editing zoom in view

[thunderchero]

Been test with good results for larger map tweaks. Most of the time 2 -3 races have 27 - 30 systems, and one race 20 - 25 systems (mostly feds, but not always).

I was just wondering what would be the best trigger to test to increase AI systems held? I still see many empty systems AI could colonize.

I still need to look at the version/registry changes and see what changes need to be made.

[Spocks-cuddly-tribble]

I was just wondering what would be the best trigger to test to increase AI systems held? I still see many empty systems AI could colonize.

Depends on the issue at hand:

- too few colony ships due to fleet buildup agenda (e.g. vanilla war/jihad)
- colony ships not used
- other issue like range vs star density


In my tests this year - ECM(base) [T1, large map, no minors] all games ended with max 2-4 uncolonized systems.
Larger maps AI tests get unplayable due to turn lag...


Here is an insufficiently tested option to remove limits for active colonize (+bldbase child tasks for it), attack and invade tasks.
Also shared for train and repair tasks, but shouldn't matter since default limit for both is 16 active map tasks per AI.

Code: Select all

40DE00   push  ecx    // 0xD200 51 -> C3 retn

If you load my vanilla saved game from page one with AI glitch (caused by the code the patch skips) you see the difference:

check out this vanilla AI glitch (romulan buildbase - colonize child task):
game1.sav download/file.php?id=3204

Updated database: https://workupload.com/file/eSmWMMPpNXa

[thunderchero]

In my tests this year - ECM(base) [T1, large map, no minors] all games ended with max 2-4 uncolonized systems.
Larger maps AI tests get unplayable due to turn lag...

Here is an insufficiently tested option to remove limits for active colonize (+bldbase child tasks for it), attack and invade tasks.
Also shared for train and repair tasks, but shouldn't matter since default limit for both is 16 active map tasks per AI.

Code: Select all

40DE00   push  ecx    // 0xD200 51 -> C3 retn

tested with change above, it helped....

here is my latest test game feds have almost taken out cardassians (5 systems left) and feds have 47 systems, romulans have 40 systems, klingons have 32 systems

latest.jpg (530.78 KiB) Viewed 1953 times

[Spocks-cuddly-tribble]

Nice, but keep an eye open for unwanted side effects (not fully terraformed AI systems, fleet fragmentations vs serious enemies).

Keep in mind this patch and 'Dynamic Danger Check for AI Colonize, Tform & BldBase Tasks' viewtopic.php?p=50865#p50865 makes the AI colonizing more careless towards other empires' 'Sphere Of Influence' i.e. more AI vs AI might be easier for the player (albeit good for MUM concept)?

Sometimes some AIs do very good due to glitch behavior of enemy AIs (like unneeded moving the whole main fleet vs minors, exposing the own territory for invasions -> see patch).

You might want to remove this for MUM:

Code: Select all

0040EA8C           jz    short dont_invade (minors)

No invade minor tasks when agenda Offensive War (to focus fleets vs main enemies?) -> Not needed with my AI invasion fleet fix vs minors.

[thunderchero]

T3 (tech 8 of 20 techs) looked real good all terraformed 38, 44, 41, 38 systems. 6, 7, 6, 7 (26 of 30 minors under majors race control) none are membered.
ferengi and klingons at peace, romulan and carassians are at war

t3.jpg (468.01 KiB) Viewed 1916 times

looks very good to me, best I have seen AI play

and yes they still keep main fleet together unless they are moving. each race had avg fleet of 275 ships.

[Spocks-cuddly-tribble]

looks very good to me, best I have seen AI play

and yes they still keep main fleet together unless they are moving. each race had avg fleet of 275 ships.

I run an ECM large test with AI task patches and colonize (+ bldbase child tasks) is improved for larger maps, but invade/attack looks semi-broken.
--> we should skip code for colonize + child tasks only (nop 2 calls @ 4101B7 / 40FF77 -> remove old 1 byte patch)

Also many nearby idle TTs take too long to get assigned to invasions (disabling pointless repair didn't solve this).
-> we should force TTs to join main fleet staging

[thunderchero]

I run an ECM large test with AI task patches and colonize (+ bldbase child tasks) is improved for larger maps, but invade/attack looks semi-broken.
--> we should skip code for colonize + child tasks only (nop 2 calls @ 4101B7 / 40FF77 -> remove old 1 byte patch)

I restored code

Code: Select all

40DE00   push  ecx    // 0xD200 51 -> C3 retn 

tested with both nop 2 calls @ 4101B7 / 40FF77 and got no subjugation or membering of minors

but when I only nop 4101B7 I got good results. 41, 43, 33, 32

4101B7.jpg (866.06 KiB) Viewed 1875 times

I also started to notice almost every system they leave un-colonized is under 300 pop...

[Spocks-cuddly-tribble]

tested with both nop 2 calls @ 4101B7 / 40FF77 and got no subjugation or membering of minors

but when I only nop 4101B7 I got good results. 41, 43, 33, 32

This strange feature needs more analysis... (plus tests with vanilla-ish AI codes - MUM might change too many AI features)

EDIT: If the one byte patch works so well for MUM you might settle for it (albeit this code isn't used for diplomacy/memberships?)

I also started to notice almost every system they leave un-colonized is under 300 pop...

You have a keen eye for observation.

4101B7 is limit for colonize main code (the other for bldbase child tasks if all colonize targets are out of range).

AI colonize system classification checks pop boundaries of 150/300, bonuses like dilithium & colvar.wtf: viewtopic.php?p=34204#p34204

B 'System Colonization Value (%s)' (base layer)
C 'System Colonization Priority (%s)' -> B impacted by empty nearby systems

Here is Romulan map AI data views B & C from my vanilla glitch game:

x_b.png (198.7 KiB) Viewed 1857 times

x_c.png (194.54 KiB) Viewed 1857 times

[thunderchero]

today I went back and edited the vanilla trek.exe's to all use 4:3 maps with Dynamic Widescreen Zoom-In for Any Galaxy Grid patch.
I also added back the tutorials for widescreen.

EDIT: If the one byte patch works so well for MUM you might settle for it (albeit this code isn't used for diplomacy/memberships?)

just to be clear

Not Used
40DE00 push ecx // 0xD200 51 -> C3 retn

Used
nop call @ 4101B7

Not Used
nop call @ 40FF77

if I nop 2 calls @ 4101B7 / 40FF77 at both offsets I got no subjugation

if I nop call @ 4101B7 only subjugation was normal for MUM again (90% subjugation of minors)

[thunderchero]

Dynamic Widescreen Zoom-In for Any Galaxy Grid worked for MUM 1024

zoom.jpg (258.03 KiB) Viewed 1766 times

full3.jpg (202.99 KiB) Viewed 1766 times

[Spocks-cuddly-tribble]

looks very good to me, best I have seen AI play

and yes they still keep main fleet together unless they are moving. each race had avg fleet of 275 ships.

I run an ECM large test with AI task patches and colonize (+ bldbase child tasks) is improved for larger maps, but invade/attack looks semi-broken.

Good news, I found the missing counter parts for active tasks limits (the one byte patch):

Code: Select all

NAME: Invade, Attack & Colonize increased 'set Tasks' Limits

# invade

>> 0xDFFF 03
<< 0xDFFF 7F

# attack

>> 0xF7AA 02
<< 0xF7AA 0F

# Colonize

>> 0xF4AF 02
<< 0xF4AF 0F

+

Code: Select all

NAME: Remove Active AI Tasks Limits

>> 0xD200 51
<< 0xD200 C3

With both the ECM AI virtually owns larger maps (still issues but most time-lapse tests looked really great for BotF AI).

+ viewtopic.php?p=57537#p57537 & AI_Invade_Minor_Fix.patch (+allow invade minor at war for MUM?)


To ensure a certain AI level of sustainable development i.e.to prevent this: viewtopic.php?p=56647#p56647

Add also:



- AI tweak 6 adapting AI trade goods
- AI ship sorting by function deviation fix
- all patches protecting AI from random events, monsters & anomalies
- food pop up fro max growth fix + viewtopic.php?p=56457#p56457
- force AI to tform improve systems workaround (above remove task limits will override this if needed)
- retreat fix is nice for combat


You can/should use most but these:

aft torpedo launchers -> great feature but much work (shiplist.sst, same as shield level)
auto-upgrade main buildings fix -> do you want issue we talked about or ignored-tech levels? viewtopic.php?p=57537#p57537
colony starting pop/buildings fix -> will change all numbers to lower index
Fairer starting positions -> No (increases min empire distances)
Morale chasing ships away -> MP issue!!!
phantom dilithium bug fix -> version? BoP disable ship sorting issue -> OK for MUM & MP viewtopic.php?p=17544#p17544

Hope this helps.

[thunderchero]

played with some patches today

tech 1 start impossible, 30 minors, borg and random off,
large 40% density

at turn 196 other races had a tech level of
cards tech 18, feds tech 16 klingons tech 14, romulans tech 16

60.jpg (753.14 KiB) Viewed 1639 times

60a.jpg (79.11 KiB) Viewed 1639 times

that was good news but here is the bad news

turns 1 - 170 took 30 min
turns 171 - 180 took 20 min more
turns 181 - 196 took 1 hour and 44 min more

[Spocks-cuddly-tribble]

Also think about tactical combat lag.
150 vs 150 max tech ships in ECM (default map large) was fun, but still the lag started to get annoying.
So larger maps (i.e. massive AI fleets) will never be playable on a regular base....

[thunderchero]

Also think about tactical combat lag.
150 vs 150 max tech ships in ECM (default map large) was fun, but still the lag started to get annoying.
So larger maps (i.e. massive AI fleets) will never be playable on a regular base....

just a quick count of large fleets (all races) on that map and it was over 1600 ships.

[Spocks-cuddly-tribble]

Looks very peaceful (no invading fleets in other territories). Are you sure you develop the right mod (MUM)?
In ECM (base & larger maps with tweaks) AIs eliminate each other given the opportunity....


I just removed a whole bunch of broken AI turn lag code without any effect (but no improvement on the issue).
It worsens if a massive fleet AI runs out of reasonable objectives/targets (big lag but no useful actions).
So for pet-universe players we could use a time-out limit workaround (say 5 seconds shipUnits tests per AI = 20 seconds max + other AI calculations)?


You question to disable shared range via treaty reminded me to fix code to relocate stranded (AI) ships when betrayed by treaty partners (in ECM this happens sometimes). I fixed the closed border treaty code for this (another bug ignores contested sectors) and call it also when ending empire treaties to move stranded task forces.