Fans suggestion thread for future releases

[Dooz]

What is that understanding based on? There's proof AI retrains in hotseat campaigns, can see it yourself.

[ptolemaiosVI]

Are settlement change suggestions welcome? I always wanted a Pelusion settlement (with garrison script) blocking Egypt from land invasion when playing as the Seleucids. Also, if I were to do it on my own, what would be the easiest settlement to substitute for Pelusion? thanks

[aaronp]

I think the current movement penalty is no longer enough for client rulers. I like the idea of them being able to move farther, since before they often couldn't effectively move to engage threats in their own territory. But as it is now, the difference is too slight to tell visually that you've got a client ruler selected. 50% instead of 20% might be better and give a good visual indicator that you shouldn't be sending this general out on campaign.

[Dooz]

That's a good point. Do you happen to know how to edit client ruler movement penalties?

[aaronp]

Haven't got a clue. Lets see. Grepping around I can see them in the export_descr_character_traits* files. For some reason it's duplicated in three different files. Anyway there are two different entries under ClientRuler for each file. The first is for I guess rebel governors that have Effect MovementPoints -100 and the other one is for I'm guessing player client rulers and that one has Effect MovementPoints -4. But I have no clue what the -4 means. Multiples of 5? So is -10 50%? Not sure.

There is also the description for the trait in text/export_vnvs.txt which is where the pop-up telling the player the penalty is 20% comes from

[Dooz]

Nice, thanks for looking into that. Well if the 20% popup is accurate, perhaps each "point" of the -4 is 5%? So maybe a -10 would give a 50% reduction. Although movement points seems to suggest a more direct value, and what we're looking for might be in a traits file instead or as well, but who knows (as between us). I'll have to fiddle around a bit myself and figure some stuff out.

[makute]

This is my understanding of the Med2 engine.

What is that understanding based on? There's proof AI retrains in hotseat campaigns, can see it yourself.

Jurand is right. I'm not a mod expert, but in my experience after years of playing M2:TW, AFAIK AI doesn't reinforce.

That's the reason why AI generals in Stainless Steel have traits like "Field Medic 40", to compensate for that limitation.

[Dooz]

Ok, but what specific experiences and observations are you basing that on? And how is the fact that AI factions have units retrained in their recruitment report scrolls explained?

In a hotseat campaign, you can assume and relinquish control of any faction. I start a campaign, relinquish control of my one picked starting faction (i.e., set to AI control), and let the game automatically run for many turns to see the game develop and test how various tweaks or whatever I'm working on are playing out.

Every couple of hours I stop by and alternately take control of various factions to see their armies and the map and all that. When those factions' turns come up, and their recruitment report scroll appears (because I'm now playing that faction so it's like a regular turn), there are retrained units included in them just like if you're playing a regular campaign and retrain units yourself. That scroll is based on recruitment queues from the turn before and means the AI queued those units to be retrained, from before I assumed control.

If there is another explanation for this, I'd like to know it. Do hotseat campaigns have a completely different system that allows AI to retrain while regular campaigns don't? Is there anything that can be pointed to to show that? I'd take any sort of proof more than just a feeling based on playing against AI and thinking that it seems like they don't retrain because they always seem to have so many depleted units.

[makute]

Ok, but what specific experiences and observations are you basing that on? And how is the fact that AI factions have units retrained in their recruitment report scrolls explained?

Tbh, I can't give a conclusive answer. But every mod out there had to find a alternate solution to AI retraining and reinforcing because that's a very well known issue with the TW engine.

I have no experience with hotseat games, but I doubt the AI is different in that particular case. Maybe what you're seeing are scripted reinforcements?

[Domaje]

AI retrain units in regular single player campaign. If you doubt, just look at the remaining eleutheroi province after turn 100 or so, you'll see units at full strenght and with gold chevrons thanks to repelling several assaults over the course of the campaign.

And I don't think the AI can merge units, because it can't click and drag... yes that's seems silly. A simple experiment is : next time you see a stack with 2 similar units and that their combined number of men is less than a regular unit, you'll get your proof.

[Dooz]

What I'm wondering is what exactly is known and how exactly it came to be known.

Not sure what you mean by scripted reinforcements. They're not new units, those are also separately listed in the scroll as they are in regular campaigns. I mean units with (retrained) in parentheses next to them, just like you see the turn after you retrain units.

Domaje, I think they can merge and possibly do (I'll have to pay more specific closer attention in future to see if true). You don't need to click and drag, M is the keyboard shortcut for auto merging all available units in a stack.

[makute]

It's known that the M2:TW AI never retrain units, and it became widelly known because modders had to implement their own solutions to this issue.

That said, maybe the increased movement made the AI more willing to retreat home and retrain its units as a side effect. Something that weren't able to do before the increase.

If that's the case, we still need an answer as to why the AI doesn't detects the need to retrain units and plan his turns accordingly.

[Dooz]

It's known that the M2:TW AI never retrain units, and it became widelly known because modders had to implement their own solutions to this issue.

How did modders know AI never retrains units? I'm suggesting they didn't know, they only believed that was the case based on anecdotal evidence. And that AI does retrain, based on direct evidence, so knowing it doesn't retrain is impossible (because it does).

Unless there is some factual explanation for the hotseat recruitment scrolls showing retrained units. If that scenario has some knowable reason to exist other than the AI having queued those units to be retrained the turn before (maybe the EBII team scripted it to make that happen), then there can be room for doubt. Until then, I'm satisfied that AI can and does retrain units.

[QuintusSertorius]

The AI definitely merges, I always assumed they didn't retrain but it's difficult to observe what they're up to when units are in garrisons.

We didn't script anything there, though.

[Dooz]

Quintus, are you aware of any differences in AI or scripting or anything at all between singleplayer and hotseat campaigns? Or do all files load and execute properly and identically in both? Any reason to believe they wouldn't?

[z3n]

Well like Dooz said, you can see for yourself in the hotseat campaigns what the AI chooses and does. A lot of things are very dependent on mod setup. There's variables to consider, for example, is the overall strategy invade_buildup being used in the mods? This decision could result in retraining of troops; the added variable of movement range can also come into effect; along with cost, availability, and defense priorities within a region. These are very mod specific things, so I don't think blanket statements work.

In my opinion, since the feature has been in since RTW it works in basically every mod unless they were doing everything possible to prevent the AI from retraining units successfully. A more accurate statement is probably that "the ai doesn't retrain units as often as a player".

[QuintusSertorius]

Quintus, are you aware of any differences in AI or scripting or anything at all between singleplayer and hotseat campaigns? Or do all files load and execute properly and identically in both? Any reason to believe they wouldn't?

The campaign_script isn't really compatible with hotseat; with the exceptions of the common stuff like the money scripts and logging, a lot of stuff will terminate when you start, then if you change faction that will too.

[Dooz]

Do you mean terminate only for the faction you start as, or for the entire game and all factions? I know the starting faction always has a hard time, but the rest progress normally far as I can tell. I've been using the hidden New Guys faction to test stuff so nothing would interact with the actual AI factions, does that insulate me from any of these potentialities? I also reload saves from before taking over other factions to check specific situations, in case of something just like this.

[z3n]

Far as I'm aware the most recent mod testing indicates scripting in hotseats works the exact same
https://www.twcenter.net/forums/show...1#post13975923

Some time ago one of our team members indicated that there were issues with LocalFaction testing in hotseats, so I believe the decision was made to simply terminate most of the scripts in hotseats rather than test for them. As the scripters on our team mainly wrote them for single player rather than hotseats. Some of them can actually unbalance a hotseat as well, so it's probably best they're terminated.

[Dooz]

I'm not sure I follow, I don't know anything about this stuff lol.

In a nutshell, as of the current release, if I start a hotseat campaign as the non-playable The New Guys faction and don't ever switch another faction to human control, will the AI progress according to all the same influences as a singleplayer campaign?

Edit: barring whatever influence a regular player would have in a campaign, I just mean mechanical, script and mod files, that sort of thing. Whether any features of the mod would be missing at all in the above scenario.