Matrix Games Forums

Forums  Register  Login  Photo Gallery  Member List  Search  Calendars  FAQ 

My Profile  Inbox  Address Book  My Subscription  My Forums  Log Out

RE: New Pacific War Bug Fixes / Enhancements

 
View related threads: (in this forum | in all forums)

Logged in as: Guest
Users viewing this topic: none
  Printable Version
All Forums >> [Classic (Free) Games] >> Pacific War: The Matrix Edition >> RE: New Pacific War Bug Fixes / Enhancements Page: <<   < prev  1 [2]
Login
Message << Older Topic   Newer Topic >>
RE: New Pacific War Bug Fixes / Enhancements - 4/23/2016 2:35:05 PM   
Hattori Hanzo


Posts: 604
Joined: 3/21/2011
From: Okinawa
Status: offline
a GREAT Labour of Love !!!!!

(in reply to Rich Dionne)
Post #: 31
RE: New Pacific War Bug Fixes / Enhancements - 7/6/2016 5:45:14 PM   
zeke99


Posts: 385
Joined: 11/26/2005
Status: offline
Bug: Flak of ships in harbour

Dear Rich,

first I'd like to thank you for your work. Just testing the latest version in the game with Istfemer.

One thing that needs to be fixed is the decrease in Flak value for ships in port. This is an old issue that could be circumvented by keeping ships in TFs. However, since now they degrade it needs to be fixed.

IMHO the Flak value should not go down in ports at all. What should increase instead is the hit probability of guns, bombs and torpedos due to stationary targets.

Just my 5cents,
Zeke

(in reply to Rich Dionne)
Post #: 32
RE: New Pacific War Bug Fixes / Enhancements - 7/6/2016 8:00:39 PM   
Istfemer

 

Posts: 157
Joined: 7/19/2014
From: Kyiv, Ukraine
Status: offline
I can already sense the problems that will pop up if your suggestion becomes reality, Zeke.
Raid on Pearl Harbor will be a spectacularly costly failure for the IJNAF. Their bombers will all get shredded by Allied flak and Allies won't lose any ships.
With flak values not reduced, on 12/7/41 the ships in Pearl Harbor port will have a combined flak value of 5599.
(non-historical start, the flak value will be slightly lower in other scenarios)

< Message edited by Istfemer -- 7/6/2016 8:11:07 PM >

(in reply to zeke99)
Post #: 33
RE: New Pacific War Bug Fixes / Enhancements - 7/7/2016 3:46:57 AM   
Capt. Harlock


Posts: 5146
Joined: 9/15/2001
From: Los Angeles
Status: offline
quote:

Raid on Pearl Harbor will be a spectacularly costly failure for the IJNAF. Their bombers will all get shredded by Allied flak and Allies won't lose any ships.
With flak values not reduced, on 12/7/41 the ships in Pearl Harbor port will have a combined flak value of 5599.


Interesting -- that much even before the upgrades? And isn't there an averaging algorithm to compensate for there being many more ships in harbor than are allowed in a TF?

_____________________________

Civil war? What does that mean? Is there any foreign war? Isn't every war fought between men, between brothers?

--Victor Hugo

(in reply to Istfemer)
Post #: 34
RE: New Pacific War Bug Fixes / Enhancements - 7/7/2016 3:56:18 AM   
zeke99


Posts: 385
Joined: 11/26/2005
Status: offline
Hi Istfemer, I thought you will bring up this point

It can be dealt with algos like Capt.Harlock proposed but it brings up a very interesting point, the TF Flak value.
As I understand it, it is calculated by simply adding all Flak and use it against incoming aircraft. This is technically WRONG. Small calibres will only protect the ship they are installed on due to range and distance of the other ships. The bigger the TF the lower their impact on air defence. This could be taken care of by an algo too calculating the protection envelope.



< Message edited by zeke99 -- 7/20/2016 6:13:37 AM >

(in reply to Capt. Harlock)
Post #: 35
RE: New Pacific War Bug Fixes / Enhancements - 7/9/2016 3:32:14 PM   
Rich Dionne

 

Posts: 376
Joined: 7/11/2000
Status: offline
quote:

ORIGINAL: Istfemer

I can already sense the problems that will pop up if your suggestion becomes reality, Zeke.
Raid on Pearl Harbor will be a spectacularly costly failure for the IJNAF. Their bombers will all get shredded by Allied flak and Allies won't lose any ships.
With flak values not reduced, on 12/7/41 the ships in Pearl Harbor port will have a combined flak value of 5599.
(non-historical start, the flak value will be slightly lower in other scenarios)


Thanks Zeke. Let me think about what can be done about this. Istfemer raises some very legitimate concerns about simply adding up all flak values for ships in harbor, and an algorithm such as Capt. Harlock mentions would probably be necessary, as well as taking into account the stationary status of the ships.

Others have mentioned a 2 part flak calculation: 1st part giving full flak credit for the individual ship under attack, and 2nd part modifying in some way the value of the other ships. Perhaps something like full flak for ship under attack plus average value of other ships times number of ships in the TF or port, up to a maximum of say 14. This would have the effect of reducing the value of flak in a port to a maximum of about 15 ships in close proximity to the ship under attack. This wouldn't account for your small calibre weapon effect, but it would perhaps be an improvement. We'd still need to account for stationary ships in port. Also need to account for the flak value of the port itself.

I'll dig into the code and see if I can find the existing algorithm, and what modifications are reasonably possible.

Regards,

Rich

< Message edited by Rich Dionne -- 7/9/2016 3:47:14 PM >

(in reply to Istfemer)
Post #: 36
RE: New Pacific War Bug Fixes / Enhancements - 7/9/2016 4:54:19 PM   
zeke99


Posts: 385
Joined: 11/26/2005
Status: offline
Dear Rich,

your approach with the 14/15 ships sounds OK for me.

To account for the stationary ships I suggest to use a multiplier for the hit probability of bombs and torpedos eg x2.

Thanks again,
Zeke

(in reply to Rich Dionne)
Post #: 37
RE: New Pacific War Bug Fixes / Enhancements - 7/10/2016 4:39:06 PM   
Istfemer

 

Posts: 157
Joined: 7/19/2014
From: Kyiv, Ukraine
Status: offline
quote:

ORIGINAL: Capt. Harlock
Interesting -- that much even before the upgrades? And isn't there an averaging algorithm to compensate for there being many more ships in harbor than are allowed in a TF?

Actually more. 5671. I forgot to add the flak value of Gato-class submarines. (six subs should be worth 72)
Regarding the possible averaging algorithm you mention, no, I don't think there is such a thing in Pacwar.

(in reply to zeke99)
Post #: 38
RE: New Pacific War Bug Fixes / Enhancements - 7/12/2016 1:56:56 AM   
Capt. Harlock


Posts: 5146
Joined: 9/15/2001
From: Los Angeles
Status: offline
quote:

Regarding the possible averaging algorithm you mention, no, I don't think there is such a thing in Pacwar.


I'm a little surprised -- I think Gary Grigsby had such an algorithm way back in "Guadalcanal Campaign". Of course, in that game there was no effective limit on the number of ships in a TF.

_____________________________

Civil war? What does that mean? Is there any foreign war? Isn't every war fought between men, between brothers?

--Victor Hugo

(in reply to Istfemer)
Post #: 39
RE: New Pacific War Bug Fixes / Enhancements - 7/14/2016 3:30:35 PM   
Rich Dionne

 

Posts: 376
Joined: 7/11/2000
Status: offline
Here’s what I’ve found out about flak for ships in port:

1) The AI cycles through all ships and only adds flak for the 1st 15 ships found to be in the port. So like TFs, a maximum of 15 ships can add their individual flak values for the total defense.

2) Low preparation points can lower a ships flak value. A TFs prep pts are compared against a random value between 0 and 50. If the random value is greater than the TFs prep pts, then flak is multiplied by a factor of 0.67. Ships in port have 0 prep pts, so their flak value is always multiplied by this factor.

3) Ships in port have an additional flak reduction multiplier of 0.125, so the combined flak reduction factor compared to TFs is 0.67 x 0.125 = 0.08375 (1/12).

4) The port itself does not add any flak of its own.


Other flak multipliers that apply to both TFs and ships in port consist of the following:

1) All Japanese ships and non-combat Allied ships (class types greater than 13) get a flak factor of 0.9.

2) In June 42, USN ships get a flak factor of 1.1.

3) In Jan 43, USN ships get a flak factor of 1.25.

4) Ships that are surprised get a flak factor of 0.25.

5) Ships with 0 ammunition get a flak factor of 0.25.

6) Japanese ships firing flak at night get a flak factor of 0.25, Allied ships at night get a flak factor of 0.5.

7) All flak against heavy bombers is reduced by a factor of 10.


So, as an example, in Jan 1943, a USN combat ship surprised in port would have applied the following flak factors: 1.25 (USN 1943) x 0.25 (surprise) x 0.67 (low prep pts) x 0.125 (in port) = 0.026.

The combined, factored flak value is compared against a random value between 0 and 8000. If the combined flak value is greater than this random number, the aircraft is hit, although not necessarily destroyed.

Another factor that affects the ability of aircraft to hit ships in port is the evasion rating, which is basically built around a ship's speed. Ships in port have this value divided by 10, making them 10 times easier to hit than ships in TFs.

Looking at the code, it would be very difficult to make any significant changes beyond adjusting the factors. I did run a few tests altering the port flak factor from 0.125 to 0.5, and there wasn't much impact on number of ship hits or aircraft lost...

Regards,

Rich

< Message edited by Rich Dionne -- 7/14/2016 4:01:21 PM >

(in reply to Capt. Harlock)
Post #: 40
RE: New Pacific War Bug Fixes / Enhancements - 7/14/2016 5:00:27 PM   
Istfemer

 

Posts: 157
Joined: 7/19/2014
From: Kyiv, Ukraine
Status: offline
^ Nice & clear explanation ^
I was looking at the same function (sub_20e09) two days ago.
quote:

ORIGINAL: Rich Dionne
1) The AI cycles through all ships and only adds flak for the 1st 15 ships found to be in the port. So like TFs, a maximum of 15 ships can add their individual flak values for the total defense.

Didn't notice this part. Of course you mean ship units, not individual ships. I assume the usual limit on flak-capable ships in a ship unit (= 6 ships per unit a.k.a . stack) still applies. Anyway, I have been proven wrong:
quote:

ORIGINAL: Capt. Harlock
And isn't there an averaging algorithm to compensate for there being many more ships in harbor than are allowed in a TF?
ORIGINAL: Istfemer
Regarding the possible averaging algorithm you mention, no, I don't think there is such a thing in Pacwar.

---
I suspect that even more randomisation may be involved.
I ran a test with USN ships in Manila. The game date was 12 or 21 Dec 1941.
I did two resolutions of the same turn with minimal changes between them. (I switched some Malayan bases to different HQs, that was all I did)
In the first case the flak value vs. a Zero squadron was 0. In the second case it was 16.

I don't think the USN ships in Manila were surprised in either case. What, then, could have caused the flak value to become 0?
---
Or maybe I just don't fully understand how the thing below works and what the FLAK:xxxx number, which flashes each time a hostile squadron begins a strafe/bombing run, represents:
quote:

The combined, factored flak value is compared against a random value between 0 and 8000. If the combined flak value is greater than this random number, the aircraft is hit, although not necessarily destroyed.


< Message edited by Istfemer -- 7/14/2016 5:23:47 PM >

(in reply to Rich Dionne)
Post #: 41
RE: New Pacific War Bug Fixes / Enhancements - 7/14/2016 5:37:02 PM   
Rich Dionne

 

Posts: 376
Joined: 7/11/2000
Status: offline

quote:

ORIGINAL: Istfemer

^ Nice & clear explanation ^
I was looking at the same function (sub_20e09) two days ago.
quote:

ORIGINAL: Rich Dionne
1) The AI cycles through all ships and only adds flak for the 1st 15 ships found to be in the port. So like TFs, a maximum of 15 ships can add their individual flak values for the total defense.

Didn't notice this part. Of course you mean ship units, not individual ships. I assume the usual limit on flak-capable ships in a ship unit (= 6 ships per unit a.k.a . stack) still applies. Anyway, I have been proven wrong:


You are correct Istfemer; I was referring to ship units, with a maximum of 6 ships in the unit contributing flak.

quote:


I suspect that even more randomisation may be involved.
I ran a test with USN ships in Manila. The game date was 12 or 21 Dec 1941.
I did two resolutions of the same turn with minimal changes between them. (I switched some Malayan bases to different HQs, that was all I did)
In the first case the flak value vs. a Zero squadron was 0. In the second case it was 16.

I don't think the USN ships in Manila were surprised in either case. What, then, could have caused the flak value to become 0?


I don't know; but I think you're right, there's probably a lot more randomization going on.

By the way the port limitation on number of ship units counted for flak is located at:

seg007:52C5 cmp [bp+var_C], 0Eh
seg007:52C9 jbe short loc_210A0 ; Limit # of ships adding flak at location to 15

Regards,

Rich

(in reply to Istfemer)
Post #: 42
RE: New Pacific War Bug Fixes / Enhancements - 7/15/2016 5:58:48 AM   
zeke99


Posts: 385
Joined: 11/26/2005
Status: offline
Hi Rich

"3) Ships in port have an additional flak reduction multiplier of 0.125, so the combined flak reduction factor compared to TFs is 0.67 x 0.125 = 0.08375 (1/12)."

I think the easiest way to fix it would be to set this multiplier close to 1.0

(in reply to Rich Dionne)
Post #: 43
RE: New Pacific War Bug Fixes / Enhancements - 7/29/2016 5:54:23 AM   
zeke99


Posts: 385
Joined: 11/26/2005
Status: offline
Test report from V3.2.13

While the new TF routine, ship get damage while in TF, makes it more realistic the ports have problems to keep up with repairs.

Japanese CV do not fill up ac sqn to max allowance of carriers. I think this was always the case in this game. Would be nice if it could be fixed.

I still have the impression that replenishment of losses in LCUs & sgns is slower than it used to be although the pool is overflowing with manpower and gear.


(in reply to zeke99)
Post #: 44
RE: New Pacific War Bug Fixes / Enhancements - 7/31/2016 9:47:43 PM   
Rich Dionne

 

Posts: 376
Joined: 7/11/2000
Status: offline
Thanks for testing Zeke!

I'll look into the CV aircraft allowance issue. I've noticed this occasionally, but not sure why it happens.

I may have done something to slow down LCU refit / replacement to prevent rapid conquering of territory. Combat units historically took more time to recover from previous combat mission than is seen it the game. I'll review my code changes to confirm whether I've done anything here.

Regards,

Rich

(in reply to zeke99)
Post #: 45
RE: New Pacific War Bug Fixes / Enhancements - 8/1/2016 5:08:22 AM   
zeke99


Posts: 385
Joined: 11/26/2005
Status: offline
Thanks for your work Rich, much appreciated.

Have you also changed something in the routine convoys? I get notifications that bases get low 1-2digit numbers (eg 6 or 11) in supplies, hardly worth to send one MCS.

(in reply to Rich Dionne)
Post #: 46
RE: New Pacific War Bug Fixes / Enhancements - 10/21/2016 7:04:50 PM   
Rich Dionne

 

Posts: 376
Joined: 7/11/2000
Status: offline
Zeke, Sorry for the long wait for a response to your question. I did alter the routine convoys, but I've now hopefully fixed the issue. See below!

Here's version 3.2.14 of PAC.EXE available for download. A list of ongoing fixes / modifications plus a new required file is also included in the zip file. You can download it here:

Pacific War Update Version 3.2.14

The latest version 3.2.14 EXE fixes / features are shown below. Note in particular item 64 and a new file “eport.cbt” that you must now include in the same directory as the EXE for the game to work. The new file is included in the zip file.

57) Fixed a couple of bugs that occasionally caused game “freezing”. These include the freeze that occurred when pressing “shift-4” to look at Japanese merchant ship counts, and the freeze that occurred if the “gus” setting in the Dosbox config.sys file is set to “false”. Hopefully these fixes will eliminate Zeke’s replenishment routine freeze issues.

58) Removed previous change 24b whereby “The number of ships required in routine convoys is now affected by the distance from the source to the destination.” And “The maximum number of ships allowed in one routine convoy ship unit has been increased from 50 to 75 to help with this change.” In looking at AI performance, the Allied AI appeared to be severely handicapped by this change. I’ll look for less intrusive ways to properly show the impact of distance on routine convoy capabilities.

59) Made some more changes to the way the AI sets paths for Allied TFs. You should see a little less suicide moves into Japanese territory.

60) Altered the port flak modifier from 0.125 to 0.5. This increases flak effectiveness in port by a factor of 4, but it is still ½ as effective as TF flak. I tested this change on the Pearl Harbor attack; it doesn’t affect the outcome, but a few more Japanese aircraft are lost.

61) Altered TF flak to make it more effective against small air strikes and less effective against large air strikes. Currently, percentage aircraft losses against TF flak are the same whether the air strike is large or small. For example, a TF flak rating of 4000 will hit on average about 50% of the attacking aircraft. So a strike of 200 aircraft will have about 100 hit, and a strike of 20 aircraft will have about 10 hit. This doesn’t seem right to me. All that flak concentrating on a strike of 20 aircraft should hit a higher percentage. So I’ve change the code as follows: a strike of 80 aircraft remaining after CAP will lose the same number of aircraft as currently. A smaller strike will lose a larger percentage of its aircraft, approaching about 30% more than a strike of 80 as the number of aircraft is decreased linearly to zero. Likewise, a larger strike will lose a smaller percentage of its aircraft, approaching about 30% less than a strike of 80 as the number of aircraft is increased linearly to 240 aircraft.

62) Increased the flak value of LCUs. You should now take some aircraft losses against large LCU groupings.

63) Fixed code that was frequently isolating Adelaide. You should see this occur less frequently.

64) I’ve made my first foray into changing how the AI behaves related to setting HQ targets. You should now see some differences in the how the AI conducts its offensive moves. So far, the changes are all on the offensive side; so you will see changes mostly in the Japanese offensive period (Dec 41 to around Jan 43), and the Allied offensive period (Mid 43 to end of war). You’ll need a new file in your game directory, which I’ve included, “eport.cbt” (not report.cbt). The game won’t run properly if you don’t have this file in the game directory. I’d call this change still a work in progress, but I’d like you all to try some test games against the AI, with you on the defensive (Allies in early war, Japan in late war) and let me know what you think. This change should not affect human vs. human games.

65) Changed the civilian kill point divisor from 100 to 20. So losing 1000 civilians nets 50 victory points instead of 10 victory points. This puts civilian kill victory points equivalent to infantry losses: currently losing 100 squads = 1000 soldiers net 50 victory points.

As usual, please let me know if you find any issues / bugs.

Best Regards,

Rich

(in reply to zeke99)
Post #: 47
RE: New Pacific War Bug Fixes / Enhancements - 1/24/2017 3:49:43 AM   
zeke99


Posts: 385
Joined: 11/26/2005
Status: offline
Thanks Rich & no problem.

Just downloaded the zip and will try to run a test game over next weekend.

Cheers

(in reply to Rich Dionne)
Post #: 48
RE: New Pacific War Bug Fixes / Enhancements - 1/29/2017 2:35:40 AM   
zeke99


Posts: 385
Joined: 11/26/2005
Status: offline
Test report turn 1, Raising sun scenario, Jap AI, max help
Pearl attack: AD starts 187, looks low
Did 2 starts, losses between 1-4BBs plus other, looks OK

Note LCU: There is NO Aussi LCU to be loaded and shipped to Port Moresby, did I miss it?

Notes ac: If don't like the auto upgrades, can't find how to switch to human factory ctrl. It would be better if it stops when HQ is set to human ctrl.

Vilderbeast & Albacore TB, none in the pool and no changes possible (old problem.
Fullmar, does anybody use this plane? Can the factory be changed to Hurricane?

Will post if I encounter strange things or bugs.
Cheers

(in reply to zeke99)
Post #: 49
RE: New Pacific War Bug Fixes / Enhancements - 1/29/2017 4:51:17 PM   
Capt. Harlock


Posts: 5146
Joined: 9/15/2001
From: Los Angeles
Status: offline
quote:

Notes ac: If don't like the auto upgrades, can't find how to switch to human factory ctrl. It would be better if it stops when HQ is set to human ctrl.


If memory serves, press the Alt and N keys.

quote:

Vilderbeast & Albacore TB, none in the pool and no changes possible (old problem.
Fullmar, does anybody use this plane?


I believe one of the early British CV's has a Fulmar squadron -- which is historical, but suicide against Zeroes.



_____________________________

Civil war? What does that mean? Is there any foreign war? Isn't every war fought between men, between brothers?

--Victor Hugo

(in reply to zeke99)
Post #: 50
RE: New Pacific War Bug Fixes / Enhancements - 1/30/2017 3:05:12 AM   
zeke99


Posts: 385
Joined: 11/26/2005
Status: offline
Alt N worked, many thanks

HMS Indomitable, I always change to F4s asap.

Turn 01/18/42 Wirraway can not change to P40 only Hurricane. #27 EXE Version update.RTF says otherwise.

LCU 7th & 30th Inf Brig appeared in Sydney assigned to SW Pac @ 18% readiness. Loaded one on AP and had to unload. Readiness went up to 51%, strange.

(in reply to Capt. Harlock)
Post #: 51
RE: New Pacific War Bug Fixes / Enhancements - 2/14/2017 12:48:44 PM   
zeke99


Posts: 385
Joined: 11/26/2005
Status: offline
Realised that it makes more sense to test both sides I restarted the test, Scenario Raising Sun, Hist 1st move:

Remarks: In a Human vs Human game all HQs should be @ Full Human Ctrl by default

Many changes with LCU have been done. Needs some rethinking of deployment.

Japanese side: 17th & 18th Army a subordinate to 8th Area Army which is not there yet. 3LCU @ Okinawa are assigned to 32nd Army, also not there. I hope this will not cause supply issues.

(in reply to zeke99)
Post #: 52
RE: New Pacific War Bug Fixes / Enhancements - 2/15/2017 12:36:45 PM   
KurtC


Posts: 1189
Joined: 7/1/2011
Status: offline
I bought this game when I was a teenager and I used that teleport bug because it made me mad that some places I thought should have supply didn't have supply. Mostly in the India area.

_____________________________

My known forum name is H Gilmer

He that has a mind to fight, let him fight, for now is the time. - Anacreon

(in reply to Rich Dionne)
Post #: 53
Page:   <<   < prev  1 [2]
All Forums >> [Classic (Free) Games] >> Pacific War: The Matrix Edition >> RE: New Pacific War Bug Fixes / Enhancements Page: <<   < prev  1 [2]
Jump to:





New Messages No New Messages
Hot Topic w/ New Messages Hot Topic w/o New Messages
Locked w/ New Messages Locked w/o New Messages
 Post New Thread
 Reply to Message
 Post New Poll
 Submit Vote
 Delete My Own Post
 Delete My Own Thread
 Rate Posts


Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI

0.157