The same way you did.  Shrink, Blind, Curse and then pound on it with fire-resistant units until hopefully it dies.

Does the faction AI cast any of those?

usually my dragons cast fear, then waddle up to within two squares of the army and breaths fire onto the entire enemy army and FRIES most if not all of the army, and the accompanying archers fire one or two volleys of arrows and battle over

A champion can take a dragon down just fine, and so can a well designed player army.

But I don't think I've often seen the AI field an army that can take on a dragon. Yithril may, but anyone else?

OK, good to hear. Since this dragon is one-shotting everyone I'm assuming we're talking late game armies, which is fine. I think the evil, treacherous Relias currently has "3/4-game" armies, if you will. But I am skeptical that the AI will put it together that he needs to prioritize this and design armies and equip heroes around that fact. (I'm on challenging difficulty all around.) Does the AI design units or just use what already exists (including player-made units--it was pretty cool to see units from my last game coming at me, even if they were only I-don't-have-crystal-so-lets-modify-the-default-unit-a-tiny-bit)?

Also, since you guys clearly are better at this than I, what would you put together for this? With slow-shrink-curse-etc (with high spell mastery) and grow-haste-heal-etc on some melee champs even I could take down a dragon, so let's do it with mainly troops. What kind of troops if you only had one supporting champion (of any type)? Two champions? None?

Or do the end-game troops with the best weapons and armor basically allow you to go toe-to-toe? I've never made it that far before winning (in my one previous game). What is the minimum tech you would need?

I would go for well-armored or high-initiative archers as the mainstay of my army.

If I go for well-armored archers, then I want to firstly cast hast on my archers and secondly spread them out so that the dragon can pick one to beat up for a while, then go on to the next. I would also try to slow the dragon after hastening my archers, but only if 1) I think there is a reasonable chance that I can slow the dragon

If I go for high initiative archers, I would also include a fast, high-initiative, low-health melee (or maybe ranged, but I want that initiative) unit for the dragon to chase around for a while while the archers turn the dragon into a pincushion. I would still scatter the archers, but I might try to slow the dragon before hastening the troops if I think there is a good chance that the spell will stick. As for why the melee unit is low-health - I don't intend for it to fight the dragon, I want it to distract the dragon, and the AI seems to prioritize low-health targets over high health targets (which incidentally makes your supporting champion a reasonable candidate for this role, since it probably has less health than the archers and, since it's a champion, can probably be made to achieve a higher initiative than most of the stuff you can train for less than the cost of training anything - just get a dagger from somewhere, and anything else that boosts initiative). Thus the melee unit, if it is a trained unit, will probably be a minimum-size group on horses (for the movement) or wargs (for the initiative) armed with daggers (for the initiative) and equipped with a minimal amount of armor (if any) to avoid any encumbrance penalties.

Technologies that I would want: some form of ranged weapon, whether a staff or a bow, and the higher tech the better; a mount, probably a horse for the mobility, but wargs are good to counteract the initiative penalty on a bow; and some form of armor. Larger group sizes are also better, as your troops will deal more damage that way, but it isn't necessary and may make it more likely that the dragon will go after your hero first - which I would generally want to avoid with a supporting-type hero. Any technology that gives equipment

Note that it is possible to beat dragons using troops equipped with only leather armor (or Robes of the Aegis), spears, and horses - I've done it, but I put it off until I had the 3rd of the four group sizes and wasn't all that concerned about losses, as I had about 20 cities and had focused research in the Civilization tree and thus had a large income with which to replace losses. Lower-tech armor (or even no armor, though that severely limits the utility of your troops against most other things) may even be better than higher tech armor, because you don't get the same initiative penalties in leather that you will in plate, and plate doesn't do any more against a fire attack than leather does (assuming the dragon uses the fire attack) - acting more often is better in this case than taking less damage from melee/ranged attacks, because 'less damage' only leaves you as 'mostly dead' instead of dead, and you may do better by training lots of weak sacrificial troops and sending armies off as if they were attack waves to wear the dragon down than sending that really expensive army you spent all game and all your resources building off to get killed by the dragon (you could also combine the tactics).

Essentially, the minimum tech I would need is 'None' - I just need to have enough rapidly-trained, unsupported, weak troops to wear down the dragon or otherwise occupy it until something of mine can actually pose a legitimate threat to it. Magnar works really well for this as you can train up a bunch of cheap slaves quickly (preferably armed with spears or better, but anything that can consistently do at least one damage which you have in enough quantity to deal more than the dragon's level worth of damage each strategic turn will work - I like spears because they ignore some of the dragon's defense and as such are much better than any other early weapon against high defense - they probably still only deal fractional amounts of damage per spear, but it's a larger fraction). More importantly, slaves cost you nothing except time (if you use the build queue) or a single payment (if you rush them), and this can be used, with a sufficiently high-production city or a sufficiently low-cost slave, to make two slaves per turn from a single city - rush one for the miniscule cost (relative to a real unit) of a slave, train one properly in the city queue, repeat on next turn, and if it doesn't complete training on the next turn, rush a slave that hasn't begun training yet, since you won't lose any production time spent on the incomplete unit. This can be a bit expensive, but it will certainly speed your troop production and works with any unit - it just takes longer with most normal troops and is more expensive, and if we're sacrificing units to kill a dragon I want those units to be cheap

Considering low tech (and, possibly, low-level champions):

Useful Spells - Curse and Graveseal (in Death Magic) are also great against a dragon (assuming you can get them to stick) if you've got low-end troops, because they increase the damage suffered by the dragon (especially Curse). I would not use anything out of Life Magic against a dragon if I have low-end troops, because Shrink just makes the dragon harder to kill without significantly improving the survivability of my units (Yay! I'm half-dead instead of all dead! Heal me so that I can die on the next hit anyways!), Heal is useless if a) the dragon can kill your units in one hit or can kill the units in two or three hits even if you heal them, and Growth is only marginally useful on low level weapons, since low level weapons have attack values ranging from 4 to 9, add 50% and we get 6 to 14, which is still much less than the 20+ defense of a dragon - although it is more useful than anything in Death Magic aside from Berserk (they are going to die in one or two hits anyway, who cares about the 1-point of damage per turn?) unless you have a high enough spell mastery to get the curses to stick to the dragon.

What magic to have on the supporting Champion? Air. No other school can give you as much with a low-level caster as Air can if you face a dragon. Haste your entire army as soon as you can, spread your forces out so that you can get the dragon to run around in the middle after something fast in the center if you have ranged troops or sacrifice your melee troops individually to the dragon. Life + Air might work if your troops can survive a hit and you can rotate weakened units out of the line fast enough to keep them from dying and long enough to heal them to full health, but it will cost a lot of mana and you'll need Wellspring on your mage to do it properly - heal is alright if you only need to rotate between two or three units, but you probably don't have troops strong enough to go toe-to-toe with a dragon for more than one or two attacks by the dragon, even late in the game.

Unless you have a high amount of spell mastery or are very lucky on your spell resistance checks, curse-type spells (Slow, Wither, Curse, Blind, Shatter) and direct damage spells (unless you have a lot of the shards which increase the spell damage) are wastes of mana, because the dragon will either resist the spell (which wastes the mana for curse spells and halves the effect of damage spells) or it won't matter much anyway - Curse lasts three turns, but the dragon probably has enough health to survive those three turns and you probably blew all your mana getting that first Curse to apply, anyway, and once Curse ends you're back to doing about 1 damage each time a unit attacks since you don't have good weapons; Blind and Slow prolong the battle but won't save you in the end, though they might allow you to weaken the dragon more than you would have otherwise; Wither, unless you have a lot of Death Shards isn't going to significantly impact the dragon's attack value; and Shatter is a relatively high-level Earth spell that costs a fair amount of mana, does a little damage, and reduces target armor slightly - but since your weapons are crap it doesn't matter, because if the dragon had 20 defense before being Shattered, it has 15 afterwards, and your weapons are not that much better against 15 defense than 20 - I think you're better off hoping that Curse gets gets through the spell resistance than that Shatter gets through, unless you've already lost most of your army. Damage spells are equally useless because you will probably go through most of your mana to deal one or two damage per cast, which you could have done by whacking the dragon with that dagger you found in the spider nest and the result would have been the same but at a much lower cost (assuming we have a low-level champion and few shards, or a high-level champion without Flame Dart without many shards backing him up). Damage spells and Wither become much more viable if you have the shards to back them up, because one-half of whatever large amount of damage you would have dealt had it gotten through resistance is still much better than the one or two you'd deal with normal attacks, and a Wither lucky enough to stick to a dragon when you have a lot of Death Shards might actually be worth sacrificing the champion to cast it (and you probably will lose the caster - that casting time immobilizes the caster for two turns, and the caster is probably the thing on the field with the least health, so is likely the target of the dragon). Conversely, curse-type spells (other than wither) and Flame Dart become much more viable with high level champions as the curses are more likely to stick and Flame Dart does more damage, which is more likely to deal its full amount since your spell mastery is likely higher. Anything with a casting time is probably not worth casting (this includes Wither) unless you have a high spell mastery (to make the spell affect the dragon at its full value) and a lot of shards of the type that make that particular spell better, because any spell with a casting time probably means that you have one and only one chance at applying it to the dragon, and then your champion is dead for the duration of the battle (if you took Path of the Defender, then I don't think your champion really counts as a support-type champion, unless you count the army improvement skills - Path of the Defender is more the 'I want to go toe-to-toe with anything the world can throw at me' path than the support role).

It is sort-of possible to kite a dragon. You need to have both high initiative and a high amount of movement on your troops in order to do it, and you almost certainly will not be able to do the 'strike and run away' kind of kiting that works for melee units. What works best is to have something fast and weak to tempt the dragon into chasing it around for a while, and keep the rest of your forces scattered around the edges of the map (these troops should have ranged attacks, otherwise there is not any point to giving the dragon something to chase - it will fight anything that is next to it on the beginning of its turn or that it runs by trying to chase your bait unit). Keeping the troops separated means that the dragon may have to spend two or even three turns to come eat the next group of soldiers, and you can spend that time moving the others even further away (if possible). Since the monster AI will usually target the enemy with the lowest health first, keeping something fast with a small group size (say, a dagger-armed group of 3 horsemen with initiative bonuses) will give you an opportunity to lead the dragon around the middle (or edges, if you prefer) of the map for a while while the rest of your forces pepper it with arrows/spells/whatever comes out of the magic staffs.

It is certainly possible to wear down the dragons, with enough troops and basic spears. It is very expensive and time-consuming, but it is possible, and you don't necessarily have to do it all in one turn, assuming that you either engage it far enough from your city that you have time before it reaches the place or you don't care if it destroys a city or three. However, at the time I did it, I already had the 7 people/group technology (I think this one is now 5 per group), so my spearmen (who were on horses) could survive a hit or two and managed to deal perhaps 5 to 10 damage per hit. With smaller groups (and, possibly, without mounts) this becomes much worse, since the damage dealt per unit goes down because it a) dies faster and doesn't do as much per hit.

Also, if we assume that the dragon has 20 defense (I don't remember how much defense a dragon has, but that seems about right), a spear (7 attack) should be able to do a maximum of about 2.4 damage per hit. This probably ends up averaging at about 1 damage per hit when we consider 'Blocked' 'Clink' and proper hits, multiplied by the number of troops in the unit. By contrast, a dagger (6 attack) will do at most about 1.4 damage per hit (probably averaging about 0.75 damage per hit), while a warhammer or axe can do at most about 2.8 damage per hit (probably averaging about 1.3 damage per troop in the unit per hit). Spears, of all the early-game weapons, are probably the best option because they don't have any initiative penalty and daggers don't do enough damage despite having a relatively high speed, unless you want to hope that you get lucky with bash (but I think dragons are immune to that). This also doesn't include the dodge-vs-accuracy aspect, which will bring down the average damages a bit, but I'm a bit surprised that you were only seeing about .75 damage per unit each attack when using longbows (although that could be about right if dragons have defense nearer to 40 than 20, I don't recall).

The situation of a Fell Dragon against a Storm Dragon and supporting army is entirely different than the early game situation that the second part of my previous post referred to. In the situation that you have units that can survive multiple hits from a dragon (such as another dragon, a high-level hero, or a large group of normal troops, preferably at high level), healing becomes much more valuable. It only isn't valuable in the case where your troops cannot survive more than a few hits from a dragon regardless of healing, which is generally the case with early-game units.

I would say that the only practical way to have enough troops to wear down a dragon is if you're Magnar and have been training huge slave armies from the start of the game, or if you're rich enough to support large armies (though if that's the case, you should also be rich enough to support quality troops that can go against a dragon on a relatively even footing).

60 groups seems a little low to me. I suspect that with an early game army (say, four units of 3 level 1 spearmen), you should probably hit the dragon about half the time, and lose one unit per time the dragon acts. In this situation, I think the dragon would kill a group of spearmen, then the remaining troops would attack to deal 9 damage (assuming 1 damage per spear), which we'll halve due to the 50% hit rate and round to 5 damage. The dragon will then kill the next group, and your next turn you can do three, and then two, for a total of about 10 damage by four groups of three spearmen each battle (these are estimates, though and will be greatly thrown off if dragons have more defense than I remember or you go against one which casts Fear on your army at the start of the battle - in that case, find yourself a place to hide and stop wasting armies, because you won't hurt it enough anyway). By the above estimate, I would think that 120 groups of 3 spearmen should be sufficient - if they can hit about half the time for 1 damage per spear, and act as often as the dragon does. Also - don't bother giving these guys armor, it won't help them survive against a dragon if they are level one 3-man groups. If you cannot do at least one more point of damage to the dragon each turn than it has levels, don't bother with the attempt to kill it this way - it'll be full health next turn and you won't gain anything for your efferts. Instead, go kill off the AIs or hope the dragon doesn't lay waste to your entire nation.

I would certainly say that your estimate of 5 attacks per group is over-optimistic - the average I expect is two attacks, and one of the groups dies without ever making any attacks (assuming basic 3-man groups). My expectation is 2.5 damage per group with the basic spear and 3-man groups, but it'll be much better if you can get larger groups, because each group will do more damage per attack and, assuming the group is large enough, the group might actually get in an extra attack against the dragon before dying.

Because of the above, Magnar is really well suited to this - slave units are cheaper and faster to produce, and it doesn't matter too much if you lose them because they are cannon fodder anyway, and since they don't cost any wages, you can train them continuously from the start of the game with the only problem being that you then have no city development (but who cares - with your massive army of weak garbage you can go attack the AI and watch them crush your slaves! Yay! They'll be buried under the corpses and you can take the cities they spent all game making nice enough to be worth taking!).

I have also seen a storm dragon killed by two jugernaughts and NO champions, one of the jugernaughts died, the other was at half health, but managed to maul the dragon to death.

harpo

Pomally - unless your 0.75 damage is an average for each successful hit rather than an average over all attacks (or is an average total damage dealt per attack by the unit instead of per bow), 0.75 damage is not unreasonably outside of my estimates for spears. You'll note that my estimate was 1 damage per spear per successful hit, and I assumed that I'd miss half the time, which gives an average of 0.5 damage per attack per spear. Applying the same 50% miss rate and an assumption of an average damage being half of maximum on successful hits, we get that longbows should average 1.5 damage per bow on each successful hit, or a maximum of 3 damage per bow. Using an attack rating of 11 for a longbow, and assuming that the dragon has 20 defense, the damage formula shows a maximum damage of 2.4 per bow, which is a about the same as what spears can do. This does assume that only about half of your attacks were successful, which only you are in a position to know (and yes, for the time and effort it takes to get them, longbows suck).

The main reason why spears can perform similarly to longbows is that they ignore one-third of the defense of the target, whereas the bow faces the full defense of the target. So when attacking a 20 defense dragon with bows, the damage calculations use the full 20 defense, but with spears it is as if you are facing a dragon with only 14 defense.

The the maximum damage formula, unless I am mistaken, is:

            maximum_damage = (Attacker_attack)*(Attacker_attack)/(Attacker_attack + (Defender_defense)*(1 - Attacker_defense_bypass_percentage))

After that you get a bit of stuff with random number generators to determine if if hits, misses, is blocked, gets a 'clink' (deals one damage), or deals damage in the range given in the tooltip.

harpo99999: That sucks. I've been terrified of juggernaughts in both the games I've played, but somehow Yithril always gets defeated. My only strategy against them is to hammer them with everything I've got before they can attack. I'd say that insane damage + maul + splash is overpowered, but they've been among the only late-game challenges I've had, so I wouldn't change them if I had the option.

joeball123: I think you are correct. I had another storm + hero + archers vs ashwake, and they were doing a pretty solid 2 damage for four shortbows (I checked), so 0.5 per shortbow. I still think that it would be near impossible to do 300 points worth of damage that way if all your troops are getting one-shotted, especially if your army is afraid for the first three turns.

Also, I just conquered a city 2-cities-deep into Altar's territory and it has an (unoccupied) dragon roost in it's territory. I understand that Relias has made it to the end of the weapons tech tree (and I haven't) but it seems relatively inexcusable for him to to have beelined that tech. Oh well.

Cuurently winning against Yithril. I have two armies each w one SD along w some champions each along  with ASOK and Warg Archers - Yithril started with stream of Juggernauts with support troops. About 40 juggernauts later the steam is now a trickle of skirmishers, while the dragons have fond memories of roast juggernuts.