Events Thread

[Boco]

What do you mean by strength? Total civ power, as mentioned above? If so, I haven't seen any evidence - but that's only because everybody seems to be behaving themselves (WRT MU) in my scenario.

No, I wasn't thinking so globally. Perhaps I should.

I have the impression that each unit makes a decision whether to attack as directed by a MU command. That decision appears to be depend on at least the combat strength ratios of attacker and defender. I've seen many examples of the 'directed' unit in question refusing to make a suicidal attack. It will if there are other friendly units around to make follow-up attacks, but it won't if it's the only attacker around. Have you seen behavior to refute this?

Or, what I was hoping for was a magic Catfish answer that has seen this and worked out some examples that demonstrate some factors involved in this decision.

 

[Catfish]

I have the impression that each unit makes a decision whether to attack as directed by a MU command. That decision appears to be depend on at least the combat strength ratios of attacker and defender. I've seen many examples of the 'directed' unit in question refusing to make a suicidal attack. It will if there are other friendly units around to make follow-up attacks, but it won't if it's the only attacker around. Have you seen behavior to refute this?

No, I see the same thing. I haven't sat down and worked out any numbers for thresholds though. In WotR I have units with low attack values (those with the defender role) make their way to the front (usually a city) via MU. When they get there they'll occasionally attack (eg, when the defenders have been weakened), but generally they'll wait or fortify.

Or, what I was hoping for was a magic Catfish answer that has seen this and worked out some examples that demonstrate some factors involved in this decision.

What? Do you take me for some kind of nerd or something?

 

[yoshi]

I'll try and do all that stuff I said over the holiday.


MODIFYREPUTATION:

Also reputation (0-100) is not to be confused with the reputation from the Foreign Minister screen and Cheat menu (0-8), ie, the number of betrayals.

AFAIK, the 'betray' parameter only refers to the 'who' civ's general rep (0-8) and the optional 'modifier' parameter refers to the 'whom' civ's attitude towards the 'who' civ. What other values are there?


Will go over what each value means for those who don't know.

In the game, reputation is categorized like so (Foreign Minister screen - F3):

7+= Atrocious (?)

7 = Atrocious
6 = Despicable
5 = Poor
4 = Dishonorable
3 = Questionable
2 = Honorable
1 = Excellent
0 = Spotless

(Note that this normally remains at the same level as the civ's last committed atrocity throughout the game unless your civ builds the Eiffel Tower wonder, which causes the number to drop back down to zero over time.)


AI attitude (if reputation is Spotless):

Peace

0 - 10 = Enthusiastic
11 - 25 = Cordial
26 - 38 = Receptive
39 - 61 = Neutral
62 - 74 = Uncooperative
75 - 89 = Icy
90 - 100 = Hostile

War

0 - 13 = Receptive
14 - 36 = Neutral
37 - 49 = Uncooperative
50 - 64 = Icy
65 - 74 = Hostile
75 - 100 = Enraged

(You can change this value using the Cheat Menu's 'Edit King' - Ctrl+Shift+K - feature.)

Here's an example of the ModifyReputation action in an event that changes reputation (i.e. the number of atrocities other civs believe the 'who' civ' has committed):

@IF
Turn
turn=2
@THEN
Text
^Roman reputation goes from Spotless to Poor.
Endtext
ModifyReputation
who=Romans
betray=5
@ENDIF

If you check in the F3 screen for the 'who' civ after this event has triggered, you'll see that the advisor's message at the top says something like, "Sire, our power is Weak and our reputation is Poor."


Where the possible values for the modifier parameter are concerned, the min-max values should be 0-100, not -100-100.

This is what the description in Macro.txt says:

"Changes the way other empires feel toward a specified civilization. Who is the civilization whose reputation is to be changed. Of the other parameters, you must have either Betray or the combination of Whom and Modifier. Betray sets the number of times that all other empires believe this civilization has betrayed allies. The higher this number is, the lower their opinion of the civilization will be. Use Whom if you only want to change the opinion of a specific other empire toward the Who civilization. Modifier is the amount by which you want to increase or decrease whom's disgust with who."

[Edited: modifier changes reputation, not attitude.]

This is what it should say:

"Changes the way other empires feel toward a specified civilization.

Who is the civilization whose reputation is to be changed.

Of the other parameters, you must have either 'Betray' or the combination of 'Whom' and 'Modifier.'

'Betray' sets the number of times that *all* other civs believe the 'who' civ has betrayed allies (i.e. any civs previously in Alliance with this civ). The higher this number is (i.e. number of atrocities 'who' civ has committed, to a maximum of 8--'who' civ's reputation is atrocious), the lower the opinion of all other civs will be toward the 'who' civ.

Use 'Whom' if you only want to change the 'who' civ's reputation as viewed ONLY by the 'whom' civ.

'Modifier' determines the change to the 'who' civ's present reputation as viewed by the 'whom' civ."


While testing the AI in the AWAW 1.2 scen, I found that no matter what I did, the Italians and French would always make war shortly after game start (no thanks to MGE's overly-aggressive AI). I tried setting both civ's reps and attitudes to zero and attitudes just switched back to their previous levels (mad as bloody hell). Alliances (Italy to Germany, France to Britain) may automatically drop attitute towards enemies of allies, regardless of other factors. But I've seen unexplained drops in attitude before, where nothing seems to appease the AI for more than a few turns (as with the testing done on DST's AI).

The ModifyReputation action should technically do exactly what the option in the game's Cheat menu does.

I thought that ModifyReputation set an absolute value, not modify the existing value

Macro.txt says the latter but if you can't enter negative values, then this makes no sense (because 0 would be your present rep--another civ's attitude towards you--and 100 would put you over the max, thus illogical. I think this was just improperly explained in the macro (see above).

MakeAggression isn't good enough as a cease fire can be arranged (and renewed) at the beginning of each player turn.

You could use Negotiation and MakeAggression in the same turn--repeat for each turn of war or use a flag or a delay. That way the two civ make war and don't talk until after x turns have passed. (AFAIK, MakeAggression has the AI consideritself at war with another player, regardless of whether the two talk or not--the AI will just sneak attack nect turn if it doesn't declare war.) You shouldn't even have to use ModifyReputation.

Forcing peace, on the other hand, is another story. (Being able to modify attitude would be essential for this--otherwise, it's all up to the AI...*shudder*.)




MOVEUNIT:

BTW, since MoveUnit gets a mention in this thread, I'll add that I've solved the worst of the AI's wandering habits in WotR. Conditions are tough for the AI on this map because of its large size, low city density and high number of impassable terrain barriers. You'd often find the AI wandering all over the place: in the far corners of Harad and in the northern wastelands. With judicious use of MoveUnit events and landmass numbering (the map is effectively divided into a number of theatres of war) it's no longer a major issue; you'll get the occasional miscreant, but they're usually brought back into line. There'll always be some arseclown trapped behind a terrain barrier, but that's nowhere near as bad as it was, either; they generally extricate themselves - eventually.

I thought low city density was a good thing: AI-controlled units apparently sometimes reset themselves when passing though a city. It also means that units being moved (via event) are less likely to have their orders cleared by bumping into a rival's cities).

There's the 'capturing' thing where you use Moveunit to redirect wandering units from areas of the map you don't want them to be in.

In the past, the most successful trigger it tech. Since you can take away tech in ToT, you can then just keep giving and taking the same tech each turn to trigger the MoveUnit every turn (AI seems to ignore when you use the simple turn trigger).

Landmass numbering: You divided MiddleEarth into continents (water dividing land)? Does this have anoticable effect on the AI (aside from obviously preventing movement of land units over water)?

AI-controlled units will wander less with additional landmasses (not referring to divisions)?

[I'm very interested in figuring out all the aspects of the MoveUnit action so please post everything you have.]

I have the impression that each unit makes a decision whether to attack as directed by a MU command. That decision appears to be depend on at least the combat strength ratios of attacker and defender. I've seen many examples of the 'directed' unit in question refusing to make a suicidal attack.

Since a unit's orders are cleared when adjacent (or in visial range) of a rival unit, it means you will never be able to force the AI to do anything it doesn't want to do. But as Wobbegong said, it depends on the unit's AI role; personally, I've never seen a unit with the Attack role not attack an adjacent enemy unit, regardless of the odds.

 

[Catfish]

AFAIK, the 'betray' parameter only refers to the 'who' civ's general rep (0-8) and the optional 'modifier' parameter refers to the 'whom' civ's attitude towards the 'who' civ. What other values are there?

Ahem, did you read all of my post? I refer to three types of values: reputation (0-8), ie, the number of betrayals, reputation (0-100) and attitude (0-100). I even posted their offsets to make the distinction.

Here's an example of the ModifyReputation action in an event that changes attitude (i.e. how a specific civ feels toward the 'who' civ):

No, the 'modifier' parameter for ModifyReputation does not modify attitude (0-100), it modifies reputation (0-100). This is the value that Mick Uhl describes in the link. If you look at the table of offsets I posted you will see that they are two separate things.

Unfortunately, the attitude modifier seems to be broken (I tried triggering it using different triggers and it had no effect on the AI).

The event correctly adjusts the reputation (0-100) value upwards, but how this value actually manifests itself in the game is currently unknown to me.

The ModifyReputation action should technically do exactly what the option in the game's Cheat menu does.

Well, it would be nice if there was a ModifyAttitude event and negative values were legal.

You could use Negotiation and MakeAggression in the same turn--repeat for each turn of war or use a flag or a delay.

How? Negotiation is an unusual trigger in that it also contains a type of action, ie, if the trigger's parameters are found to be true then contact between the specified tribes is automatically barred. If I use the Negotiation trigger it will block negotiations for the entire game regardless of what I use for the second trigger.

(AFAIK, MakeAggression has the AI consideritself at war with another player, regardless of whether the two talk or not--the AI will just sneak attack nect turn if it doesn't declare war.)

Yes, but that's next turn. In my scenario the human player can avoid considerable damage by accepting a cease fire for even part of a turn. What's more cease fires can be renewed in the following turns, generally at the behest of the AI and despite the activation of MakeAggression. In my example, the AI won't make a sneak attack because it's not powerful enough relative to the human player. This means that the human player can keep negotiating until they're in a suitable position to make a decisive attack themselves.

I thought low city density was a good thing: AI-controlled units apparently sometimes reset themselves when passing though a city.

I mean with respect to typical AI behaviour, not that imposed by MoveUnit.

In the past, the most successful trigger it tech. Since you can take away tech in ToT, you can then just keep giving and taking the same tech each turn to trigger the MoveUnit every turn (AI seems to ignore when you use the simple turn trigger).

In WotR, where there is a lot of fighting, I use the UnitKilled trigger with AnyUnit and Anybody wildcards for almost all of the MoveUnit events. That way units get bombarded with the event multiple times per turn (and during all players' turns). In versions MGE or less, you have the Anybody bug, so that wildcard is fairly limited for those scenarios.

Landmass numbering: You divided MiddleEarth into continents (water dividing land)?

No. This was done by hex editing. It took me a couple of hours of mind-numbing work, but at that point I was close to canning the scenario.

Does this have anoticable effect on the AI (aside from obviously preventing movement of land units over water)?

Yes. Landmass index has a significant effect on determining the AI's goals. Of course there are no artificial water barriers (see previous). They no longer idiotically target inaccessible tiles over the other side of the map. There are now plenty of areas that don't even require MoveUnit to generate sensible behaviour.

AI-controlled units will wander less with additional landmasses (not referring to divisions)?

Some wander from time to time, but with the combination of repeating MoveUnit events and landmass numbering, as a whole they are more focused. Wanderers are almost always recaptured by MoveUnit events - but there are always a few exceptions.

 

[Boco]

No. This was done by hex editing. It took me a couple of hours of mind-numbing work, but at that point I was close to canning the scenario.
Yes. Landmass index has a significant effect on determining the AI's goals. Of course there are no artificial water barriers (see previous). They no longer idiotically target inaccessible tiles over the other side of the map. There are now plenty of areas that don't even require MoveUnit to generate sensible behaviour.
Some wander from time to time, but with the combination of repeating MoveUnit events and landmass numbering, as a whole they are more focused. Wanderers are almost always recaptured by MoveUnit events - but there are always a few exceptions.

Curse you...the rewards you describe sound well worth the effort. EA2 could benefit from that a bit, too. Did you use any aids like 'marking' squares in ToT before hexing?

 

[Boco]

What does the AI do on its with air units in your doctored map? What does it do with MoveUnit commands?

 

[Boco]

With judicious use of MoveUnit events and landmass numbering (the map is effectively divided into a number of theatres of war)

How many continents do you have on the main map? I don't know how many the AI can handle.

 

[yoshi]

Ahem, did you read all of my post? I refer to three types of values: reputation (0-8), ie, the number of betrayals, reputation (0-100) and attitude (0-100). I even posted their offsets to make the distinction.

I read the post...sorta ...but I understood from the description in Macro.txt that the modify parameter (the 0-100 reputation value) affects two civs; who and whom:

Use Whom if you only want to change the opinion of a specific other empire toward the Who civilization. Modifier is the amount by which you want to increase or decrease whom's disgust with who.

If it modifies reputation, then shouldn't the min/max values for the modify parameter also be 0-8, not 0-100 (i.e. or -100-100 as it mistakenly says in the macro file)?

If I use the Negotiation trigger it will block negotiations for the entire game regardless of what I use for the second trigger.

File swap (one file with negotiation, one without).

They no longer idiotically target inaccessible tiles over the other side of the map. There are now plenty of areas that don't even require MoveUnit to generate sensible behaviour.

So you're changing values in the sav file in order to make the AI think it's on a seperate landmass? Where? Unless it's inhumanly tedious, I must know how to do this.

Using UnitKilled is good. As you said, the beauty of it is that you can trigger the same event mutiple times during a single turn. I should add an example of that. Would you mind posting one?

 

[Boco]

Do you observe AI units attacking adjacent units that lie in a different land mass?

[Edit]
@Yoshi: it depends on your definition of inhumanly tedious. There's a hexing document floating around indicating addresses where map information (among other things) is stored. If you don't already have it, I'll post a link when I have more time. Hexing land masses square by square is definitely a multiple-hour exercise of very repetitive finger motions. I'm hoping Wobbe has a trick or two.
[/Edit]

 

[Catfish]

What's this? Some kind of questionnaire? I thought I was the one asking the question here.

Curse you...

Well, nobody's forcing you to put a hex on your map.

Did you use any aids like 'marking' squares in ToT before hexing?

Nope, I just set the column width to 6 bytes per row and the jump length to 6 x map width. You can easily see the type of terrain and any improvements in the first 2 bytes of each tile. Reload the saved game in Civ2 if you get lost and look at the landmass numbers.

What does the AI do on its with air units in your doctored map?

I'd say the air units (read Winged Nazgul and Spiders of Mirkwood) are pretty much unaffected by the landmass changes, since air units will target areas on other landmasses anyway.

Air units have never really been a problem in this scenario. The Spiders attack units in and around Mirkwood. The Winged Nazgul, which spawn in Barad-dur, attack Osgiliath, Minas Tirith and Cair Andros. Generally they base themselves (and rehome ) in Osgiliath once it is taken and attack from there. They occasionally attack cities in Gondor's rear if no archers (x4 vs air) are present. If Gondor loses those frontline cities, the Nazgul push further west. Note that the whole Gorgoroth to Gondor area is a single landmass, with the exception of the Anduin and mountain ranges (impassable).

I've had games where some of the Winged Nazgul have targeted Lorien from Mordor (different landmasses) and then based themselves in Dol Guldur. Sometimes they have attacked (and killed) some of my Fellowship members en route, both on the ground and in longboats. This adds variation so I don't mind – let the Elves deal with them, I say. It also adds a threat level to this region. Khamul the Easterling and one lesser Nazgul were based in Dol Guldur anyway.

BTW, if you've got any tips on how to prevent flying units from rehoming, let me know.

What does it do with MoveUnit commands?

Are you referring specifically to air units?

How many continents do you have on the main map? I don't know how many the AI can handle.

Twenty two. Certainly I wouldn't exceed the allocated 64 per map.

If it modifies reputation, then shouldn't the min/max values for the modify parameter also be 0-8, not 0-100 (i.e. or -100-100 as it mistakenly says in the macro file)?

C'mon yoshi, you're getting lazy. There are two types of reputation: 0-8, the value from the Foreign Minister screen (ie, the number of betrayals) and 0-100, a value that each tribe has for every other tribe. ModifyReputation's 'betray' parameter modifies the first, its 'modifier' parameter modifies the second. The second reputation value (0-100) is not the same as attitude (0-100). I don't know how it works; that's what my initial post was about.

File swap (one file with negotiation, one without).

No can do. That will clear all of my flags and delays. Apart from that, like Kobayashi, I'm allergic to file-swapping. I'll try the 'PowerGraph' method I mentioned previously. This will be done by creating additional cities on one of the secondary maps. These will have relatively high growth rates.

So you're changing values in the sav file in order to make the AI think it's on a seperate landmass? Where? Unless it's inhumanly tedious, I must know how to do this.

That's right. And yes, it was inhumanly tedious – my map is exceptionally large. To edit the .sav file, you'll need a hex editor and the offsets (locations of bytes in the file). It's not rocket science, especially if someone tells you where to find the offsets.

Using UnitKilled is good. As you said, the beauty of it is that you can trigger the same event mutiple times during a single turn. I should add an example of that. Would you mind posting one?

Here's one straight out of WotR:

@IF
Checkflag
who=Mordor
flag=20
state=off
@AND
UnitKilled
unit=AnyUnit
attacker=Anybody
defender=Anybody
@THEN
MoveUnit
unit=AnyUnit
owner=Harad
maprect
129,213,163,213,163,237,129,237
moveto
152,202
numbertomove=All
@ENDIF

Those units with the 'defender' role are not affected wholesale by the 'numbertomove' parameter being set to 'All'. Common sense prevails here and most ignore it.

Do you observe AI units attacking adjacent units that lie in a different land mass?

For ground units: yes, if they stumble on them. They will even wander into adjacent landmasses from time to time. The difference is that they don't explicitly target tiles (with GoTo orders) on different landmasses. As stated above, air units will target and attack units on other landmasses.

There's a hexing document floating around indicating addresses where map information (among other things) is stored. If you don't already have it, I'll post a link when I have more time. Hexing land masses square by square is definitely a multiple-hour exercise of very repetitive finger motions. I'm hoping Wobbe has a trick or two.

For editing the map bytes, he could try using that spreadsheet I posted. It's what I use. It's definitely a chimp-style keyboard-bashing exercise. Earphones or speakers are a must.

 

[Mercator]

That's right. And yes, it was inhumanly tedious – my map is exceptionally large. To edit the .sav file, you'll need a hex editor and the offsets (locations of bytes in the file). It's not rocket science, especially if someone tells you where to find the offsets.

Why didn't you use MapCopy?

You can let it copy the body counter from another map. You could chop up the continents in that map using Ocean boundaries. You can then copy its counters to the savegame. That will leave you with the land tiles used for Ocean boundaries being badly numbered. To fix that properly, if the AI really minds it, you'd have to hex out those tiles back to land counts. But that should save a hell of a lot of work compared to renumbering all the land.

 

[Catfish]

Why didn't you use MapCopy?

Actually I started out doing it that way, but gave it away part way through the map-editing stage. Chopping up the 'continents' in the Civ2 map editor proved to be a painstaking task in itself. Certainly this was the case for my map where a large number of the 'continents' are sprawling and irregular. As you said, after running MapCopy you'd still need to manually repair the borders with a hex editor. I decided it would be simpler to just hex-edit the file. However, if you have a simple 'continent' structure, then it's probably the faster method. I completely forgot about it.

 

[yoshi]

I'm hoping Wobbe has a trick or two.

I'm hoping he does too. (By all means, post the link.)

BTW, if you've got any tips on how to prevent flying units from rehoming, let me know.

I though only land units rehomed--damn. Well, AFAIK there are certain slots that the AI won't rehome. (I can only assume the AI will treat air units like it does land units for this.)

There are two types of reputation: 0-8, the value from the Foreign Minister screen (ie, the number of betrayals) and 0-100, a value that each tribe has for every other tribe. ModifyReputation's 'betray' parameter modifies the first, its 'modifier' parameter modifies the second. The second reputation value (0-100) is not the same as attitude (0-100). I don't know how it works; that's what my initial post was about.

Okay, let me get something cleared up here first:

# of betrayals=reputation (0-8)
attitude per civ ranges 0-100
modifier parameter=also reputation (0-100) ???

This is illogical: why have two repuation parameters (betrayal and modifier) when they both technically do the same thing. The only possible answer is that modifier changes a civ's rep as seen ONLY by the whom civ. But if that were the case what's the point of having a 100 limit when the worst rep a civ can have is 8?

The only thing that has a limit like that is attitude. So none of this makes any sense to me, except that it seems to change some value(s) in the sav file.

That said, if the optional modifier parameter does anything then it should have done something in the following event and yet there seemed to be no effect on the AI's behaviour or any other aspect of the game:

@IF
Turn
turn=2
@THEN
ModifyReputation
who=Romans
whom=Persians
modifier=89
@ENDIF

Neither the who nor whom civ seemed to be affected in any way--the turn played out exactly the same each time with no change, regardless of what value was set.

I can only assume that the optional 'modifier' parameter is broken (i.e. was initially meant to do something but MP opted not to support it and didn't bother removing it).

That will clear all of my flags and delays.

Oh right, I forgot Delevent clears everything.

To edit the .sav file, you'll need a hex editor and the offsets (locations of bytes in the file). It's not rocket science, especially if someone tells you where to find the offsets.

Would that someone be you? (Couldn't find a cute flashing eyelashes smilie.)

 

[Catfish]

I'm hoping he does too.

Mercator just provided an alternative method. If your landmass divisions aren't too complex, I'd recommend that one.

Well, AFAIK there are certain slots that the AI won't rehome.

Do you remember what they are?

This is illogical: why have two repuation parameters (betrayal and modifier) when they both technically do the same thing.

Well, we don't know what exactly the 0-100 values do so you can't make that assumption. Even Mick Uhl didn't know how they worked. What I do know is that they exist.

The only possible answer is that modifier changes a civ's rep as seen ONLY by the whom civ.

That's correct. Look at the offset table I posted. You'll see that there are 8 values for each tribe. It's those values that change when you use the 'modifier' parameter with ModifyReputation.

But if that were the case what's the point of having a 100 limit when the worst rep a civ can have is 8?

Because they use a scale of 0-100, not 0-8.

So none of this makes any sense to me, except that it seems to change some value(s) in the sav file. (...) I can only assume that the optional 'modifier' parameter is broken (i.e. was initially meant to do something but MP opted not to support it and didn't bother removing it).

Well, it wouldn't be the first time something was broken in Civ2. It's either that or it's just too subtle for us to detect. I'll throw in Mick Uhl's quote from the link I posted earlier:

Reputation or betrayal are not the only factors that influence AI behavior. Many, many other variables are involved. As a result, you often don't see an immediate and obvious response to a betrayal or change in reputation. I wish I could rattle off the algorithms that decide these things but, unfortunately, I've been informed they are multitudinous and live in many places.

Would that someone be you? (Couldn't find a cute flashing eyelashes smilie.)

I can show you how to do it, but unless you're a perfectionist and/or crazy, it should really be a last resort method, ie, if you're having severe problems with the AI. You'll need a hex editor for starters. If you're unfamiliar with hex-editing then I'd read this. There's an Excel spreadsheet I posted which tells you how to find the exact location of map tile offsets in the ToT .sav file. If you've got any questions, fire away.

 

[yoshi]

Mercator just provided an alternative method. If your landmass divisions aren't too complex, I'd recommend that one.

Depends on what do you mean when you say, 'complex.'


As for the modifier, I don't know how else to test for it.

Thanks for the hex-editing info; I suppose I should get of my butt and learn to do this stuff.

 

[Catfish]

Well, AFAIK there are certain slots that the AI won't rehome.

Do you remember what they are?

BTW, do you mean unit slots, or domain or AI role slots?

Depends on what do you mean when you say, 'complex.'

I guess it does; a few big, uncomplicated blocks at one extreme and numerous long, sprawling dendritic types at the other. Also there's no reason why you couldn't use both methods, eg, use the MapCopy method for any large, blocky 'continents' and finish the rest manually with a hex editor. Both methods involve some degree of hex editing. As to the crossover point where one method is faster than the other, I couldn't tell you.

As for the modifier, I don't know how else to test for it.

Whatever it is involves more time than I'm prepared to invest – which is why I asked in the first place.

 

[yoshi]

BTW, do you mean unit slots, or domain or AI role slots?

(Sorry, I forgot to reply to your previous question.)

When I say 'slots' I'm referring to the unit's line under @UNITS in Rules.txt.

I posted a thread on this a while ago (I'm too lazy to look through the Civ2 forums); you can check by placing units in an empty city (or alternatively, one with 3 garrisoned units) and seeing which units rehome. For instance, I think the Marine unit slot was one that didn't re-home during this test whereas the Armor unit did (at first, I thought it was the attack role that was the determining factor but it doesn't look like it).

I didn't do much work on this so I can't say anything for sure. (I got pissed off that the AI was rehoming and causing other units with that home city to disband when the shields were completely consumed--as if the already bloody long list of AI stupidities wasn't bloody long enough.) Feel free to test this yourself; maybe I dreamt it and the AI makes no exceptions, so you'd better make sure.

I remember that ships don't rehome and I was under the impression that the same applied to air units--guess I was wrong.

I guess it does; a few big, uncomplicated blocks at one extreme and numerous long, sprawling dendritic types at the other. Also there's no reason why you couldn't use both methods, eg, use the MapCopy method for any large, blocky 'continents' and finish the rest manually with a hex editor. Both methods involve some degree of hex editing. As to the crossover point where one method is faster than the other, I couldn't tell you.

Well, the only project I'm presently working--and I use that term lightly--on is that WW2-Europe scen I'm always on about. I'm mostly worried about the wandering; I'm using 'formation' units (e.g. Armoured Division, Armoured Battalion as opposed to just Armour or Panzer IV) so I can't afford to have units be lost in some distant part of the map when they are needed for an assault--as opposed to just sending large masses of units and hoping most of them will go where they're supposed to at the expense of some wanderers.

There are other things like the AI trying to get to N. Africa from Europe by way of the Caucuses (i.e. the shortest land route), but I haven't seen the AI do that in other scens on the topic--distance may play some role in targeting--so it's really a secondary concern.

As long as nothing needs to be done before the other aspects of the scen are complete (i.e. hex editing can be done at any time), I'll leave it for now.

Whatever it is involves more time than I'm prepared to invest – which is why I asked in the first place.

Sorry I couldn't be of any help on this but when I play with the AI, it's either a war scen or a modded epic scen where anything goes, so this has never come up (I tested it once but I just used the 'betray' parameter).

 

[Catfish]

When I say 'slots' I'm referring to the unit's line under @UNITS in Rules.txt.

OK, thanks. I'll test it, erm, when I'm next in the mood for that sort of thing.

Sorry I couldn't be of any help on this...

No need to be apologetic. It was an open question to anyone who might have some experience using that event. I guess it will have to remain a mystery... for the time being. Thanks for showing the interest.

 

[Boco]

When you've solved this one, guys. Post away! May the mood strike soon.

It's been a minor bane for me, too. Unit slots 62, 63, and 71 have rehomed too much for me. Ground domain; Role = 0 and 1. Also have seen occasional disbanding of a role 0 unit after rehoming, too. That forced me to use CityTaken/CreateUnit events in cities where the AI should not have disbanded the initial deployment.

 

[Boco]

Wobby (Yoshi, or anyone, excluding Mer-"what's a spreadsheet?"-cator ), do you use A.X.E., v2? I've tweaked your xls by adding a button to copy the offset, switch to AXE, and go to the offset. If you use it, I'll post the xls.