Addressing the AI Precursor Bug

[z3n]

This is a pretty quick tutorial, just explaining a concept.
Some of you may have encountered the endless AI loop with 'reloading' after the AI suddenly switches into a from a fighting state with an enemy into a reloading state for no apparent reason.

Instead of having to be stuck with this bug you can do something else in the export_descr_unit.
It requires two things

1) the cannot_skirmish attribute
2) thrown after stat_pri_attr instead of prec

; COMMENTS              First Cohors Reformata (Post-Marian Legionary First Cohort) / Unit ID 193 / Start exp: 0 / Refr rate: 0,048 / Pool cap: 6
type                    roman infantry reformata first
dictionary              roman_infantry_reformata_first
category                infantry
class                   heavy
voice_type              General
accent                  Roman_Latin
banner faction          main_infantry
banner holy             crusade
soldier                 Marian_First_Cohort, 96, 0, 1.25, 0.3
officer                 officer_roman_marian
officer                 officer_roman_aquilifer
officer					officer_roman_legion_vexillarius
attributes              sea_faring, hide_forest, hardy, can_withdraw, command, cannot_skirmish
move_speed_mod          0.83
formation               0.89, 1.19, 2.7, 2.7, 4, square
stat_health             1, 5
stat_pri                20, 0, pilum, 60, 2, thrown, missile_mechanical, piercing, none, 0, 1
stat_pri_attr           thrown
stat_sec                10, 5, no, 0, 0, melee, melee_blade, piercing, sword, 0, 1
stat_sec_attr           no
stat_pri_armour         9, 6, 9, metal
stat_sec_armour         0, 0, flesh
stat_heat               3
stat_ground             -1, -3, -3, -5
stat_mental             7, disciplined, highly_trained
stat_charge_dist        14
stat_fire_delay         0
stat_food               60, 300
stat_cost               1, 2500, 625, 81, 591, 2500, 6, 625
stat_stl                9
armour_ug_levels        0
armour_ug_models        Marian_First_Cohort
ownership               f_rome, f_carthage, f_makedonia, f_epeiros, f_kh, f_seleukid, f_ptolemaioi, f_baktria, f_hayasdan, f_pontos, f_parthia, f_aedui, f_arverni, f_casse, f_lusotannan, f_sweboz, f_getai, f_saba, f_saka, f_sauromatae, f_numidia, f_pergamon, f_gandhara, f_lugia, f_bosporan, f_arevaci, f_boii, f_nabatu, slave, f_bellouakoi
era 0                   f_rome
era 1                   f_rome
era 2                   f_rome, f_carthage, f_makedonia, f_epeiros, f_kh, f_seleukid, f_ptolemaioi, f_baktria, f_hayasdan, f_pontos, f_parthia, f_aedui, f_arverni, f_casse, f_lusotannan, f_sweboz, f_getai, f_saba, f_saka, f_sauromatae, f_numidia, f_pergamon, f_gandhara, f_lugia, f_bosporan, f_arevaci, f_boii, f_nabatu
info_pic_dir            f_all
card_pic_dir            f_all
recruit_priority_offset 50

Note you will be able to give the fake precursor units as many javelins as you want and they will use their javelins in a relatively realistic and effective manner. *Although the play testers/game mechanics team which tested this were not using vanilla/other AI files. So I am not sure whether it will be the same unless you use the BAI located in the Skynet AI resource.

In any case, this was the solution that I came up with to help cut down on the behaviour we were having in EBII, hopefully it helps others as well with this issue. They're all very excited and happy with the results (especially getting to add more than one javelin/pilum, instead of the previous limit of just 1), so hopefully some of you will be as well.

[GRANTO]

This is a pretty quick tutorial, just explaining a concept.
Some of you may have encountered the endless AI loop with 'reloading' after the AI suddenly switches into a from a fighting state with an enemy into a reloading state for no apparent reason.

Instead of having to be stuck with this bug you can do something else in the export_descr_unit.
It requires two things

1) the cannot_skirmish attribute
2) thrown after stat_pri_attr instead of prec

; COMMENTS              First Cohors Reformata (Post-Marian Legionary First Cohort) / Unit ID 193 / Start exp: 0 / Refr rate: 0,048 / Pool cap: 6
type                    roman infantry reformata first
dictionary              roman_infantry_reformata_first
category                infantry
class                   heavy
voice_type              General
accent                  Roman_Latin
banner faction          main_infantry
banner holy             crusade
soldier                 Marian_First_Cohort, 96, 0, 1.25, 0.3
officer                 officer_roman_marian
officer                 officer_roman_aquilifer
officer                    officer_roman_legion_vexillarius
attributes              sea_faring, hide_forest, hardy, can_withdraw, command, cannot_skirmish
move_speed_mod          0.83
formation               0.89, 1.19, 2.7, 2.7, 4, square
stat_health             1, 5
stat_pri                20, 0, pilum, 60, 2, thrown, missile_mechanical, piercing, none, 0, 1
stat_pri_attr           thrown
stat_sec                10, 5, no, 0, 0, melee, melee_blade, piercing, sword, 0, 1
stat_sec_attr           no
stat_pri_armour         9, 6, 9, metal
stat_sec_armour         0, 0, flesh
stat_heat               3
stat_ground             -1, -3, -3, -5
stat_mental             7, disciplined, highly_trained
stat_charge_dist        14
stat_fire_delay         0
stat_food               60, 300
stat_cost               1, 2500, 625, 81, 591, 2500, 6, 625
stat_stl                9
armour_ug_levels        0
armour_ug_models        Marian_First_Cohort
ownership               f_rome, f_carthage, f_makedonia, f_epeiros, f_kh, f_seleukid, f_ptolemaioi, f_baktria, f_hayasdan, f_pontos, f_parthia, f_aedui, f_arverni, f_casse, f_lusotannan, f_sweboz, f_getai, f_saba, f_saka, f_sauromatae, f_numidia, f_pergamon, f_gandhara, f_lugia, f_bosporan, f_arevaci, f_boii, f_nabatu, slave, f_bellouakoi
era 0                   f_rome
era 1                   f_rome
era 2                   f_rome, f_carthage, f_makedonia, f_epeiros, f_kh, f_seleukid, f_ptolemaioi, f_baktria, f_hayasdan, f_pontos, f_parthia, f_aedui, f_arverni, f_casse, f_lusotannan, f_sweboz, f_getai, f_saba, f_saka, f_sauromatae, f_numidia, f_pergamon, f_gandhara, f_lugia, f_bosporan, f_arevaci, f_boii, f_nabatu
info_pic_dir            f_all
card_pic_dir            f_all
recruit_priority_offset 50

Note you will be able to give the fake precursor units as many javelins as you want and they will use their javelins in a relatively realistic and effective manner. *Although the play testers/game mechanics team which tested this were not using vanilla/other AI files. So I am not sure whether it will be the same unless you use the BAI located in the Skynet AI resource.

In any case, this was the solution that I came up with to help cut down on the behaviour we were having in EBII, hopefully it helps others as well with this issue. They're all very excited and happy with the results (especially getting to add more than one javelin/pilum, instead of the previous limit of just 1), so hopefully some of you will be as well.

Great work Z3n , that always bugged me in a few mods.

[bitterhowl]

Prec attribute confirmed works for cavalry (disables skirmish mode by itself without cannot_skirmish). Unit throws projectiles while running to enemy and charge without any additional animations or stooping.

In another hand it finally proves that prec is not functional for infantry.

[QuintusSertorius]

Prec attribute confirmed works for cavalry (disables skirmish mode by itself without cannot_skirmish). Unit throws projectiles while running to enemy and charge without any additional animations or stooping.

In another hand it finally proves that prec is not functional for infantry.

Sorry, in what respect is prec "not functional for infantry"? We're using it and it does exactly what you'd expect - units throw javelins then charge. They also behave much better than they do without it, insofar as actually performing charges.

[z3n]

Sorry, in what respect is prec "not functional for infantry"? We're using it and it does exactly what you'd expect - units throw javelins then charge. They also behave much better than they do without it, insofar as actually performing charges.

I believe what he refers to is the endless "reloading" loop which occurs when units have the prec attribute.

[bitterhowl]

Does prec by itself disable skirmish mode in infantry units?

This thread OP is about removing prec.

[QuintusSertorius]

Does prec by itself disable skirmish mode in infantry units?

No, but frankly that's a trivial matter given all you need to do is give them cannot_skirmish.

This thread OP is about removing prec.

The only reason to remove it was to address the freezing bug - which in fact is mostly resolved by giving the unit 2 missiles. Removing prec causes a much worse problem than freezing, namely that charges don't work at all without it.

[bitterhowl]

type roman infantry principes camillan
dictionary roman_infantry_principes_camillan
category infantry
class spearmen
voice_type General
accent Roman_Latin
banner faction main_spear
banner holy crusade
soldier Principes_Camillan, 64, 0, 1.1, 0.25
officer officer_roman_camillan
officer officer_roman_principes_sb
attributes sea_faring, hide_forest, free_upkeep_unit, can_withdraw, cannot_skirmish
move_speed_mod 0.83
formation 0.9, 0.9, 0.9, 1.8, 8, square
stat_health 1, 5
stat_pri 20, 0, pilum, 60, 2, thrown, missile_mechanical, piercing, none, 0, 1
stat_pri_attr thrown
stat_sec 8, 4, no, 0, 0, melee, melee_simple, piercing, spear, 0, 1
stat_sec_attr light_spear, spear_bonus_4
stat_pri_armour 6, 6, 6, metal
stat_sec_armour 0, 0, flesh
stat_heat 3
stat_ground -1, -3, -3, -5
stat_mental 6, disciplined, highly_trained
stat_charge_dist 14
stat_fire_delay 0
stat_food 60, 300
stat_cost 1, 1650, 309, 82, 165, 1650, 6, 309
stat_stl 9
armour_ug_levels 0
armour_ug_models Principes_Camillan
ownership f_rome, f_carthage, f_makedonia, f_epeiros, f_kh, f_seleukid, f_ptolemaioi, f_baktria, f_hayasdan, f_pontos, f_parthia, f_aedui, f_arverni, f_casse, f_lusotannan, f_sweboz, f_getai, f_saba, f_saka, f_sauromatae, f_numidia, f_pergamon, f_gandhara, f_lugia, f_bosporan, f_arevaci, f_boii, f_nabatu, slave, f_belgai
era 0 f_rome
;era 1 f_rome
;era 2 f_rome
info_pic_dir f_all
card_pic_dir f_all
recruit_priority_offset 60

I made many experiments in my own Westeros mod and find that charge is possible when after throwing unit situated in charge distance area. You have 60 and 14 here, also you got perfect throwing animation, when unit runs towards the enemy, so become closer after throwing. I use your animation and have 45 and 35 distance - and unit charges just fine. In my mod - with prec reloading bug persisted until I did as OP says.

"Quod hominem tot sententiae", everyone has it's own mod, I'm just sharing my experience. With cavalry prec I saw how should it be ideally with infantry, but it doesn't (in my mod).

[QuintusSertorius]

type roman infantry principes camillan
dictionary roman_infantry_principes_camillan
category infantry
class spearmen
voice_type General
accent Roman_Latin
banner faction main_spear
banner holy crusade
soldier Principes_Camillan, 64, 0, 1.1, 0.25
officer officer_roman_camillan
officer officer_roman_principes_sb
attributes sea_faring, hide_forest, free_upkeep_unit, can_withdraw, cannot_skirmish
move_speed_mod 0.83
formation 0.9, 0.9, 0.9, 1.8, 8, square
stat_health 1, 5
stat_pri 20, 0, pilum, 60, 2, thrown, missile_mechanical, piercing, none, 0, 1
stat_pri_attr thrown
stat_sec 8, 4, no, 0, 0, melee, melee_simple, piercing, spear, 0, 1
stat_sec_attr light_spear, spear_bonus_4
stat_pri_armour 6, 6, 6, metal
stat_sec_armour 0, 0, flesh
stat_heat 3
stat_ground -1, -3, -3, -5
stat_mental 6, disciplined, highly_trained
stat_charge_dist 14
stat_fire_delay 0
stat_food 60, 300
stat_cost 1, 1650, 309, 82, 165, 1650, 6, 309
stat_stl 9
armour_ug_levels 0
armour_ug_models Principes_Camillan
ownership f_rome, f_carthage, f_makedonia, f_epeiros, f_kh, f_seleukid, f_ptolemaioi, f_baktria, f_hayasdan, f_pontos, f_parthia, f_aedui, f_arverni, f_casse, f_lusotannan, f_sweboz, f_getai, f_saba, f_saka, f_sauromatae, f_numidia, f_pergamon, f_gandhara, f_lugia, f_bosporan, f_arevaci, f_boii, f_nabatu, slave, f_belgai
era 0 f_rome
;era 1 f_rome
;era 2 f_rome
info_pic_dir f_all
card_pic_dir f_all
recruit_priority_offset 60

I made many experiments in my own Westeros mod and find that charge is possible when after throwing unit situated in charge distance area. You have 60 and 14 here, also you got perfect throwing animation, when unit runs towards the enemy, so become closer after throwing. I use your animation and have 45 and 35 distance - and unit charges just fine. In my mod - with prec reloading bug persisted until I did as OP says.

"Quod hominem tot sententiae", everyone has it's own mod, I'm just sharing my experience. With cavalry prec I saw how should it be ideally with infantry, but it doesn't (in my mod).

What we found with prec removed was that units would run up, stop a short distance away, switch weapons from their javelins and have no momentum whatsoever left in making the charge.

Does a longer charge distance remove the issue with weapon-switching, because it happens further out?

[bitterhowl]

Got fresh news about prec/thrown (fresh for me, didn't noticed that before) - prec unit has "fire_at_will" button disabled by default, when prec changed to thrown "fire_at_will" enables back by default.

As for charge_distance, I set it to high numbers according to an old Rome tutorial, and noticed strange thing with missile infantry - they switched to secondary wearpon and stopped firing while enemy was on charge distance area. I spent some time before find out what happened. Also knight charge was broken with charge_distance 145 - they swithed to secondary just before enemy instead continuing lance strike. Charge_distance lower 70 healed that for knights in my case.

[QuintusSertorius]

Interesting; we used short charge distance, because that seems to result in better cohesion.

[Razor]

Ok, so what's the update on this issue? This back and forth between you two is all a bit confusing to me.

[bitterhowl]

Studying started here

https://www.twcenter.net/forums/show...0#post15786780

[Razor]

Ah, so no definite conclusions as of yet. Thanks for the link.

I see that this is being tested with the EBII mod. I'm curious to see if we can get similar results on vanilla M2TW without the (IMO annoying) throwing animation of EBII which causes units to move forward.

[bitterhowl]

Without EB2 animation we should make throw distance even less (about 35) and charge distance about 25 - now we just wait for enemies that comes closer during their attack. Don't forget about cannot_skirmish attribute. I did prec effect like this for my mod before adding new animation.

[QuintusSertorius]

The testing we're doing in EBII is on charge distances, not the presence/absence of prec.

[bitterhowl]

Well, seems I got some progress here. My Ironbors fire their axes then run to enemy, no manually scripted behavior.

https://youtu.be/fMVX-70_hwo

Just copied Rome I EDU settings for pilum unit, + set horde instead square (in square sometimes only first line fought in melee, and back tryed to fire volleys and only then fight).

Spoiler Alert, click show to read: 

type Dismounted Huscarls
dictionary Dismounted_Huscarls ; Warriors
category infantry
class heavy
voice_type Heavy
banner faction main_infantry
banner holy crusade
soldier Dismounted_Huscarls, 48, 0, 1
attributes sea_faring, hide_forest, very_hardy, can_withdraw, frighten_foot, frighten_mounted, ;cannot_skirmish
move_speed_mod 1.1
formation 1.2, 1.2, 2.4, 2.4, 4, horde
stat_health 1, 3
stat_pri 11, 1, iron_axe, 35, 2, thrown, missile_mechanical, slashing, axe, 35, 1
stat_pri_attr prec, thrown, ap, launching
stat_sec 5, 3, no, 0, 0, melee, melee_blade, slashing, axe, 25, 1
stat_sec_attr ap
stat_pri_armour 4, 10, 4, metal
stat_sec_armour 0, 0, flesh
stat_heat 4
stat_ground 2, -4, 0, 0
stat_mental 7, impetuous, untrained
stat_charge_dist 30
stat_fire_delay 5000
stat_food 60, 300
stat_cost 1, 850, 175, 290, 70, 850, 4, 170
armour_ug_levels 2, 3, 4
armour_ug_models Dismounted_Huscarls, Dismounted_Huscarls_ug1, Dismounted_Druchima_ug1, ;Norse_Swordsmen_ug2
ownership hre, slave
era 0 hre
recruit_priority_offset 8

When I set only prec without thrown first throw was correct, then unit ran to enemy but then I mentioned that back lines stood still with constant "reloading" mark, so unit was uneffective in close combat. I already wanted to leave prec mechanics but then made last try to copy Rome I EDU. Made about 10 tests, everything works as on video.

[Lusitanio]

That seems great Bitter, +1 rep from me. I will test that myself today!

[Lusitanio]

Indeed, I have just confirmed what bitterhowl said, setting horde instead of square completely fixes the issue with infantry charges.
I tested it several times with 6 units vs 6 units in which half had a horde formation and the other half had square formation (3 squares and 3 hordes vs 3 squares and 3 hordes).
Time and time again the units with the square formation charged incorrectly or badly, sometimes they didn't charge at all. The units with horde always charged correctly at the enemy units. The only issue I could report is that when running at the enemy the units would get behind the leaders and would divide in corridors which would leave some spaces open but that was really temporary. They also seem to penetrate more on the enemy unit and try to surround them more.

[bitterhowl]

Well, have to say that once again I made my common mistake - posting before fully testing. So, in a battle 1 on 1 everything works as on video. When we have more units reload bug appears again, it has no system, sometimes units fight as they should after throwing, sometimes they just constantly reload even being surrounded by player's units. I changed this line from 40

<heavy-infantry>
<max-engage-dist>70</max-engage-dist>
</heavy-infantry>

And units act more adequate. When I set EDU as in OP everything work just fine without any reloads, but no prec charge on enemy, as you understand. Need more ingame experience with that to close this theme.