New 'forgiving' fatigue model

[Kafka BC]

Originally Posted by sub_optimal

Back on topic, though, does anybody know an easy way to access NGYM's fatigue model?

I think all you need is the basic.cfg file. I've extracted it uploaded it to Mediafire for you here. It is from the NYGM3.3B patch from a year ago. The 3.3C patch didn't have it, So I think its the latest one. If something more is needed, maybe somebody in the know can get it for you.

...the cleanest lobster boats on Earth.

Living in a fishing village (Eastern Passage) that is now part of Halifax, I can tell you that is no mean feat.

Since then I've been studying as a naval historian, and have spent some time academically observing (and occasionally teaching) militaries of various Western countries.

Then, perhaps you know of this character. I've followed him since the '70s when he used to publish my eagerly awaited Strategy and Tactics magazine and great wargames. He is now a Military Analyst and lecturer, and on occasion appears on CNN.

...and can show me up for the fraud that I am.

You probably have more creds than a lot of them.

And here is my shameless attempt at Naval History education.

- Kafka BC

[sub_optimal]

Thanks Kafka, that's great.

These values seem very interesting, and quite different from what I would have thought of myself, although they've clearly been thought through with great depth. For example, some of the SpecificFactor values seem really... well specific: For compartment 5, the value is 0.027225. I'd really like to understand where that number comes from. Also, unless I'm reading this wrong, there's no weather modification whatsoever, which seems strange.

Any comments?

[Stiebler]

@Sub_Optimal:

Hitman is correct, Observer was a real-life American ex-submariner. He departed from this forum about three years ago.

Observer spent a very long time with mathematical modelling/balancing of the various compartments for fatigue - and also for morale - which is why some of the figures look so precise.

I found after much experience that the morale figures were crippling any U-boat that went on an extremely long cruise, such as an IXd2 to the Indian Ocean and back to France. Similarly, very long schnorchel patrols in 1945 during stormy weather around the British coast were crippling the engine-room crew. For this reason, I toned down the effects of morale a lot, but retained the fatigue components.

I have no knowledge of the effect of weather patterns, I don't remember seeing any separate crew fatigue components for weather that could be adjusted.

Good luck, anyway,
Stiebler.

[Volk2]

What sections in basic.cfg are responsible for fatigue model? I'm playing from a long time with no fatigue, but maybe NYGM model would be good. I wish to try to copy it to GWX basic.cfg. Certainly its the sections [CREW_0]-[CREW_8]. What else? [COMPARTMENT]?

[sub_optimal]

Originally Posted by Stiebler

Observer spent a very long time with mathematical modelling/balancing of the various compartments for fatigue - and also for morale - which is why some of the figures look so precise

Very intersting. Does anybody know where morale is configured?

Originally Posted by Stiebler

I found after much experience that the morale figures were crippling any U-boat that went on an extremely long cruise, such as an IXd2 to the Indian Ocean and back to France.

That's an excelent point. Perhaps one sollution would be to have several different fatigue models to balance realism and playability in different conditions. An 'early war' and 'late war' war model might be an example, or different models for different classes (which would seem to make intuative sense anyway).

Does anybody know whether a schnorcheing sub is counted as 'surfaced' or 'submerged'?

[sub_optimal]

Originally Posted by волк

What sections in basic.cfg are responsible for fatigue model? I'm playing from a long time with no fatigue, but maybe NYGM model would be good. I wish to try to copy it to GWX basic.cfg. Certainly its the sections [CREW_0]-[CREW_8]. What else? [COMPARTMENT]?

This turns out to be a little complicated. The main section is under the heading [FATIGUE_COEF], but the values there are somewhat delphic, which is why I tried to work out what they exactly mean (detailed in my opening comment to this thread). There are also, as you point out, values describing how these fatigue values effect different rank catagories under [MEDALS_CREW] and the various [CREW_x] sections.

A simpler way to modify these values seems to be to use the excelent SH3 Commander programme, which allows you to easily switch fatigue models before launching the game. If you want to take this route, I can easily post the instructions, which are in the SH3 Commander help files.

I'd like to further add that the NYGM values are (at least as far as I can work out) indeed 'forgiving', but not necesarily in the way I had in mind, in which the point would be to allow fatigue to rise and fall quickly, creating a broad range of conditions over the course of the cruise. Not that this is neccesarily what you're looking for, but I'm still working on the problem for my own, personal satisfaction. :-)

[sub_optimal]

Originally Posted by Kafka BC

And here is my shameless attempt at Naval History education.

How very appropriately shameless, given that it reflects only credit upon you!

[Volk2]

Thanks for the info. Indeed, there is a "Crew fatigue models.cfg" in SH3 commander and as I see there are only options CREW_0... and FATIGUE_COEF... to copy from basic.cfg. I wonder if NYGM model would work this way in GWX and if its better than GWX 8 hours. The RUb is also interesting, but this "combat fatigue" model is not really convincing.

And how do you see the situation, that fatigue works in game only at less than 32x time speed?

[Stiebler]

Concerning adjustment of morale:
I did this a year or two ago, but find that my extensive list of SH3 mod-change notes does not address this issue for some reason. So I regret that I can no longer help here.

The problem with schnorchelling U-boats occurs in high winds (15 m/s), when the schnorchel dips repeatedly below the waves. This causes the engine room to transfer repeatedly between the diesel and electric engine rooms, which creates high fatigue.
The best solution is NOT to alter the fatigue settings, but instead to lower the maximum extent permitted for the rise and fall of the waves.

Fatigue changes are only seen at time control = 32x or less, but I have often had a suspicion that fatigue 'catches up' with its real value after you change time control from a high level to 32x. That is, after a long patrol at 1024x, there is no effect on fatigue or morale, but after returning to 32x fatigue and morale effects seem to change more quickly than usual. I am still not sure if this a real effect or my imagination. Has anyone else noticed the same effect?

EDIT:
Just checked - it is very easy to locate the coefficients for both morale and bad weather in file data\cfg\basic.cfg. Just open it up in Notepad, and search for 'morale' and 'weather'. No other file is required, if I remember correctly.

Stiebler.

[Kafka BC]

Originally Posted by sub_optimal

How very appropriately shameless, given that it reflects only credit upon you!

Thanks. Wasn't looking for any, but was trying to give it to the boys of the Royal Canadian Navy Volunteer Reserve, the Naval Reservists, and the good officers that commanded them.

[LGN1]

Hi,

a while ago I made a long-term fatigue model for myself. From reading old threads (thanks to all who investigated this earlier!), experimenting,... I found out the following things:

In basic.cfg every rank has this section

[CREW_0] ;SEAMAN
MoraleMin=0.30 <--- value below which morale cannot drop (see comment below)
MoraleMax=0.60 <--- value above which morale cannot rise
MoraleStep=0.05 <--- amount by which morale is increased in the crew comp. every 10 min.
FatigueMin=0 <--- value below which fatigue cannot drop
FatigueMax=0.80 <--- value above which fatigue cannot rise
FatigueStep=0.031 <--- amount by which fatigue is reduced in the crew comp. every 10 min.
CoefMorale=0.4 <--- see formula below
CoefFatigue=0.21 <--- see formula below
QualEffect=1 <--- see formula below
Hp=10
Wounded=-0.02 <--- amount by which morale is reduced if someone is wounded in the compartment
Dead=-0.05 <--- amount by which morale is reduced if someone dies in the compartment
SunkShips=0.1 <--- amount by which morale increases if a ship is sunk
TorpedoHit=0.05 <--- broken, does nothing
Experience=0

Now the question is how do these values influence the efficiency in a compartment. The answer is (assuming all men are healthy, never tested wounded men):

Efficiency = (M*(1-c_m)+c_m)*(F*(c_f-1)+1)

M: morale of crew member
c_m: CoefMorale in basic.cfg
F: fatigue of crew member
c_f: CoefFatigue in basic.cfg

The total efficiency is the sum of all the efficiencies of the crew members in the compartment. For petty officers the efficiency is multiplied with the QualEffect factor from basic.cfg (if they have the qualification corresponding to the compartment). For officers the total efficiency in the compartment is multiplied by QualEffect. Thus, having an officer with the right qualification is most important.

Ok, now we need to know how efficiency influences the different tasks. This is where my knowledge is limited. Here is what I think. We have several tasks: Repairs, torpedo reload, steering commands, sensor work.

Repair times: Repair time is given by (this is without a repair team, i.e. only the crew in the compartment is doing the repairs. I think you need at least two men in the compartment.)

T = Tfix(depends on damage amount)*Interval3/Eff,
Tfix: hardcoded 'basic' time
Eff: efficiency in compartment (formula above)
Interval3: specified for each compartment in basic.cfg
Interval1 and Interval2 seem to have no influence


Torpedo reload time:
T = Tfix*factor
Tfix: 'basic' time in sim file
factor=1 for efficiency in compartment >= Interval3
factor=3/2 for Interval2 < efficiency in compartment < Interval3
factor=3 for Interval1 < efficiency in compartment < Interval2
infinity for efficiency in compartment < Interval1 (i.e., no reloading)

I did not notice any influence of efficiency on giving commands. I also do not know for sure how efficiency influences the sensors. It seems that the Interval values have no influence. So, I guess a continuous scale is used (as with repair times, not as in the case of torpedo reloading). Engines will not work if efficiency in compartment < Interval1. It seems the other Intervals have no influence.

Finally, basic.cfg contains the [FATIGUE_COEF] values for each compartment:

RegularFactor50=0.001 <--- amount by which fatigue is decreased when a crew member is moved into the compartment
RegularFactor51=0.001 <--- unknown. No idea.
SpecificFactor50=0.045 <--- amount by which fatigue decreases every 10 min. when the sub is surfaced
SpecificFactor51=0.045 <--- amount by which fatigue decreases every 10 min. when the sub is submerged
BadWeather5=0.02 <--- amount by which fatigue is decreased in bad weather (I'm not sure whether it's added to RegularFactor or used without it. But this can be easily checked.)

That's how I think it works. If someone has more information or thinks something is wrong I would love to know it. Especially, it would be great to know more about how sensors are influenced by the efficiency.

Cheers, LGN1

PS: One has to be careful when playing with MoraleStep. If you make it negative the morale can become negative because it is not stopped by MoraleMin. IIRC, you can also get negative fatigue if you choose some strange settings.

[sub_optimal]

That's fantastic, LGN1. That saves a ton of work.

Do you mind if I ask a few questions?

Firstly, are you sure that "Efficiency = (M*(1-c_m)+c_m)*(F*(c_f-1)+1)" shouldn't read "Efficiency = ((M*1-c_m)+c_m)*((F*c_f-1)+1)"?

Second, what leads you to the conclusion that SpecificFactor is envoked when crews are moved from compartment to compartment? I've been sitting here moving my crews around, their fatigue doesn't seem to go up. Furthermore, in the default settings, the SpecificFactor for compartments 3 and 4 (diesel and electric engines respectively) are mirrored for the x0 and x1 conditions (ie 30 is 0.4, 31 0.1; 40 is 0.1, 41 0.4). This would seem to fit the model that these factors should reflect the submerged and surfaced postures of the submarine, as the electrics are only used submerged and the diesels surfaced. Thus:

30 - diesel surfaced - in use = 0.4 (high)
31 - diesel submerged - not in use = 0.1 (low)
40 - electric sufaced - not in use = 0.1 (low)
41 - electric submerged - in use = 0.4 (high)

[divingbluefrog]

LGN1, I see you have fine tuned your model since your discovery of the formula that changes my life... Thank you again for this fantastic breakthrough.
I can give you my two cents :
Regular - noncombat situation
Specific – Combat situation
This is not my own conclusion, I found this somewhere in the inmost depths of this forum, and copied it to my own Tips 'd Tricks file.

For the Tfix in the repair time formula, I incline toward a relation with the Hit Point of the equipment/compartment, even with the amount of hp lost or remaining...
Could you confirm the formula is T = Tfix/(Eff*Interval3) ?

By the way, do you have any idea about the Resilience factor? All I know is that it is influenced by crewman HP, the more HP, the less resilience. Should have an influence on Fatigue or Efficiency...

[LGN1]

Originally Posted by sub_optimal

That's fantastic, LGN1. That saves a ton of work.

Do you mind if I ask a few questions?

Firstly, are you sure that "Efficiency = (M*(1-c_m)+c_m)*(F*(c_f-1)+1)" shouldn't read "Efficiency = ((M*1-c_m)+c_m)*((F*c_f-1)+1)"?

Second, what leads you to the conclusion that SpecificFactor is envoked when crews are moved from compartment to compartment? I've been sitting here moving my crews around, their fatigue doesn't seem to go up. Furthermore, in the default settings, the SpecificFactor for compartments 3 and 4 (diesel and electric engines respectively) are mirrored for the x0 and x1 conditions (ie 30 is 0.4, 31 0.1; 40 is 0.1, 41 0.4). This would seem to fit the model that these factors should reflect the submerged and surfaced postures of the submarine, as the electrics are only used submerged and the diesels surfaced. Thus:

30 - diesel surfaced - in use = 0.4 (high)
31 - diesel submerged - not in use = 0.1 (low)
40 - electric sufaced - not in use = 0.1 (low)
41 - electric submerged - in use = 0.4 (high)

Hi,

I am pretty sure about the formula. Your version can be simplified to

Efficiency = (M-c_m+c_m)*(F*c_f-1+1) = M*F*c_f

which means for c_f=0 you have no efficiency. Also c_m has no influence. But in-game you can choose a purely morale-dependent model by choosing c_f=1. By choosing c_m=1 you get a pure fatigue model.

Sorry, about the regular and specific factors. I confused the two. It's the other way round (corrected it above). Look in this old thread:

http://www.subsim.com/radioroom//showthread.php?t=85656

Be careful, the formula given there for the efficiency is wrong, IMHO.

Originally Posted by divingbluefrog

LGN1, I see you have fine tuned your model since your discovery of the formula that changes my life... Thank you again for this fantastic breakthrough.
I can give you my two cents :
Regular - noncombat situation
Specific – Combat situation
This is not my own conclusion, I found this somewhere in the inmost depths of this forum, and copied it to my own Tips 'd Tricks file.

For the Tfix in the repair time formula, I incline toward a relation with the Hit Point of the equipment/compartment, even with the amount of hp lost or remaining...
Could you confirm the formula is T = Tfix/(Eff*Interval3) ?

By the way, do you have any idea about the Resilience factor? All I know is that it is influenced by crewman HP, the more HP, the less resilience. Should have an influence on Fatigue or Efficiency...

About the specific and regular factors, see the thread I linked above. It's quite old, but I don't know any definite conclusion about the regular factor.

You are right about the repair times. I am pretty sure that the status of the damaged item (via hitpoints) plays a role. However, how hitpoints exactly enter is not known to me. The formula is just about how the efficiency enters (what happens if you send more men to repair). A more correct version would be

T = Tfix(damage/hitpoints)*Interval3/Eff

,i.e., Tfix is an unknown function of the damage/hitpoints. If you increase the Interval3 value, the time will also increase.

Which Resilience factor do you mean? The colored bar for each crew member? I think it's just 1-fatigue. I don't think that hitpoints influence the efficiency. At least I can predict the efficiency (green bar above the compartment) without using any hitpoint information.

Cheers, LGN1

[sub_optimal]

Originally Posted by LGN1

I am pretty sure about the formula. Your version can be simplified to

Efficiency = (M-c_m+c_m)*(F*c_f-1+1) = M*F*c_f

Wait... maybe I'm going crazy, but I thought that you had to do the addition and subtraction first, and then the multiplication and division.