Hi h.sie,

thanks for your post. I have to examine your equations closer before commenting further on them.

Just a bit more food for thoughts concerning weather effects:

Here you can find a table that provides a rough connection between wave height and wind:

http://en.wikipedia.org/wiki/Beaufort_scale

As you can see at 8m/s wind speed you get a wave height of 1-2m. Furthermore:

Wind speed: Wave height:

8 - 11m/s ............ 2-3m
11-14m/s ............ 3-4m
14-17m/s ............ 4-5.5m

Wave turbulences will go deeper, but I guess they are quickly damped (maybe on a length of 1m :06:). Torpedoes set at a depth shallower than the above values should have a dud probability of close to 100% (surface runners). For torpedo depths below the values above + 1m I would set the probability close to 0% (the wind speed related probability).

Regards, LGN1

Approved http://www8.pic-upload.de/06.11.11/xmahaowjr661.jpg

:salute:

@LGN1: Thanks for that table, but comparison with real windspeeds is difficult.

We have to consider, that our in-game windspeed of 15m/s is the maximum possible (worst case), while in reality, 15m/s is moderate.

Thus, in my opinion, the influence of in-game windspeed must be boosted compared to real weather.


So you prefer a "digital" solution:

If torpedo breaks water surface => Dud rate factor = 1,0
If not => Dud rate factor = 0,0

----

I cannot believe that 1m below a rough water surface the turbulences are completely damped.

I agree, h.sie.

But if you want to interpret 15m/s wind in-game with x m/s wind in real-life the question arises whether x m/s in nature are as common as 15m/s in SH3.

Breaking the surface is certainly not the only factor. Turbulences below the surface lead to forces on the torpedo which the steering mechanism has to correct. The higher the forces the more problematic it becomes. However, I think these turbulences are damped quite fast.

What might be a good idea is not to have a true digital solution / step function, but a 'Fermi distribution':

p = 1 / ( exp( (TD-f(WS)) / width) + 1 )

with TD: torpedo depth, f(WS) a function of the windspeed (according to the numbers above a linear function works quite well), and width a parameter that determines how fast the probability drops from 100% to 0% (simulating the turbulence layer).

Edit: The turbulences are certainly not completely damped after 1m, however, a torpedo is also quite heavy.

I'll check whether that could be approximated by the Boltzmann-distribution, in order to make programming easier.

Or, much easier....a linear function. But Boltzmann sounds more interesting.

You should be able to approximate it with a Boltzmann-distribution (for high temperatures the Fermi-distribution is well approximated by the Boltzmann-distribution), but in our case you will have a problem with the plateau for torpedo depths shallower than the wave height. This you might circumvent by using a piece-wise defined function, e.g.,

100% for TD >= WH
Boltzmann-distribution for TD < WH

I was 50% joking / 50% serious. I think we are creating a very coarse model with many unknown factors. If we would now use sophisticated formulae this would be inappropriate since it suggests an accuracy that is not really given. Piecewise-Linear functions are surely sufficient.

My formula in post #2611 was simple and intentionally not physically-based, because that is too sophisticated IMHO. It was my mathematical description of the consensus between Hitman's and your (LGN1) suggestions.

But I'm open for better formula.....especially for the torpedo-depth-term

I agree completely (but I thought you prefer a smooth, analytical function). I think the best would be:

p = 1 ...................................for TD < 0.3*WS
P = 1-(TD - 0.3*WS)/width .....for 0.3*WS < TD < 0.3*WS+width
p = 0 ...................................for TD > 0.3*WS+width

just a short comment about gameplay without any further fixes on torpedoe duds:

I'm actually playing my 6th -whole war- career, now I'm in june '44. I'm always using manual targeting and trying to hit targets at an angle of nearly 90 degrees, fire distance between 500 - 2000m. I never had a patrol without torpedoe duds, worst attack was in dec 1939, when I attacked a "Renown" with a type II U-boat and none of my 3 launched torpedoes exploded.

My experiences with duds in stock sh3 so far:
9/39-6/40 - 20-30% - best 16% - worst 66%
7/40-6/42 - 15-25% - best 14% - worst 43%
later - 10-20% - best 9% - worst 26%

So, who means the numbers of duds are too small in stock sh3? In most comments I read here it's pointed out as a fact that nearly every shot is a hit - I cannot agree with this.

I think gameplay is alreday well balanced, especially when you use manual targeting and it's possible that some of your shots miss their targets . It's already hard to hit under bad weather conditions (difficult to set the correct values for the TDC depending of limits of optics and a great chance that torpdoes hit the targets over waterline or running too deep) and please remember that not every hitted unit is sinking later or loosing speed.

What I don't like is that it's not important which pistol you use, especially when the Kriegsmarine had the most problems with magnetic pistols during invasion of Norway. So I would like to have a fix with only small changes on the numbers of duds, but great changes on pistol type your are using.

Greetz

Hi tekai,

thanks for your input :up: I guess you are always using the magnetic pistol, aren't you? Because if you use the contact pistol and set the torpedo depth to 2 or 3 meters you should not see any duds in SH3 (except of misses because of manual targeting, but that does not count as a torpedo failure).

My test suggests a failure rate of 6% for a range <1000m and a failure rate of 30% for a range of 1000-2000m early in the war. This is indeed quite high, however, I think it's effictively still too low because especially early in the war you can shoot from quite short distances and therefore obtain a failure rate around 6% (or 0% if are using the contact pistol). BTW, NYGM improves the whole situation a bit by forcing you to use the magnetic pistol :up:

Cheers, LGN1

PS: On a related note, I don't remember ever reading a thread in the SH3 section of subsim about too many torpedo failures in SH3. However, if you go to the SH4 section, you will see quite a few :D And if you read the historical accounts from the US skippers and the German commanders the situation was quite similar back then. So, I think that either SH4 or SH3 must be off quite a bit.

@tekai:

Thanks for that comment. It shows that you are satisfied with the current situation, but I know at least 4 guys who would like more duds, and for those I program that mod. It is not intended to make all happy. Those who are happy with the current situation, should disable this mod in the OptionsSelector.

Regarding torpedo pistol: Unfortunately, until now I was not able to analyse the individual pistol setting of each torpedo, thus, ATM I cannot fulfil your wish, even if it makes sense to me.

NYGM locks the pistol type via the dials.cfg file to magnetic. However, some people don't like that because in this case you have to use the magnetic pistol even when it was banned officially (and you can make 'magnetic' shots from bad angles although it was not possible historically).

Hi LGN1,

I agree, a term describing influence of torpedo depth as follows

p = 1 ...................................for TD < A*WS
P = 1-(TD - A*WS)/width .....for A*WS < TD < A*WS+width
p = 0 ...................................for TD > A *WS+width

is more realistic than my initial linear term.

Interval 1: Torpedo breaks water surface. 100% dud.

Interval 2: Torpedo not breaking surface, but affected by turbulences, depending on torpedo depth.

Interval 3: Torpedo too deep. Not affected. 0% dud.

Now we need to talk about the parameters A and width. Maybe A=0,3 is a good one, but width?

Hmm, if magnetic pistol already has that high dud rates, it will be overkill, if we program additional duds. Seems we need to differ between the pistols, but ATM I cannot find where the pistol setting is stored for each torpedo.

@hsie
It was not intended to critizise your work in a bad way, I only wanted to give a few numbers from my experiences with stock settings of sh3, as explained in my first sentence.
So please go on with your mod!
BTW, there's no reason why I should'nt test it, when it's ready.
Regarding torpedoe pistols: I gave you no order to change something in the way I would prefer it, I only pointed out what could be improved some day by anyone.

This is exactly what I meant in my last post.

@LGN1
I'm using contact pistol in about 80-85% of my attacks , but I have duds enabled in my settings - and as you can see - these are the results I got. Torpedoe depth is depending of the ship I want to attack, the deeper the hit the greater the chance the hidden unit begins to sink or looses speed. So normally I set the depth at about 5-7m, for great tankers I take 7-9m.

Greetz