Too many torpedo duds

[host249]

Hi, I fired 5 torpedos against a merchant and obtained 5 hits, but only 2 exploded, the 3 duds were with contact exploder. Next merchant 4 hits, 2 duds, contact exploder. Almost 90 degree angle. Well, the german navy had a torpedo crisis early in the war but the difficulties were more or less solved in 1941. Despite that its really not fun playing with a more than 50% failure quote...this has to be fixed soon please.

[gerald2]

Originally posted by McDewgle:

You want to fix historically accurate torpedoes?

yes i do ,in real there was about 50% duds on last patrol i got 100% duds so really those 50% would make difference for me ...they screwed something at last update.

[boris.glevrk]

Originally posted by Tanker D18:

I actually dig the increased dud rate. There was a reason many skippers (especially in the early war) would launch all 4 bow tubes at a single ship.

The true reason was the low hit chance, not high dud rate.
Even the American Mk 14 has "only" like 60-80% dud rate while some of us are experiencing 100%.

Originally posted by McDewgle:

DudChance = 0.3

OK so now you switched to 30% again. What about the 19% you used to claim?

For 30% dud to demonstrate 3 dud and then 1 hit:
0.3*0.3*0.3*0.7=0.0189, less than 2%.
Yet there's 3 of us now experiencing such dud rate or worse.

[Dewgle]

Originally posted by boris.glevrk:

Originally posted by Tanker D18:

I actually dig the increased dud rate. There was a reason many skippers (especially in the early war) would launch all 4 bow tubes at a single ship.

The true reason was the low hit chance, not high dud rate.

I think Wolf310ii, PafunaMT, Stosstrupp, Prophet or any of number the guys on the Crush Depth team would disagree with that.

Originally posted by PafunaMT:

~Wooden Torpedoes - The German Crisis, article, Uboat Aces, http://www.uboataces.com/articles-wooden-torpedoes.shtml

Originally posted by boris.glevrk:

Originally posted by McDewgle:

DudChance = 0.3

OK so now you switched to 30% again. What about the 19% you used to claim?

You must've missed this part huh

Originally posted by McDewgle:

G7e Torpedo T2 - Pi1<-<-<-<-<----
Range1 = 3000, Range2 = 3000, Range3 = 3000, Speed1 = 15.5555, Speed2 = 15.5555, Speed3 = 15.5555, DudChance = 0.3, Damage = 4.9, CrewDamage = 0.8, DamageRadius = 40, DamageEffectsRadius = 7.0, DamageEffectsIntensity = 1.0, MaintenanceCooldown = 1200, MinPistolActivationAngle = 40, MaxPistolActivationAngle = 69, MagneticExplosionOnArm = 0.1, MagneticExplosionAfterArm = 0.005, MagneticExplosionFail = 0.1

This is what I used to 'claim'
T1 = DudChance = 0.19
T2 = DudChance = 0.3
T3 = DudChance = 0.16
T5 = DudChance = 0.14

Which you can still find out for yourself in the datasheets. Which I highly encourage you go do instead of just trying to start something.

Originally posted by boris.glevrk:

For 30% dud to demonstrate 3 dud and then 1 hit:
0.3*0.3*0.3*0.7=0.0189, less than 2%.
Yet there's 3 of us now experiencing such dud rate or worse.

I don't.. What? That's not how probability works. Here hope this helps
https://youtu.be/2tKC55XfYMA

[boris.glevrk]

Originally posted by McDewgle:

I don't.. What?

Are you really that intellectually challenged to understand that?
You kept saying that it's normal "for a 50% coin to show 4 heads in a row", so I calculated for you the chance of a 30% dud torp to experience 3 duds and one good detonation.

And the result is clearly conflicting with what we are experiencing.
I now have at least three guys experiencing 75% or higher dud chances in "small sample", and according to that calculation, even in small sample it's close to impossible for all three of us to see such an abnormal dud rate. All the calculations show that IF the dud rate is working as intended, the chance for us to experience this is very slim, let alone to have 3 of us experiencing this.
Therefore we reach the conclusion that the "actual" dud chance for impact pistol is not as game file shows, and is too high.

[boris.glevrk]

Originally posted by McDewgle:

That's not how probability works. Here hope this helps

You need to dig your brain out and reinstall stuff.

Known 1: We have a 30% dud torp.
Known 2: Torpedo detonator only has two result: dud or detonation.

Question: What is the chance for the first 3 torps to be dud and the fourth to be good detonation?
Ans: 0.3*0.3*0.3*0.7 = 0.0189, or 1.89%.
(first torp dud rate)*(second torp dud rate)*(third torp dud rate)*(fourth torp detonation rate).
This IS how probability works.

And now I have two other guys with even worse ACTUAL combat dud rates, all with 90 degrees impact angle and warmed (though according to you, warming doesn't affect it)

Therefore we reach the conclusion that the claimed dud rate doesn't match in-game results, and thus we are calling it a bug.

[Vlad Tepes]

This thread has made me wonder, are there like 5000 torpedoes (not exploded) sitting down at the bottom of the English Channel?

[link]

Originally posted by boris.glevrk:

Originally posted by McDewgle:

That's not how probability works. Here hope this helps

Known 1: We have a 30% dud torp.
Known 2: Torpedo detonator only has two result: dud or detonation.

Question: What is the chance for the first 3 torps to be dud and the fourth to be good detonation?
Ans: 0.3*0.3*0.3*0.7 = 0.0189, or 1.89%.
(first torp dud rate)*(second torp dud rate)*(third torp dud rate)*(fourth torp detonation rate).
This IS how probability works.

And now I have two other guys with even worse ACTUAL combat dud rates, all with 90 degrees impact angle and warmed (though according to you, warming doesn't affect it)

Therefore we reach the conclusion that the claimed dud rate doesn't match in-game results, and thus we are calling it a bug.

and correct depth right? didn't read the total chain.
the torpedo's didn't pass under? < Assuming this and all of the factors you listed are correct and I assume they are. then yes there is likely an issue.

My suggestion is try and record a video of the 4 torpedo in an ideal situation not exploding. (ensuring to capture all the listed variables: depth, angle, detonator type, hell sea state? ) and post a link in the Patch discussion so the devs can evaluate the situation. Might even throw in a output log just to be safe. I would laugh (not at anyone) if the exceptions that are occurring with torpedos in P7 are the cause of this - i have seen a few.

Secondly (this is directed to the entire thread), lets not start insulting other people that are trying to help (not implying anyone is innocent here either)
And before someone goes and says "he started it" just dont pass insults back and forth, this just hurts this community as a whole, and does literally nothing to the person you are talking to.
If needed, report the comment the first time, and remain professional.

[Dewgle]

So I'm not familiar with this formula of probability, sue me ¯\_(ツ)_/¯
(0.3*0.3*0.3*0.7 = 0.0189, or 1.89%.)

But I know the datasheets, and you can too if you have a program that read excel files. I spent most of tonight talking to Salamander, creator of the magnetic detonators mod and realistic T5 mod and with his help we're going to try and make a guide to help explain the new torpedo failures.

[Vlad Tepes]

That's just the probability of EXACTLY getting 3 duds and a good shot when firing 4. It does not change the fact that each shot still has a 30% chance of being a dud. The problem is that dud statistics are taken from thousands of attempts, and the smaller sample size totally messes up the results (sure the calculation is the same). After you've fired 1000 torpedoes, if there are no bugs in the game, you should see approximately 300 duds. .. ish :)

[boris.glevrk]

Originally posted by McDewgle:

But I know the datasheets, and you can too if you have a program that read excel files.

If we know that the game doesn't work as how datasheets claim, editing datasheets will not work.
My guess is this:
"best detonation happens within X-Y degrees from perpendicular"
ideal explination: Best detonation happens when the impact angle is BELOW a certain angle threshold, and the THRESHOLD ranges between X and Y.
Actual explaination: Best detonation happens when the impact angle is between X and Y.

In other words, assume the best detonation angle threshold is between 60 and 45 degrees, what we expect is that detonation rate begin to decline between 60 and 45 (thus 90 is always optimal angle), while the system is somehow coded as the detonation only happens reliably between 60 and 45 (thus 90 is outside of this range and thus result in lots of duds).

Originally posted by Voltaj Tepes:

The problem is that dud statistics are taken from thousands of attempts, and the smaller sample size totally messes up the results (sure the calculation is the same).

The problem is that the result has well exceeded what can be explained as small sample size.
Having 3 straight duds on a 30% torp would be .3*.3*.3=2.7% chance. Already very slim, and here we have 5-8 straight dud reports, which lowers the probability (under the assumption that the dud rates are working as intended to sub-percent levels.

What you don't understand is that if I shoot 1000 torpedoes I will not get 300 duds. What is happening in B129P7 is that we will get 800 duds.

[wolf310ii]

Originally posted by boris.glevrk:

The problem is that the result has well exceeded what can be explained as small sample size.
Having 3 straight duds on a 30% torp would be .3*.3*.3=2.7% chance.

2.7% is still a chance. Just because its a very slim chance dosent mean it cant happen.
The chances to win the lottery is by far smaller and there are people who won millions twice.

[Dark Sheep]

I think what we overlook in a discussion like this is, what ever the dud percentage is, the distribution of faulty vs good torpedoes is random.

Think of it this way: on average if you flip coins on average you'll get 50/50 heads or tails. But that doesn't mean if you toss heads, the next toss must be tails. You can get sequences of only tails or only heads wich, in aggregate, will average out to be 50/50 heads/tails.

Same with the dud torpedoes. It might be 19% duds overall. but you might get unlucky and have a boatload of dud torpedoes. You can't know in advance.

I think the question is: do we want a more fun game, or a more realistic game. I don't think there's a right or wrong answer to that. I can see the appeal of both.

[nihilcat]

Hi,

I want to clarify a few things about how duds work in the new system in the case of the impact (AZ) setting.

1. On the side of the data: DudChance parameter is no longer used in the game at all. It may have a comeback in some form in the future, hence why we left it for the time being :)

2. Please keep in mind that the impact angle is computed relatively to the hit surface in 3D. That means that even, if you fire the torpedo at the perfect 90 degrees relative to the ship's side looking at it in 2D on the map, but the torpedo runs deep enough to hit the hull at a low point where the hulls usually have a steep curvature, the actual hit angle may be far from perfect and the detonation will not be triggered.

For this reason the torpedoes with AZ setting are the most reliable at the shallowest settings possible, which have a drawback of inflicting less damage.

3. The revised Pi1 pistols were not intended for research in the scenarios currently present in the game. It was a result of mistake and we corrected that in Preview 8. This should noticeably decrease the dud rates right after starting the campaign.

4. Since B129 remains on the unstable branch it's still very well possible that there are also some unwanted effects and bugs accounted here that we will be doing our best to eliminate soon.

[Dewgle]

Thank you for the info, Nihilcat :)

[bartbandy]

The beauty of this game is that it could range from arcade to full out simulator if it wanted to be. Regardless of the fixed torpedo failure rate you will never make everyone happy. The arcade style players who wanna blow ♥♥♥♥ up are not going to like the failures and sim folks want the realism. This could all be achieved through the AI setup so easily. As you increase game difficulty, the RNG goes to more historic levels. Make the difficulty slider mean more. I am sure the devs have thought about this.... why not implement?