![]() |
Quote:
1.) standardized test will hopefully give us a controlled testing enviorment. 2.) speeds up testing. Running a patrol takes time. Sooner this gets hashed out the sooner we can start enjoying the game. Although if you could come up with a 2nd version that has say, 3 or 4 escorts to go along with the single escort mission, that would help greatly. Specficially id like to test single DD detection, and then also under a "normal "convoy situation. EDIT: BTW whats the crew rating? 3, 4, or is it remarked out? Id like to experiment with that. In specfic, a crew rating of 4 vs a remarked out crew rating. |
Thought I might throw in my 2 cents, I generally find that dd's don't detect me unless I surface or something! however once they have detected me they hound me endlessly, dropping dc's with pinpoint accuracy, finally after 30mins or more I give up!! because they sure don't. :-?
This is in a carreer, I don't play the single missions, however I do play LAN coop game and even on the easiest settings (1939) they are the same - uber! if anything 1943 gameplay is much easier & escaping dd's easier :dead: Keep up the good work modders & don't give in - please! :lol: |
Posting this now while its still fresh in my mind:
Using this mission: http://rapidshare.de/files/8027236/s...-test.rar.html I edited the starting point back so theres like 11K meters distance between the DD and the player sub to give me time to dive. Parameters of my testing so far: Sim.cfg: Hydrophone] Detection time=0.0 ;[s] Sensitivity=0.03 ;(0..1) Height factor=0 ;[m] Waves factor=0.0 ;[>=0] Speed factor=15 ;[kt] Noise factor=0.0 ;[>=0] AI_sensors.dat: I took all default passive sonar values and multiplied it by 3. Ballpark figure, puts most of them in the 20 to 30K range. AI_Hydrophones set to 30K, later reduced to default value of 6K. Observations: Test 1: Escort was at crew rating 4. Instant i hit 10 meters i was detected at ahead slow, doing about 4 knot. DD detected me before the sonar operator acknowledged he was there (about 2 seconds later). About 5 seconds later i got him targeted in periscope at about 9800 meters. THATS FAR. Test 2: I dumbed down the escort to crew rating 1. Results were the same as test 1. After this i re-edited the AI_sensor.dat and changed the AI hydrophone from 30K back to its default value of 6K Ran test 3: escort of crew rating 1 (unchanged from previous test) Same results, detected almost immediatly, under same conditions, only difference is the AI_hydrophone was set to 6K while all specific passives were tripled in value. Next, going to change sim.cfg hydrophones back to default values and run another 2 tests. If results are the same then going to lower hydrophone max range from Default *3, to default *2. More later. EDIT: More testing: Parameters used: Sim.cfg: -vanilla default AI_sensors.dat - AI_hydrophones = default of 6K - All passives = default *3 Test 4: DD crew rating =4 Was not detected on dive, ran at ahead slow at 50 meters. Recieved no detection notice, DD just starting pinging when within active sonar range without any warning. Test 5: DD crew rating =1 Similar results, not detected on dive, ran ahead at 50 meters at ahead slow. However this time i did recieve a detection notice, im guessing at about 1500-2K meters, Just before he starting pinging. In neither test did i try silent running, i just ran at ahead slow at 50 meters running head on into the DD to see if/when hed detect me. Clearly the settings in the sim.cfg play a MAJOR role. My thought now is to tone down the passive sonars to default, or default*2 and fine tune the CFG. I should also note that all tests were with the above linked mission, calm seas, and clear weather, seems like ideal sonar conditions. I think ill run another test before chaning anything at ahead 1/3rd and see when im detected. |
Hi Ducimus
I am sorry but I have appointment in a bit so cannot do any testing this morning. Therefore my next comments are speculation only. Lets for 1 second say my Hunch is right and that crew rating does in fact affect sensitivity and sensitivity then goes on to affect the rest i.e. noise, detection time etc etc. If you leave waves to 0, noise to 0, and detection time to 0 also leave all passives to the same as they was default. I can imagine you will get similar results with any crew rating. I can almost imagine that you will start to get passive contact detections closer to what the real levels were set as in the ai_sensors. This however means that it would be like this no matter what weather etc and in my book, it is just as much a game killer as when we started. I do think we will end up with one crew setting, and some minor tweaks in the sim.cfg as well as some tweaks at ranges in the ai_sensors to find a common ground. I personally do not want to remove the noise and wave settings at the moment just because I am getting results. I would rather try to find what ranges and crewratings fit best into the current settings in the sim.cfg I am after finding what crew rating works the closest to the settings after I nerf them just a tiny bit. It may be that crew ratings will work at 4, 3, 2, 1, or 0. It may be better to remove them all together but CB has already warned of a possible side effect. When I get back, I am going to set the waves and noise to a much lower level but not completely disapeared i.e. 0.001 or similar. I am going to leave all crew in the game, and leave the ranges as they are. I bet any money, I start to get similar results to what the rest of you are experiencing. I do suspect that crew ratings may not only affect sensitivity but they may affect other things and it is these that I want to discover now. Anyway Good luck guys, I am off now and will be back around 1pm uk time. |
Quote:
MORE TESTS: SIM.CFG = default AI_sensor.dat -AI_hydrophones = 6k (default) -All passives = default * 3 Test1: Crew rating = 1 I ran at ahead 1/3rd at 50 meters I was detected im guessing at twice the distance i was before. In this case if it was say.... 1K meters, he picked me up at 2K. Hell he had to step on me to find me. Test 2: crew rating =4 same parameters, only this time he picked me up at about 4K im guessing. Im going to run both tests again at periscope depth so i can get a distance to target. My conclusion as so far : 1.) SIM.CFG settings for lack of a better term seem to dictate "diminishing returns" on the passive sonar. Zero it out, he has no diminishing returns, put it back and he cant hear as far, or how much he hears. 2.) crew rating seems to effect sensitivity as to how soon the escort detects you. off to test again. EDIT: OK More findings: Test, same files. Ahead 1/3rd, at periscope depth. Test 1: Crewrating =1 DD detected me at about 1100-1200 meters. Test 2: Crewrating =4 2a.) at about 4K meters i noticed he stoped steaming straight, but started zig zagging. Obvsiouly he knows im around, but im guessing hasnt pin pointed my location yet, hence no detection message.. i decided to rerun this test. 2b) DD detects me somewhere in between 3500 and 4000 meters and starts to zig zag. Im betting by tweaking the CFG files we can increase the distance at which these events occur. Obviously crew rating is a big difference.. im gonna try it one more time with a crew rating of 3 and call it quits for the night. EDIT: Damn crew rating 4 makes a BIG difference. just ran another test with crew rating 3. Test 1: at periscope depth at 1/3rd speed he detected me at 1200 meters and started evasive zig zagging. ( I should note that crew rating 1 makes no evasive manuvers where as 3 does) Test 2: same thing only at about 1500 meters. So right now im of two thoughts: A.) Use crewrating 3, and Tweak the CFG findings so he can hear a bit farther out, Hes obviously not dumb, and not uber, might be worth working with. or.. b.) use crew rating 4, and impose various limiations to make them escapable and convoys approachable. Honestly i think B might be easier then A. im not sure. Those CFG settings are very subltle, and make a marked difference. Adjust them properly and we could probably make any DD with a crew rating of 3 find us at some appropriate range. I think another trick is to make them detect us at "ahead slow" while not in silent running. Personnaly i rarely go above that speed unless im going to miss the firing point. Going 1/3rd is like trying to ring a dinner bell. So a very sublte CFG tweak is probably whats needed. EDIT: you know, i should rerun the same tests, only instead of 1/3rd, use 2/3rd or standard. Ill bet the detecton ranges will increase, probably in scale to the detection ranges at 1/3rd. |
Exactly
Also remember when we decide to have a single crew rating, all can be adjusted to fit better with game i.e. vision, Detection, loss contact time etc etc. Also note that we now have a starting ground of rules to make the sub hydrophone work better. Also note, maybe, just maybe, we can manipulate the crews night vision on sub. Also note, If we did want random crews (I am not fussed as long as they work), then we could hope that timetraveller, Jcjones or der teddy make a randomiser for the CFG file e.g. SH3 commander. All great future prospects, but now is the time to try and get the hydrophones working at intended ranges with only very minor discrepencies between the waves and noise. A big Kudos goes to first person who gets a good result that matches the range for the equipment used on the DD :rotfl: :rotfl: :rotfl: :rotfl: :rotfl: |
My final thought for the evening:
Change passives back to default or double them.. unsure at the moment, im leaning on default, depending on ranges. I think what we have to agree on, is the crew rating to use, and at what range they should be able to detect you at any given engine settign (slow, 1/3rd, 2/3rd, etc) Right now my thoguht is at 1/3rd a DD should be able to hear me at a maximum of 5K range (NOT 1200ish) . Slow and slow w/ silent runnning is a another story. Other speed settings are up in the air. But what crew rating? Im thinking 3. Primarly because he seems to be a bit more consitant from what ive seen so far, and probably more controllable. This might require more experimentation. It might be just as simple as a fraction of a percent change in the CFG files. My question now is, what variable should id start tinkering with? Waves? or noise level? Exactly does each of these variables do? EDIT: Bah, Ran another test, crew rating 3, DD picked me up at 2500 meters. Obvisouly there is other randomization going on. I guess if i want to be through id have to get a mean median and mode stastic on each crew rating.. but i dont think its neccessary, randomization variable on detection or not, the crew level does make a marked differnce. EDIT: I need to go to bed.. heh. I just restored the AI_sensor.dat file to default, and changed the sim.cfg file. The only change i made was: Noise factor=0.5 A crew rating 3 escort just picked me up at 4000 meters at 1/3rd speed. |
i see we're going to go round and round untill we vanish up our own perfectly controlled ; perfectly standardised exhaust pipe here; ESPECIALLY if we all use the same mission---i'm out of here !!
|
Quote:
Wake-up in the morning :dead: :doh: Good morning guys. I run some test las night prior to sleep. And agree with you changes in Sim.cfg works. But changes into AI-Sensors.dat, in beam angles make a lot of diference. Reducing the depth of the beams introduce diferences at medium ranges and short ranges. Changes in beam wide looks to be more important at near dog search, anyway had some utility to maintain undetected at far distances. I reach to manage my problem wich was the Ubber Buckley and Evarts american DDs. The problem seems to be now the crew rating afecting too much the performance, because we need to stablish a normal detection ranges, and crew rating looks to affect it too much. I think so and it was my impresion, a sub running at silent running, and even at ahead slow, must to be near to undetected if it is submerged and not do any stupid thing as surface, rise snokel or periscope, reloading or repairing. If not, they didnt has the need to create or invent the active sonar. Detection ranges was long but for cruise and high speeds at surface, but for submerged speeds, those detection ranges was very shorts. Whats next step ? Do we need to adjust detection ranges for average crew quality ? Or for crew rating 1 ? or crew rating 4 ? :hmm: I had not time yet to test if crew rating affect other settings as adquire time, contact lost time, precision in the attack or in precision depth. If we adjust settings to good game play with crew rating 3, a mission with crew rating 1, will be too easy. If we adjust settings ro good gameplay in a mission with crew rating 2, when play a mission with DDs with crew rating 4, they will be Ubber Dds. Whats the best game play solution ? :hmm: |
Quote:
Came here CB, we need you, i will not stop up to test the changes in every single mission, from stock, from 3rd party, and campaign. To save time, i always tes on only 4 or 5 missions, U-505, Brhman, Happy Times, Royal Flush, Biskmark..... I use most U-505 because is the only one mission giving me problems ...... i consider it is out of game play, because it has unreal historical scape and survive probabilities. back here. :up: |
Quote:
ie suppose you make a change in sim.cfg, and run a campaign game, and run into a convoy in a storm, get detected at 1k -- is that short distance due to your settins or due to storm? re going round and round -- i agree. we're retesting stuff that's already been looked at. but maybe someone will notice something that's been missed. anyway, pls don't go away. :cry: test how you like, and post your results -- whatever results you get will be appreciated by everyone here. |
Right I am back
Quote:
I had a feeling this would happen. I knew there would come a point when either our games (do not think so) playstyles or personal traits will cause a difference of opinion on this subject. I am sorry to see you getting frustrated by some of my techniques and maybe you are on the right track and I am on the wrong track, whatever. To be honest I do not care who is right and wrong as long as the problems are improved. Do you think I am stupid, egotistic or something? Its time to be honest here CB! I recall some past comments but was not sure if they were directed at myself. Anyway I am not here to get browny points with other players or modders. I am not dumb until it comes to standard deviation/maths and I am not going top appollogise for driving anyone up the wall. I say thanks for all your contribution to the subject, some of which has been fantastic and inspiring. ALL in all I think we need to aggree to disagree on the methods of testing and should you crack it, well done. No hard feelings from my side unless you are cursing me under your breath. If people think I am hindering this project, please say so. I will gladly shut up and stay of the topic to go ahead with my other work that I have been wanting to do. I don't want to be falling out and getting nasty because of a game. |
I had missing some thing or what ?
It is my point of view, some person open a topic about a speciphic behavior in the game, and some interested people start up giving their opinion about it, slowly the topic become in kind of "team" attempting to discover and break the AI and detection secrets.......... And i can see a very good job here....... :up: Every one of us had added important information about how the detecting and AI behavior works. Of course, not all us have the same problem, i have Ubbers Buckley and Evarts DDs wich kills the intersting in the game. CB had stupid DDs, wich kills the interesting in the game too. Finally we will arrive to a medium setting wichs give us a base to satrt up each individual personalization, may be some one want to release a mod, may be more than one mod, adapted to later times DDs, or more adapted for early time DDs. The work done here is more than excellent, we know how the main part of files works, we know how to manage Bolds settings. We only need the help of TT to manage sensivity in AI_Sensors.dat, and noise into Sensors.sim. Come on guys stay together :up: Back here CB :up: |
Quote:
one group says repeatedly that their DD's are more or less completely useless-- the other group says their DD's tend to be more uber (prompting the main subject of the thread) what am i do to? start intensively and rigourously testing soloutions to a problem i don't have? -in scripted missions designed to test changes that have little or no effect? (for me?) because you can't accept that those people who have said their DD's are useless are actually telling the truth--(why wouldn't they be?) i'm concerned here that because this makes constructing a one size fits all mod virtually impossible- you would rather this fact just went away!! so would I!!!-- but it isn't going to-- so i cannot join in the tests you and folks are suggesting because they are just not relevant to my problem and the results would not be relevant to yours!!! and as usefull as strictly controlled testing CAN be for checking details - you allways end up having to adjust the settings when confronted with all the differring conditions encountered in run of the mill campaign gameplay- - it's better IMO to take a step back from the problem and try to see the bigger picture-- but this approach would be more relevant to the problem of generaly useless DD's than the problem of Uber DD's-- so the bottom line is really wether folks accept that there are folks who are basically playing a screen saver because their DD's are useless-- if not then there's nothing i could possibly say that would make any difference-- and this is what i'm finding--in effect group A and group B are actually playing different versions of SH3 with different problems |
It very clear CB, the problem is not the same for all us, somebody has Ubber detection, and some body has Dummy detection routines.
So the common problem is the detection behavior and capability. You need to rise up it, i need to drop down it, whats the problem ? discovering how the settings works both problems will be solved. If some body want to test on a speciphied or standarized mission, let them, i will test that mission too, but i will tes on many more too. Finally the guys testing on a only one mission will be enforced to go to the game and test you changes into a multiple missions. I think so this fact is not enought to disolve this common job :up: Back to the topic, i am sttoped by the settings, if i adjust them to good gameplay for Ubber Buckley and Evarts, early DDs become Dummys and have the CB problem. If i adjust values to have a hard game play with "normal" DDs, then the later Evarts and Buckley become a hell. I think so the best way may be to adjust values according CB problem, and then attempt to drop down the Evarts and Buckley sensors. |
well yes out side of the problems with useless DD's and Uber DD's
(the fact that there are two seperate problems experienecd by different groups of people in a fairly consitent fashion should eventually if taken on board yeild some use full piec of the jigsaw puzzle) using the sim.cfg uber effct and calming them down with partial cfg entrys is proving very reliable for my useless DD problem-- the DD's tend to react when a ship is torped and start searching (that in itself is a huge leap forward) what happens next is very curious--- you get into a great period of gameplay where the DD's make seriuos attempts to kill you with varying degress of accuracy--(good gameplay) if you survive this period (about 3/4 of an hour perhaps) you enter into a dead end gameplay wise-- you find that the DD's retreat from the attack slowly one by one--untill only one is left----but that solo DD is un shakeable--and continues attacking even tho it has obviously run out of DC's ad infinitum-- never gives up-- i'm spending time watching this DD to see wether anything it does gives a clue to why it hasn't re-joined the convoy as all the others did-- it can be fooled with decoys but never lost completely-- it's DC drops are no more dangerous than any of the other DD's so i don't believe it is any more "expert" than the others-- yet it remains |
Right then lets step back a refresh the problems.
1. Everyone has dumb DD when it comes to initial detection via hydrophones and different crews. Anyone who says they dont are confused, innexpirenced or just plain overlooking the issue. That is the problem that needs most fix. I am agreed on that 100% on that issue. What I am not agreeing to is the fact that we should remove the cfg entries all together as thy are there to add variation within the game. My approach is get at least get 1 crew rating for all that matches the cfg settings in one standard mission. Once I am happy that we have that nailed down the true ranges, it then becomes, like you say, another issue of all the random weather and its effects in the game. This means more tweaking, plain simple fact. There will be no good fix unless this is looked at you know it and I know it. My possible solution was to set the waves and noise factors to such low number but all the same, numbers ratyher than blanks. Now I could be very wrong here. I am guessing that sensitivity is affected by crew ratings and I have no clue if noise is also affacted i.e. it also has a longer hex value. This is the part where I am openly dumb. Either way, it is as you say. We all have different motives to fullfill but I do suggest we stay in contact and at least extract info from each other rather than all go silent on the issue. CB why not start another thread and work on your work. If my theories (and thats all they are) are wrong, I will gladly sink into line with your suggestions via campaign test etc etc. 2. I am wondering if you are mistaking my outlook on the uber DD issue. When me and Redwine mention this, I do not think we are saying the DD is brainy etc. In fact I find them quite easy to shake off. We are simply saying that a DD should not be able to turn as the uboat turns over the last 100 metres. This simply should not happen. I think it can be fixed together with the rest but you may disagree which is fair enough as you may be right. An exagerated example If you would be so kind as to set a default 505 mission and then start to run with engines at flank or slower so long as you get the DD on your tail end. Watch this with external camera and when he gets close, try a hard rudder turn. You should notice that he immedietly follow you on the turn. That is the game killer for me and others. whats more, it severly screw up my plans for a damage mod. If you do not experience this, then we must have very different games and there is another underlying problem. So there it is, thats my main two problems. A There is no uber DD but there is a issue over the last 100 mtres that make them turn with you once they have got on your tail. B Sensors are screwed with by crew ratings (my theory) What do you suggest? I suggest we all just sit back and do our own thing for a little while so we can all calm down a little. Anyway I need to do less chatting and more setting up so I can try and show proof. Hey man, Your still a great guy in my book along with redwine and everyone else who has put their time into this. Peace |
your absolutely right about the turn thing--the DD's allways react to your turn when close and directly behind you even tho they have no feasible way of knowing you've made it ( and the imaginary Captain of the vessel can't guess right EVERY time !!)
i'm not suggesting tho that any one should use the sim.cfg uber effect as any sort of soloution to this sort of behaviuor-- but only for those folks who literally have to surface at flank in the middle of a convoy before the DD's detect them (this is not an exacegeration)-and then only till some-thing better comes along-- the problem with useless DD's is not a subtle one -- it really is that bad for some people (me included) lets keep pouring info and thoughts into the thread because bit by bit a clearer picture of why all this is happenning will appear-- quick thought on the turn thing-- is it possible to increase the noise generated by the DD props this might mask the u-boat as the DD excellerates to make it's DC run? or alternatively decrease the maximum depth (if this actually works?) of the DD's sensors to 150 metres-- i rememebr i kept reading hints that most of time DD's didn't understand that the u-boats could actually dive below 150 metres and nearly allways dropped their DC's at that sort of maximum depth--even late in the war-- there was a story of a type 9 transporting parts of the ME 262 and variuos scientists to Japan in order to help with the Japanese research into nuclear weopons - and how they when attacked simply went down to 300 metres and were perfectly safe as all the DD's dropped at 150 metres max--tho dunno how much sense this makes-- this would mean that if you went below 150 metres you could lose them --- with the caveat that you ain't going to sink many ships at 150 metres! and if your damaged you can't perhaps get that deep--uber up the phones to compensate ? |
re dd detection at close range -- in terms of active sonar, i've managed to get a dead zone w/out using the min range nerf eg if an escort is charging you, and you're at silent running, then eventually (distance depends on depth) he loses contact, the stealth meter goes green, and you can turn or whatever -- at silent speed. if you accelerate to flank, he hears you, and starts compensating for your position change.
my issue, specifically, is increasing the range at which sensors pick you up. if you drop noise factor, then you can be picked up alittle further away by hydrophones -- the downside is that at close range, you can be heard even at silent running, so the dd drops on your head. also, ideally, i'd like to make your aspect wrt the escort play more of a role -- head-on or stern-on, you should be relatively hard to pick up on asdic. @cb what exactly are you doing w/ sim.cfg? deleting it? what ranges are you picked up at subsequently? |
Quote:
Any way I do not want to go onto this subject yet untill we get hydro solved but here are some possible food for thought on the future. One we do find the best crew to use, (I am currently getting all of my crew to a rating of 4) If anyone else wants to try this type of testing, I have taken the liberty to set them all in campaign files as it takes time to manually change. remember this is based on a vannilla game patched to 1.4. http://rapidshare.de/files/8386990/C...rew_4.rar.html The missions should take each person 2 seconds to do Individually so I have not done that yet. Right back to topic. DC pinpoints, It is fair to say that they should not be able to turn on the last few seconds no matter if i hit flank or not, but that is their only advanages as far a uber is concerned. There is no other uber characteristic with DD's IMHO. CB, I also encounter the lone DD that remains and is persistant. Maybe this is what Beery and Jungmann were on about as far as the 90 degree bearing is concerned and it being immpossible to lose the sonar contact due to DD fast turns. Possible solution IMHO Once we get other issues solved i.e. the passives and ranges, I have a feeling we may have to only slightly change the ai sensor.dat file to compensate the small discrepencies. (I think we will have to do this at some point anyway) What we need here, is the ability to know what sensor we are dealing with in a particular mission. We then get this a close to settings as possible via very minor tweaks in cfg or infact we could just change the sensor to fit into the CFG settings. This is the trick part and like you say CB, we also have to consider the game weather etc. The next step would be to apply the same number differences to all other passive sensors i.e. if we needed to tweak type???? by 500mtre and this is proven in game, then we add 500 to each passive sensor. We then do the same for the actives, After this we should start to get a baseline that works. Though this is still only a base line and not the complete fix due to need of campaign randomness weather etc etc. Now onto the pinpoint. I remember early in this thread when someone warned me about changing the min ranges too much as sometimes the DD lose contact and go all over the show. What I now think is the following, Even though I set 200 mtre min values, they were in fact not 200 in the game due to the nerfs i.e. noise, waves, and crew ratings. Thats why it porbably never showed any great results in the game and pinpoint seemed to still be present for some yet ok with my settings. I have a small hunch that once we get the base line, we will find the uber pinpoint relitivly easy to solve because we will understand the range nerfs etc and waves, noise. I am so confident with this hunch, I would bet we could get very good results but that is a while off yet bacause as warned, we do not want DD's to suddenly become uber actives that hear us in silent. I have not yet even contimplated the aspect issue :88) If on the other hand, we could infact add a node to the DC and then attach a decoy into it, we would get better results. It seems there is 2 decoys to use in game, we could use one for DC with very short duration and keep one for all subs with better duration. That would be a great dream come true if any model modders could try this. Remeber I mentioned earlier about setting depth presision to 100 but the dc were too kaput and starting blowing the rear of the DD. Well that maybe was because of min depth setting being 25 hense they could blow too early. My Hedgehogs at 300*300 seems to show better behaviour and give me greater confidence for survivability in the later war. I am sure we will get this all solved but at the same time, I think it requires our tollerance, patience and open mindedness on the subject. It is not the time to think about making mods on this yet as we are all still headbashing the subject with each other and this could change at any moment with a new discovery :rotfl: :rotfl: :rotfl: :rotfl: :damn: :damn: :damn: :damn: . |
All times are GMT -5. The time now is 10:41 PM. |
Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2025, Jelsoft Enterprises Ltd.
Copyright © 1995- 2025 Subsim®
"Subsim" is a registered trademark, all rights reserved.