whoops, thank you.
|
Test games played
|
Started the new game. Same as previous, except instead of Archmage I picked Alchemy as it works better for nomads (horsebowmen are a nice fast early unit worth mass producing in several cities if no alchemist guild is needed) and for werewolves (no alchemist guild for making werewolves either, and free mana so more left to raise skill - last game despite archmage I didn't have the luxury to spend all that much on skil).
Started on an island again but this time big enough for my 2 starting settlers. About same map position but this time land is nearby on both sides - to the east we have Aura, 8 Sorcery 2 Nature Specialist Archmage plus Vanilla H 6 Nature 4 Life Tactician Sage Master, and to the right we have Flandre 5 Nature, 6 Death Alchemy. Two of them are lizardmen, one is klackons, I plan to actually raze all cities for a while on any decent territory, and only keep those on small max pop locations the Nomads can't use well. I send my first few wolves to the east, so I was able to raze two major AI cities right away and several more belonging to Aura. So far I'm not at war with the third wizard but they are Maniacal so we might end up in one eventually. The area I razed has 2 mithril, so I'm planting 4 Nomad cities on that area all with the plan to pump out horsebowmen. Population numbers are fairly good too, 12,14,16 and 23. That and the wolves should be enough for most of the early and midgame. The only way this might go wrong is if I get back attacked by Flandre but I'm planning to build enough horsebowmen to defend myself everywhere. Maybe I should keep the next Klackon city, build some engineers and then let the AI take it back so I can raze if I don't find one in a place I don't want to use for nomads. The Myrran wizard is Seravy (99% he will be Life Nature). Aura seems to have plenty of Sprites which might be inconvenient for wolves, I'm going to get those horsebowmen ready ASAP. ...unsurprisingly an equal army type war declaration arrived from Flandre. I don't have enough to attack there as well, but I can at least focus on my defenses until I do. Thinks going much better this time. The AI players are doing some wars and it makes a huge difference, no stacks attacking my cities this time. Also last game my early enemies both had nagas and one had wolf riders, so holding any city was really difficult unless I had serious garrisons which I clearly couldn't have that early. This time there are much less attacks coming, and they are not from unstoppable 3-5 waterwalking/pathfinding movement units. It also feels much faster, I suspect I play a lot slower when recording since I try to be more careful (and usually end up playing worse anyway somehow...).
Uploaded part 4 and part 5 of lunatic walkthrough.
Not sure if i will do lunatic no save/load run ^_^. Although I am sure it is possible. Might even be better to start with 8 necro books to have even more tempo. Although no web/resist elements hurts. But that tempo can help crush those wizards even before resist even matters. Or there is a good chance to find new books before i need those spells. Something like 8 death, alchemy, myrran, spellweaver.
Don't worry about it being Lunatic - even Master or Expert videos are plenty of useful information.
Lunatic, as is, is pretty much kept around for testing purposes (if someone does beat it, it's a very good way to figure out which features in the game are unbalanced) for the time being, and will eventually be made easier (once the lower difficulty levels are already adequately tested, as I suspect they are also too hard currently.)
One particular interesting point this game that made a big difference, the Sorcery wizard picked Sprites over Nagas.
Now, Sprites are definitely more effective for a lot of things - defending the fortress, and stopping werewolf invasions in general (werewolves can't regenerate if a flying unit forces retreat) but they are absolutely horrible at being a threat to cities compared to Nagas. Nagas with their fast overland movement and decent strategic strength can easily reach and destroy any weakly defended point the player has, slowing down the possibility of early expansion massively. (You seriously need at least 2-3 wolves left being in each city, or some real armies like 5+ horsebowmen.) Meanwhile Sprites...don't do that. They "oh scary, if that attacked my city I would definitely lose"...but then you put a single werewolf into the city and they are afraid to attack. You put 2-3 horsebowmen and same deal, plus if they somehow attack, the horsebowmen actually obliterate them. And once you research Mana Leak, you can mostly ignore sprites all around the map, as long as you are careful enough to not attack into them. Of course, Nagas are bottom tier against Death, with just a zombie you can turn an entire unit of naga to undead which will then be super effective against other Nagas since it'll be the same unit but immune to poison, buffed by darkness, and as a human, taking better advantage of First Strike. Despite all of that Nagas still felt more of a threat than Sprites, so I wonder if we are ok with this AI preference or want to change it. (if we can, not sure since we can't have a loop of dependent choices form, they have to still be possible to arrange in a well defined order) Another reason why this might result in such a big difference is AI wars. Sprites are extra vulnerable and likely get stomped out effectively by other AI in battles, while Nagas are very unlikely to fall that way (as they tend to spend a significant amount of time on water tiles, and have a decent enough strategic rating). Quote:This one always amuses me. I haven't built city walls in CoM yet, not since I first found it, back before version 2. Poor AI.Would it be better if they only built one (mandatory or otherwise) if they had at least X ranged unit in their garrison? It does seem detrimental to melee armies defending against ranged units, but it's fairly good in all other cases (melee vs melee, ranged vs melee, ranged vs ranged).
August 30th, 2018, 21:05
(This post was last modified: August 30th, 2018, 21:06 by Nelphine.)
No, AI should always build it in their fortress. If a city wall is going to make any kind of difference, then often the fortress lightning is a significant part of their offense - in which case the extra survivability from The Wall is far more important than possibly losing out on some melee damage.
Elsewhere.. hard to say. But given that you and I have differences of opinion in what makes a good ai garrison anyway, it's probably fine to leave as is elsewhere.
I would say, the au should connect all its own cities. It should connect one of its own cities to another players cities iff it has no cities of its own it can connect to or its own city is not yet connected to a city of another species and the city is connecting to is another species (assuming that's actually something that matters for the trade bonus - it's been a really really long time since I actually looked at the mechanics of how it works). Ideally it would check this in such a manner that the AI only ever has to build roads from its cities to one other city on a given continent.
In all other cases, the AI should not build roads. |