offensive armies in this way - for example by attacking when the player is offline, or by following their army home after an attack (timed to land within a second or two of the army getting back, so they've no time to dodge).

Travian has a lot of excessive micromanagement issues, but I think this is really more of an activity issue (which is also another major problem with the game IMO). If you want to be a top player, you need to put a lot of time into the game.


I proposed much the same thing on the Travian forums. It was shot down by other players, of course (comments about how there would be no point in them playing any more if it was automated, etc). But I think it would result in a far more enjoyable experience for the more casual players.

Travian definitely doesn't cater to casual players though - in fact, they're usually ground into the dirt by the hardcore players. The reasoning given is typically along the lines of "the top players are the ones who buy most of the gold, so they're the ones who should enjoy the game the most".


Well they have two types of artillery - rams for destroying the village wall, and catapults for destroying everything else. The other troop types can only kill troops and steal resources.

This actually works pretty well, because artillery is slow (more time to organise a defence) and expensive (in terms of both training cost and upkeep). Thus destroying someone's village is usually a slow and risky process. If regular (cheap and fast) troops could destroy a village then that would make it much easier to destroy other players.

In Travian the wall doesn't prevent troops getting into a village, all it does is provide a % defence bonus to every defending troop. That's something else I'm not so struck on - raiding a walled village is one thing (you can assume they're stealing resources from the surrounding farms and countryside), but I think a serious attack should require at least some sort of siege weaponary (even if it's just scaling ladders). The Civ4 approach (artillery vs walls) sounds more like the sort of thing I've got in mind.