Realism- and gameplay-related hardcode fixes for SH3.EXE

[h.sie]

Hi raymond,

funny, just in the last minute I tried to find a flag or similar that shows if the repair team is working or not, in order to enlarge CO2 production accordingly.

Good idea to analyse the 4 or 5 states. Let's hope I find a way to do that.

h.sie


Note: Considering crew number for CO2 production already works. But it can be switched OFF easily for those who don't like it.

[Delareon]

i know its not 100% on topic but if i need to have the starforce free version of the game to use this mod i think i can ask in here
How did i find the starforce free version?
I searched a while on amazon but i cant find any information if the versions they provide are with or without Starforce.
Well i dont mind to spend a few bucks on SHIII to buy it a second time but i will mind if i buy it again with starforce ****

[SquareSteelBar]

http://www.subsim.com/radioroom/show...6&postcount=59

click the posted links...

Go to your next pc market or supermarket and look for the 'Green Pepper' edition or 'Purple Hills' edition. They're StarForce free and already patched to v1.4b...

[reaper7]

Is link down http://www.mediafire.com/hsie

Any chance of a upload to other provider .


EDIT: Ignore working now

[LGN1]

Originally Posted by raymond6751

I believe I read somewhere that the game can alter the game speed while hunting or being hunted. I suggest that such times of high stress might be a time of more use of breathable air. So then you would have four states, perhaps five if damaged.
normal, silent-running, hunting (in contact), being hunted (under attack), and damaged with DC crew working. These are suggested signals that you can get from the game, perhaps. Just a suggestion.

That's true. The max. time compression possible depends on your 'state':

[TIME COMPRESSION]
TimeStop=0
RealTime=1
LandProximity=256
CriticalDamage=1
CrewEfficiency=1
AirEnemyDetected=1
SoundEffects=2
CharacterAnim=4
EnemyDetected=1
RadioReport=8
Particles=8
PrayState=8 <--- when the enemy has detected you
HunterState=32 <--- if you have contact
3DRender=32
Maximum=1024

It's an easy way to cheat and find out whether the enemy has detected you (you can avoid it by setting it equal to HunterState).


Since you are so successful, h.sie, some more ideas for bugfixes

- Detection probability is influenced by repairs, not only rpm and torpedo reload. At the moment repairs do not increase the chance for being detected.

- Depth charge explosions influencing the sensors (I guess that's a difficult task with possibly far-reaching consequences for game play).

Cheers, LGN1

[h.sie]

Hi LGN1,

nice new ideas, I put them on the todo-ist. unfortunately my progress is very slow. still working on the CO2. it's hard for me to find out where the states (hunting, being hunted, praying, normal) are stored and how they are coded.

but: I found a very elegant, accurate and easy way to calculate the amount of repair- and flooding-recovery-work that is done per time-period. this can be used to calculate CO2 production. In my opinion the repair-work is more important for CO2 production than the above-mentioned states.

but maybe I can locate the states in the next days?

h.sie

[h.sie]

Currently I found out that the routine which calculates CO2 production is executed only every 10 minutes, e.g. at 12:00, 12:10, 12:20 and so on!! You can see this at high Time Compression when you look at the CO2 gauge. It moves step-wise every 10 minutes. This means, the current method of CO2 calculation (in V15C) is inaccurate because the state of the Silent-Running flag is also evaluated only every 10 minutes. The whole timespan inbetween these points of time does not influence CO2 production. So the CO2 calculation could be cheated easily by setting Silent-Running = ON only at 12:00, 12:10, 12:20 for some seconds and OFF for all the other time.

For a more accurate CO2 calculation, one needs to consider the whole time span inbetween these 10 minutes points. This is equal to averaging over the time. This can easily be done by tweaking the routines which are responsible for repairs and flooding-recovery, because these are executed every second. The more repairs are done - the more guys are working, the more a certain special variable is accumulated. Easy to program. This variable can then be evaluated (and resetted) every 10 minutes by the CO2 routine so that you have an accurate measurement of the work that has been done in the last 10 minutes.

So many ideas arise, but trying to keep things easy and manageable, I ask myself if it makes sense to additionally consider things like Torpedo reloading work, psychological stress when hunting or being hunted and so on, since these things are (halfway) correlated to repairs. When crew does repairs, there is/was of course psychological stress, and the UBoat was of course hunted and surely Torpedoes have been reloaded, too. This assumption is not always true, but accurate enough in my opinion.

But if I don't consider these other effects (directly), I think it's a good idea to consider them indirectly by giving for repairs a bigger CO2 penalty than only a factor of 1.5.

Phew,
h.sie

[LGN1]

Hi h.sie,

I would not spend too much time on the CO2 stuff. IMHO, it's fine if there are only two states: work (whatever it is) and rest (triggered by silent running). Usually, you will not use silent running if there is no threat. In this case there is also no need to stay submerged for ages. So, it's more or less only important if there is a threat and then work is the main difference.

I also don't think that the 10 min. update interval is a problem (btw, it's the same with fatigue. Every 10 min. the fatigue is decreased). If someone wants to use it to cheat then let him cheat.

Concerning the two points I mentioned earlier: I think there must be a variable that contains the probability for being spotted. I guess the stealth meter shows its value. Would be great to identify this variable because it would allow to easily modify many aspects (thermal layers depending on depth value , repair noise,...). In combination with a variable that would somehow indicate exploding depthcharges, deaf sensors should be easy to achieve

Anyway, thanks a lot for your continuous work! You are starting to become a legend

Cheers, LGN1

[h.sie]

My new routine to calculate CO2-production is nearly finished. While doing some tests and enhancements, I would like to describe how it works. The intention behind this is to get some feedback in order to fine-adjust some relevant parameters.

1) Crew size now proportionally affects CO2-production and thus max. diving time: 10% more crew members -> 10% higher CO2-production -> 10% shorter diving time.

2) Repairs now affect CO2-production, depending on the amount of repairs that is currently done. For extreme damage and most of the crew doing repairs, the current CO2-production rate is multiplied with a factor of 2,0. (This appears to be a little bit high, but with this I try to compensate, that things like Torpedo reloading work, psychlogical stress when being hunted, chlorid gas caused by damaged batteries and so on, are not considered directly for CO2).

3) In Silent running, your crew stops its normal work, only the minimal necessary work is done, so that the current CO2- production is reduced by a factor of 1,5. This was already modeled in Patch V15C, but now the calculation is much more accurate.

I will also rise the max. diving time of the UBoats.

Example: Type VII.
Diving Time in stock SH3 game: 36h.
Historical data (for 37 crew members): 72h.

So in V15D you'll get a diving time somewhere between the 36h and the 72h, depending on 1) actual crew size, 2) amount of repairs, 3) Silent-Running.

This current CO2-modelling is still far away from being realistic, but much better than the stock SH3 behaviour, in which CO2-production is always constant.

h.sie

[ryanwigginton]

Great work!

Just when I thought I'd played this game to death, your important revisions come along.

[Magic1111]

Originally Posted by h.sie

So in V15D

h.sie

Hi h.sie !

That sounds all good !!!

When you're planning to upload the V15D?

Best regards,
Magic

[h.sie]

Originally Posted by Magic1111

Hi h.sie !
When you're planning to upload the V15D?

I think during the next few days. Some long-time stability tests and it's done. Unfortunately my assembler routines for exact measurement of repair-work grew and grew and I feared to lose overview. So I had to make the calculation more simple as follows:

Crew number is still considered for CO2-production and diving-time.

There are now 3 states of the UBoat/Crew which influence CO2 production by a certain factor:

1) Silent Running / Crew rests: Factor 0,7
2) Normal work: Factor 1,0 (default state)
3) Repairs (and indirectly "being hunted", which is correlated with repairs): Factor: 1,5

Different from V15C, it is no more possible to cheat the CO2-calculation.

Until now I was not able to find out how the in-game states (Hunting, being hunted and so on) are coded and stored, so we have to be content with the current model described above.

Maybe in the future I could also consider these in-game states, provided I find them???

And maybe I'll add a random-generated very high CO2-production factor (when batteries damaged heavily and compartment flooded) in order to model battery damage / chloride???

But now I am tired of CO2!!!
h.sie

[SquareSteelBar]

Originally Posted by h.sie

...But now I am tired of CO2!!!...

That's very dangerous...you've to be more careful !!!

[h.sie]

@SSB:

Regarding compressed air: compressed air is stored in-game as a floating-point value between 205 (full) and 0 (empty). After a first glimpse at the algorithm that is responsible for discharging the compressed air, I think the devs wanted to model an discharge rate proportional to the pressure. high pressure - high discharge rate. low pressure - low rate. unfortunately, at low pressures the discharge rate converges to zero, so that the compressed air never gets completely empty. it does not go lower than a value of 10. but anyhow, you can still blow ballast in these situations, but without sinking pressure. that is a bug.

since changing the discharge algorithm isn't trivial, I first tried an easy fix. I simply disabled the blow-ballast button when a pressure value of 10 is reached. Works fine. now you can get out of compressed air, that means, you can blow ballast about 5 or 6 times, but then the button is disabled. unfortunately, the "E" key still works - it is not logically coupled to the blow ballast button. another bug.

[makman94]

H.Sie....you are reminding me a whisky....it says....''KEEP WALKING'' !!

Seriously, the most awesome and interesting thread of the last months (say year or it is too much ? )

this modding is just at the start....but i am sure that amazing things will come out...really amazing