• There seems to be an uptick in Political comments in recent months. Those of us who are long time members of the site know that Political and Religious content has been banned for years. Nothing has changed. Please leave all political and religious comments out of the forums.

    If you recently joined the forums you were not presented with this restriction in the terms of service. This was due to a conversion error when we went from vBulletin to Xenforo. We have updated our terms of service to reflect these corrections.

    Please note any post refering to a politician will be considered political even if it is intended to be humor. Our experience is these topics have a way of dividing the forums and causing deep resentment among members. It is a poison to the community. We appreciate compliance with the rules.

    The Staff of SOH

  • Server side Maintenance is done. We still have an update to the forum software to run but that one will have to wait for a better time.

Wtrexpl effects & damage

UncleTgt

UncleTgt

SOH-CM-2025
Way back in 2004 Pen32win discovered that the grndexpl damage values were well off, & that's why when you or your AI wingmen missed hitting their target, even slightly almost no damage seemed to be inflicted. As these were not smart munitions, that's most of the bomb drops seen in the sim.

In the AF_Bomblast package, he revised these 3 fx files, & amended a lot of the larger building DPs etc to take advantage. Now if you missed, you still saw evidence of blast damage.

Now, I'm beginning to suspect the 3 wtrexpl fx files have the same issue, & I'm wondering whether there's a compelling reason the AF Team didn't revise the water explosion fx files in the same way.

I get that there'll be less shrapnel flying around, compared to a ground explosion, but even a 250lb (Type 1) bomb going off close aboard a 4t Shohatsu landing craft should/could still swamp it/ open up the hull to leaks due to hydrostatic shocks ... At the moment it just shrugs it off, unless I get a direct hit...

JAP_Shohatsu_cargo_UTe.jpg
 
Meantime, I'm heading down the rabbit-hole of ship DPs & damage effects.

First up, a recolouring of the ugly matt black oil slick fx...
recoloured_oilslick-debris_fx_1.jpg
 
Next, a flooding/fuel leak effect that continues to damage after the initial hit (like some of the fire effects do)

flooding_oil_leak_port_fx_WIP3.jpg
 
& how a bout a large fire that billows from the hangar sides of a carrier, instead of through the flight deck...Hangarfire_fx_trial3.jpg
Hangarfire_fx_trial2.jpg
 
John,
I cannot find anything related to this in the Resources list – nothing uploaded by Pen32win, nothing called gndexpl, nothing called AF_Bomblast. If you do have a copy of the original file, perhaps you could upload it here, please.

FWIW, if modifying the gndexpl.fx files meant that some buildings needed updating to match, then I can easily imagine that doing the same for the wtrexpl.fx files would need an even bigger effort to make the ships compatible.

Anyway, your results to date look very impressive – keep up this good work . . . and thanks for all your updates . . .
 
Ro,

It's over at Simviation, along with all the other "old" AF Midway Team enhancements.

Look for: - af_bb_pak.zip

Now, back to testing wtrexpl fx files to see if there's any difference between my new ones & the CFS2 originals... bombs away!
 
Last edited:
Whilst trying to develop better DPs for my JAP landing craft, I was struck by how "indestructible" small vessels are in CFS 2. If you're toting anything lighter thasn a 500lber you can forget it, & even then, those Daihtsus in 87 Days just never seem to sink!!!

Now, way back in 2004 Pen32win revised the grndexpl fx files that modelled land explosion "near-misses", as he'd uncovered that the reason CFS2 originally shipped with such puny object DPs for ground targets was how these fx files imparted damage for bombs that missed their target (ie nearly all "dumb" bomb explosions).

From the AF Bombblast readme...

...

Here’s why:
CFS2 creates ground bomb damage in two different ways.
1. A direct hit on the infrastructure item’s DP.
2. A near miss with the bomb striking the ground.

...

Case number 2 is were things get messy. First off there are 4 different “weapon_type”s that cover all of the different bombs. Type 1 is the lightest, like the 30 & 60kg and 250Lb’ers. Type 2 covers the 250kg and 500lb bombs. Type 3 is for the 1000lb bomb and Type 4 is used by the 800kg and 2000lb bombs. Now that that’s clear let’s make mud. When a bomb strikes the ground in CFS2 the program selects one of 3, yes 3 and not 4 FX to trigger. You’d think there would be 4 FX since there are 4 different weapons types, right? Nope. Here’s how things work out. The 3 Effects are the fx_gndexpl_s, m & l. Type 1 bombs use the “s” FX. Type 2 uses the “m” FX and Types 3 & 4 both use the “l” FX. Why does this matter? In CFS2 the AI aircraft do not hit their targeted objects DP unless it is a very large object like the Factories or Hangar Type A Large. They even miss those sometimes. So 97+% of the time when an AI Aircraft is attacking with bombs the resulting damage is caused by the 3 grndexpl FX. And that is what was making CFS2’s ground bomb damage act so erratic and why the ground objects DP were set so low in the number of hit points required to destroy them.

The Stock FX numbers relating to damage they cause look like this:

Small:
Hitpoints=500.00, 600.00
Range=100.00, 200.00

Medium:
Hitpoints=150.00, 200.00
Range=100.00, 200.00

Large:
Hitpoints=2000.00, 2200.00
Range=100.00, 200.00

No, I didn’t paste them out of order and Yes the above numbers match of weapon type is correct. I know this for certain because I added different colored smoke to each of the FX for testing. Just ask my Beta testers how much that freaked them out…LOL OK, these FX are the root of the problem. They don’t match the damage numbers in the weapons DP’s as they should. CFS2 is very good at reproducing a bomb blast radius but they dropped the ball on the FX end and ended up creating wimpy DP’s for the Ground Infrastructure to compensate.

...

His solution was to revise the hitpoints & blast radius values for the 3 fx files

...

Small:
Hitpoints=1000.00, 1500.00
Range=50.00, 75.00

Medium:
Hitpoints=5000.00, 5500.00
Range=90.00, 100.00

Large:
Hitpoints=10000, 20000
Range=100.00, 150.00

Notice how they match up with the different weapon types better? I had to do some fudging with the Large FX because of the lack of a 4th FX for the Type 4 Weapons but it works out well. What these FX do is allow the creation of more realistic Ground Object DP’s without making them indestructible to bombs. BTW, this will also fix the damage done by the AF=Midway Arty Shells and any aircraft that crashes near a building.

...

=============================================================================================

Fast forward to 2024...

Now, it seems nobody ever looked at water-based explosions, but it looks to me like they were modelled in exactly the same way. Again, remember most dumb bombs miss, but land close to the intended target. So there should be a chance of a near-miss inflicting blast damage - hydrostatic overpressure popping hull rivets/welds & creating leaks, the blast itself (but no ground to create extra shrapnel from), or maybe just the sheer volume of water swamping an open-topped landing craft.

Here's the stock fx file values

Small:
Hitpoints=30.00, 60.00
Range=20.00, 40.00

Medium:
Hitpoints=50.00, 100.00
Range=30.00, 60.00

Large:
Hitpoints=60.00, 120.00
Range=50.00, 100.00

Wimpy or what - no wonder your Airacobra's single 500lber never seemed to do any damage except if your were lucky enough to actually hit something!! Any those pesky Daihatsu's are real small...

So I went ahead & created a replacement set using Pen32win's grndexpl values as a guide...

Small:
Hitpoints=500.00, 750.00
Range=50.00, 75.00

Medium:
Hitpoints=2500.00, 3000.00
Range=90.00, 120.00

Large:
Hitpoints=5000, 10000
Range=100.00, 150.00

So, I'm using Pen32win's range values, but have reduced the hit points to approx. 50% of the grndexpl fx values to account for there being fewer shrapnel pieces.

Then I set about testing my hypothesis. I created a 9 vessel convoy attack mission template & reproduced it as follows:

Landing Craft attack
A mix of my own Shohatsu, Daihatsu & Toko Daihatsu models, being attacked by 7 pairs of SBDs, each using a single 250lb bomb

Maru attack missions, 3 types: 250lb, 500lb & 1000lb

The convoy consists of:
2 x 6800t Maru (my own JapMaru1 mdl, based on Stuart's Daifuku Maru mdl)
2 x 4000t Maru (my own JapMaru2 mdl)
2 x 1200t Tankers (Stuart's TSa tankers)
2 x TB escorts (Stuart's SC13 class)
1 x Daihatsu at the convoy lead (Used as a location to watch the action unfold each time)

I just put the Player aircraft into autopilot & then F10 to get trail view. Using the hat-switch I could then cycle back/fwds to watch the different targets as required.

Here's the results: (number of vessels sunk after 14 bombs dropped)

Stock wterexpl.fx (ratio) my revised wtrexpl.fx (ratio)
===========================================================================================

JapLC attack, 250lb 5 , 3 , 5 , 4 , 4 21/70 7 , 8 , 5 , 8 , 9 37/70

JapMaru attack, 250lb 1 , 1 , 1 , 0 , 1 4/70 0 , 1 , 1 , 1 , 1 4/70

JapMaru attack, 500lb 1 , 1 , 1 , 1 , 1 5/70 1 , 2 , 1 , 1 , 1 6/70

JapMaru attack, 1000lb 1 , 2 , 3 , 1 , 1 8/70 3 , 3 , 1 , 2 , 2 11/70

Not perfect, but a definite effect there, particularly with the smaller craft!!

I like the fact that the larger vessels don't have much difference between the two sets, as I'd expect any ship kills to be driven by actual hits rather than near-misses. However, I also noted that all types craft seemed to take/display more visual damage effects when using my files than they did using the stock files, perhaps indicating the increased radius was indeed having an impact from near-misses, like I'd hoped it would.

Now aircraft gunfire or small bombs can have a real impact on small craft, without necessarily making them too easy to kill.

Any thoughts?
 
Hi UncleTgt
The effects look really good.
I do like your research. Smaller craft are hard to slow down.
A science prize should be handed out.
:medals:
Cheers
Stuart

PS
Here is an interesting one that may have already been solved.
Large ships sink quickly, whereas most large ships take a long time to sink, certainly when the attacking aircraft are long gone.
It could possibly be solved with damage point changes ie removing 'sink' and have a damage model sitting low in the water, but it may open another can of worms
 
Hi Stuart,

I think the "SINK" command animation is hard-coded, so no chance to amend.

I'm not sure the damage model idea works with Ships, only with aircraft models. I have been considering a variation of your idea for use with ships in harbour (Pearl, Taranto etc.). When ships were "sunk" in harbour locations they almost always bottomed with their deck awash, rather than sank.

So, it should be possible to:

Use the normal ship model for an undamaged vessel, placed as static infrastructure in a mission.
Maybe the DP needs tweaking to ensure it sinks when hit, but when the vessel "SINKS", a damaged model is spawned in it's place at the same location.
 
I have a different issue with small craft. If I use the Unit Type=19 (Merchantmen), they invariably attract AI attacks (because of the low hardness/protection values) & die quickly. This is especially noticeable in a mixed convoy scenario, where the AI will not attack the larger transports/cargo ships but prefers the easier target that the small craft pose.
It suggests the reason for coding small stuff as =24 amphibious craft (even when they're not) was a work around. If I code them this way, the AI flights ignore the small craft entirely (too small), & correctly go for the larger stuff.

It suggests there is a different attack profile for amphibs, rather than the standard anti-shipping attack profile, but I've yet to discover one that works.

Has anyone ever devised a mission where they got the AI to attack amphib type targets whilst moving in open water without having to issue attack directives as the Player?
 
Well, that prompted a lively discussion ...:stupid:

After a fair amount of trial & error, I was able to get the AI to drop bombs on Amphibs without prompting from me.

The AI flight has to be flying "Search & Destroy", & must overfly the target area at low level, ideally approaching at <1000ft. Only then to they spot & move to engage. Unfortunately, if flying near some interesting ground units or gsl, a significant portion of your strike will wander off, sometimes almost at right angles to your planned route, as they spot something "more interesting" to bomb!!

So I think there's still a use for small craft identifies as =19 "Transports" - I guess I'll supply both as options...JapMarus_under_attack2.jpg
 
As a follow up to this, I ran a series of experiments.

Experiment - AI Behaviour Attacking Category 24 (Amphibious Vehicles) Targets UncleTgt AUG 2024

Mission Set-up

A staggered series of seven 2 plane sections, each armed with a single 250lb bomb.

The target is a formation of 9 JAP Landing Craft of mixed types, sailing NNE just 3nm SE of Palau, beginning 5nm S of Babelthuap airstrip. Each vessel has it's course charted individually, with "do not evade" selected for each waypoint.

Each section approaches from the South, ie the stbd rear quarter of the Landing Craft. All flights waypoints are at the same altitude, & they have an "attack with bombs" waypoint at 6nm short of the target convoy. They then have a flightpath that takes them over the convoy, & circles them around to stbd. The tightest circle is 3.5nm diameter, the largest 7nm diameter. On returning across the convoy's path, they each have an "attack with all weapons" waypoint added, approximately 2.5 to 3.5nm short of the convoy's path.

The attackers were spawned in sequence, so that each Section's attack could be viewed separately more easily.

Thunder = 0 min delay
Cutie = 2 min delay
Copper = 3 min delay
Timeless = 6 min delay
Concrete = 6 min delay, but spawns further East than Timeless ( ie further out to sea)
Crossroads = 9 min delay
Bullfrog = 10 min delay

This mission was then run with different variations of:

Flight Altitude: 1000, 2000, 3000, 4000, 6000, 8000, 11000ft
Mission Directive: Strike, CAS, Search & Destroy

The initial runs were using the Donationware SBD-5, which had a unit category = 10 in the DP file.

Results

Directive: Search & Destroy

Alt Nr of Sections attacking shipping Nr of Sections attacking on 1st pass

1K 4 out of 7 100%

2K 4 out of 7 100%

3K 3.5 out of 7 100%

4K 2 out of 7 100%

6K 2 out of 7 100%

8K 2 out of 7 100%

11K 2 out of 7 100%

Comments:

There does seem to be a definite sequence in deciding the targets for each section as they spawn.
In all cases the 4th section to spawn (Timeless) always chose a land target instead of shipping.
This directive does not result in any overflying of the convoy, or any chosen land-based target, they always attack on the first pass. The only caveat is that sometimes a section's wingman would not release because his leader was "in the way".

Directive: CAS

Alt Nr of Sections attacking shipping Nr of Sections attacking on 1st pass

1K 4.5 out of 7 c.50%

2K 4.5 out of 7 c.50%

3K 5.5 out of 7 c.50%

4K 4 out of 7 c.50%

6K 5.5 out of 7 c.50%

8K 4 out of 7 c.50%

11K 5.5 out of 7 c.50%

Comments:

There does seem to be a definite sequence in deciding the targets for each section as they spawn.
In all cases the 4th section to spawn (Timeless) always chose a land target instead of shipping.
This directive seemed to have a higher number of shipping attacks, compared to the Search & Destroy directive, however this also includes the sections that overflew the convoy first, & then either followed my original waypoints to circle round for another pass, or detected a target too late & circled around to attack on the second pass. In all cases the sections followed a clockwise circuit as I had plotted, so I'm inclined to believe they were indeed following my original waypoints, & detected a target on the second approach. Often this meant they Glide Bombed Shipping in beam-on or rear quarter attacks, instead of manouevring for a position directly astern.

The key differences are the higher number of shipping engagements, but the lower number of first pass attacks. The CAS directive seemed to result in more sections returning to attack with all weapons (ie MG strafing) towards the end of the run than the other directives.

The favourite land target was the gsl encampment around Palau's South Lighthouse, if selected during the initial approach. This target lay 45-50 degrees port of the flightpath, & like the convoy, was 6nm away when selected. If a land target was selected during the second approach, it was usually the gsl bridge camp SE of Koror island, which meant flying past the rear of the convoy, which would be a perfect set up for an anti-shipping attack at the higher altitudes - go figure! This target was generally within 30 degrees of direct ahead of the flightpath & about 5-6nm away when selected. The convoy was generally closer to (c.3 - 4nm) & was more directly ahead (c.15-20 degrees off), but was not chosen.

Directive: Strike

Alt Nr of Sections attacking shipping Nr of Sections attacking on 1st pass

1K 4.5 out of 7 c.50%

2K 5 out of 7 c.50%

3K 4 out of 7 c.50%

4K 4.5 out of 7 c.50%

6K 4 out of 7 c.50%

8K 4 out of 7 c.50%

11K 4.5 out of 7 c.50%

Comments:

There does seem to be a definite sequence in deciding the targets for each section as they spawn.
In all cases the 4th section to spawn (Timeless) always chose a land target instead of shipping.
This directive seemed to have a higher number of shipping attacks, compared to the Search & Destroy directive, however this also includes the sections that overflew the convoy first, & then either followed my original waypoints to circle round for another pass, or detected a target too late & circled around to attack on the second pass. In all cases the sections followed a clockwise circuit as I had plotted, so I'm inclined to believe they were indeed following my original waypoints, & detected a target on the second approach. Often this meant they Glide Bombed Shipping in beam-on or rear quarter attacks, instead of manouevring for a position directly astern.

The key differences are the higher number of shipping engagements, but the lower number of first pass attacks. I was surprised by how similar the results were to the CAS directive, I expected "Strike" to mean the AI would prefer more static targets.

The favourite land target was the gsl encampment around Palau's South Lighthouse, if selected during the initial approach. This target lay 45-50 degrees port of the flightpath, & like the convoy, was 6nm away when selected. If a land target was selected during the second approach, it was usually the gsl bridge camp SE of Koror island, which meant flying past the rear of the convoy, which would be a perfect set up for an anti-shipping attack at the higher altitudes - go figure! This target was generally within 30 degrees of direct ahead of the flightpath & about 5-6nm away when selected. The convoy was generally closer to (c.3 - 4nm) & was more directly ahead (c.15-20 degrees off), but was not chosen.


Food for thought - or not?

Possible Next steps

My first thought is this needs repeating several times, to minimise the effect of "random" events from the observations made.
The obvious next step would be to repeat the same sequence, just changing the attacking aircraft category.

If anyone wants to have a go & report back their own observations, I've attached the mission templates I used to this post.
 

Attachments

  • JAP_LC_AI_watch_Mission_experiments.zip
    65.3 KB · Views: 3
Further update:


So I went ahead & ran the missions again, just changing the AI SBD aircraft category in it's DP file.

Runs were made using the Donationware SBD-5, which had the unit category changed to = 3 in the DP file.

Results

Directive: Search & Destroy

Alt Nr of Sections attacking shipping Nr of Sections attacking on 1st pass

1K 4 out of 7 100%

2K 4 out of 7 100%

3K 4 out of 7 100%

4K 2 out of 7 100%

6K 2 out of 7 100%

8K 2 out of 7 100%

11K 2 out of 7 100%

Comments:

Much the same as for category = 10, except when flying above 10000ft the AI will do a vertical dive. Unfortunately this category resulted in LESS ACCURATE bombing, with no vessels sunk!!! At lower levels the AI adopt glide bomb approaches, & the effectiveness improves. Conclusion: Don't use Category 3 Dive Bombers to strike small vessels, especially with a higher altitude approach!!

Directive: CAS

Alt Nr of Sections attacking shipping Nr of Sections attacking on 1st pass

1K 5 out of 7 c.50%

2K 4.5 out of 7 c.50%

3K 5 out of 7 c.50%

4K 4.5 out of 7 c.50%

6K 5 out of 7 c.50%

8K 4.5 out of 7 c.50%

11K 6 out of 7 c.50%

Comments:

Much the same as for category = 10, except when flying above 10000ft the AI will do a vertical dive. Unfortunately this category resulted in LESS ACCURATE bombing, with only 1 vessel sunk!!! At lower levels the AI adopt glide bomb approaches, & the effectiveness improves. Conclusion: Don't use Category 3 Dive Bombers to strike small vessels, especially with a higher altitude approach!!

Directive: Strike

Alt Nr of Sections attacking shipping Nr of Sections attacking on 1st pass

1K 4.5 out of 7 c.50%

2K 5 out of 7 c.50%

3K 4.5 out of 7 c.50%

4K 5 out of 7 c.50%

6K 4.5 out of 7 c.50%

8K 5 out of 7 c.50%

11K 5 out of 7 c.50%

Comments:

Much the same as for category = 10, except when flying above 10000ft the AI will do a vertical dive. Unfortunately this category resulted in LESS ACCURATE bombing, with only 1 vessel sunk!!! At lower levels the AI adopt glide bomb approaches, & the effectiveness improves. Conclusion: Don't use Category 3 Dive Bombers to strike small vessels, especially with a higher altitude approach!!
 
And again...

Runs were made using the Donationware SBD-5, which had the unit category changed to = 4 in the DP file.

Results

Directive: Search & Destroy

Alt Nr of Sections attacking shipping Nr of Sections attacking on 1st pass

1K 4 out of 7 100%

2K 4 out of 7 100%

3K 2 out of 7 100%

4K 2 out of 7 100%

6K 2 out of 7 100%

8K 2 out of 7 100%

11K 2 out of 7 100%

Comments:

Seems similar to category 10 & 3, except the target selection "distractions" kick in earlier at a little lower altitude, reducing strike effectiveness.

Directive: CAS

Alt Nr of Sections attacking shipping Nr of Sections attacking on 1st pass

1K 5 out of 7 c.50%

2K 5 out of 7 c.50%

3K 4 out of 7 c.50%

4K 4.5 out of 7 c.50%

6K 4.5 out of 7 c.50%

8K 4.5 out of 7 c.50%

11K 4.5 out of 7 c.50%

Comments:

Similar to both category 10 & 3.

Directive: Strike

Alt Nr of Sections attacking shipping Nr of Sections attacking on 1st pass

1K 5 out of 7 c.50%

2K 5 out of 7 c.50%

3K 4.5 out of 7 c.50%

4K 4.5 out of 7 c.50%

6K 5 out of 7 c.50%

8K 4.5 out of 7 c.50%

11K 5 out of 7 c.50%

Comments:

Similar to both category 10 & 3.
 
I also tried a sample of the same missions with the Donationware SBD-5 unit category changed to = 2, 5, or 6 in the DP file.

In all cases the AI behaves as dumb level bombers, with the section leader reporting bombs dropped at the first attack waypoint, & continuing to fly the waypoints set without any response to the convoy at all. Note, that although the message says bombs were dropped, the external weapons remain in place, & there is no evidence of bomb splashes in the empty ocean below. These categories seem to require a fixed point or named target in the flightplan in order to get them to manouevre.
 
Goddamn, I don't know a thing about effects in general, that look like a lot of work in programming, code, etc. Thanks for all of this work Uncle :applause:
 
You must log in or register to reply here.
Back
Top