I want to limit that to mounted melee though. Normal melee defending on the frontline can gain XP well enough, it's just the mounted ones that have trouble leveling up.
Humble opinion about mod
- Thread starter GoreBush
- Start date
Its not a matter of XP per say, its a matter of increasing general offense over defense...which is a net weakening of the human compared to the AI.
let's continue the discussion at https://github.com/LoneGazebo/Community-Patch-DLL/issues/9524
This is like the 28563th topic about the doom carpet late game warfare, and funny enough I still didn't see anyone suggesting the obviously simpler "add AoE to late game high cost units".
Admittedly I was also sitting on my hands despite suggesting I can take care of the change myself (in my defense, work was pretty busy and I didn't even have much time to play since 3.0), but shouldn't there someone else who can simply give all bombers and guided missile and artillery free splash dmg promotion and test it out on late game war ?
The idea is sound both historically and gameplaywise, as you can both simulate how modern warfare rely much more on large area bombardment from out of sight rather than the slow crawl trenches warfare slug fest , while also taking care of the doom carpet. And you don't even need to change anything to the AIs, it's a simple "more dmg to both sides" without actually give them any bonus/malus to CS/hp that would affect small scale skirmishes.
Admittedly I was also sitting on my hands despite suggesting I can take care of the change myself (in my defense, work was pretty busy and I didn't even have much time to play since 3.0), but shouldn't there someone else who can simply give all bombers and guided missile and artillery free splash dmg promotion and test it out on late game war ?
The idea is sound both historically and gameplaywise, as you can both simulate how modern warfare rely much more on large area bombardment from out of sight rather than the slow crawl trenches warfare slug fest , while also taking care of the doom carpet. And you don't even need to change anything to the AIs, it's a simple "more dmg to both sides" without actually give them any bonus/malus to CS/hp that would affect small scale skirmishes.
There's already so many changes for this congress, and the end of the year is typically crunch time for a lot of people. I don't foresee many of the current proposals being sponsored, let alone implemented in a timely matter, so a lack of suggestions isn't a concern of mine. I agree though, moving Splash on siege a little earlier is probably a healthy move, at the very least it would help give siege a use outside of city busting (and loosely anti-navy).
Because the goal is to not to "deal" with the carpet of doom, people are asking to remove the carpet of doom.
The human can already deal with AI hordes are high levels, which is why we have hordes to begin with. Adding splash would just make the horde easier to deal with....which defeats the purpose. The real ask is to provide equal or even greater challenge with fewer units, and that's why its such a tricky problem. Put another way, what people are "really" asking for, is a fundamental change to Civ 5 warring, to make it where the humans "immortal unit" play no longer lets them beat armies 2,3,4 times their size....so that we can remove the armies 2,3,4x the human's size.
But it also goes back to a fundamental question taht has been somewhat brushed aside.....does the AI actually have "hordes" now? Sure a few versions ago, it was carpets of units, but the AI has gotten a lot of key nerfs since then, the biggest one is no ABC production bonuses. That's a big deal, and will definately mean less units for the AI. So is it enough, is the AI really providing carpets of units still, or is that just people's old expectation and the problem has already been addressed?
The AI does seem to have far less units now. Unfortunately this has also made it quite a bit easier.
You can never program tactical AIs to be as good as human players, ever, thus there's no other way but accept that AIs must have more or stronger units to even out the playing field.
It's either giving them more units and deal with doom carpet, or give them much stronger units, which I think, personally, would spark more "unfair advantage" complains than doom carpet, since player can't really make use of terrain for tactical advantage to even things out like with doom carpet. Imagine the reverse of the situation the other day where 1 knight with 3 skirmishers giving flanking bonus can 1 shot another knight, means to even it out you have to gives the AI bonus CS of similar scale, and their knight with 1 or 2 units flanking can 1 shot your knight. Have fun with that.
False. It's possible, but very hard. RL could help with that.
Really? @ilteroi Is it just reduced search of best actions? Maybe it could be configurable.
Search depth is now configurable in DifficultyMod.xml under "AI Behavior Modifiers".
Cool! Do you remember the previous values with better AI, but longer computation?
the impact of those search depth / breadth settings was never systematically evaluated. if anyone wants to do it, that would be great.
the other problem is that there are lots of corner cases but no test suite. meaning there will be regressions from bug fixes.
that said i'm pretty sure the current version is the best version of the tactical AI overall. but if you have examples of bad behavior then open a issue on github.
the other problem is that there are lots of corner cases but no test suite. meaning there will be regressions from bug fixes.
that said i'm pretty sure the current version is the best version of the tactical AI overall. but if you have examples of bad behavior then open a issue on github.
Tactical AI is probably the strongest part of the current AI. They are pretty good at moving units within combat, withdrawing damaged ones. Could do with better focus fire, which would be the next level up but overall it is fine. It is just everything else that is an issue.
usadefcon1
Prince
- Joined
- Sep 24, 2016
- Messages
- 503
Code:
<TacticalOffenseMinCompletedPositions>17</TacticalOffenseMinCompletedPositions>
<TacticalOffenseMaxCompletedPositions>23</TacticalOffenseMaxCompletedPositions>
<TacticalOffenseMaxBranches>3</TacticalOffenseMaxBranches>
<TacticalOffenseMaxChoicesPerUnit>3</TacticalOffenseMaxChoicesPerUnit>
<TacticalDefenseMinCompletedPositions>7</TacticalDefenseMinCompletedPositions>
<TacticalDefenseMaxCompletedPositions>23</TacticalDefenseMaxCompletedPositions>
<TacticalDefenseMaxBranches>2</TacticalDefenseMaxBranches>
<TacticalDefenseMaxChoicesPerUnit>2</TacticalDefenseMaxChoicesPerUnit>Am I correct in my assumption that increasing TacticalOffenseMaxBranches and TacticalDefenseMaxBranches will give the AI more time to think?
The fact that early war is still the best method to snowball in higher difficulty despite not having accumulated all those elite high lv units shows that even if current tactical AI is the best it can be, it's still nowhere near a good human player, thus my original point. They need an edge, either quality or quantity, and I would prefer quantity plus a way to deal with the problem that came along (doom carpet) than the other option.
I don’t know if it is, I think medieval war is. Ancient classical just costs so much in terms of infrastructure.
Medieval your a lot more stable, i think war ar this point has high value
After wall is up it's almost too slow to take cities unless you have UU in early eras, and capturing cities early also help skipping cost of training an extra settler or two.
Medieval you have a good boost from knight and stables so maybe it's also easier, but I don't think it's as good as archer/spearmen rush before wall.
Medieval you have a good boost from knight and stables so maybe it's also easier, but I don't think it's as good as archer/spearmen rush before wall.