Jump to content

AnPetrovich

1CGS
  • Content Count

    347
  • Joined

  • Last visited

Community Reputation

1846 Excellent

7 Followers

About AnPetrovich

  • Rank
    Developer

Profile Information

  • Gender
    Not Telling

Recent Profile Visitors

1783 profile views
  1. The greatest pleasure and ultimate goal for any of us in the team is to see our players happy. This is the most pleasant motivation that keeps us work! Fly safe, have fun!
  2. Sorry guys, it was a busy time! I remember, I promised the test results for 90°, here it is: (P.S. Actually "fix is coming" is needed to read as "now in 4.006")
  3. Unfortunately, it is not possible to run this test with 4.004, since the DM was significantly changed in hundreds of files when transfering to version 4.005. I can not turn things back and moreover run the new tool with old DM. Yep, I'm also sad about this because I'm also interested in this comparison.
  4. Quick answers: Not exactly in the same spot but at the same part of the wing. Thus, the length of spar also does matter. 110 is an average number as a most likely result. In a specific situation, this number may vary. See my explanation here. From 117 mm to 220 mm in the D.VII and from 86 mm to 102 mm in the F.2B. I'm glad you're on board again! Honestly, I never heard that G-load limits the ability to use a control stick. However, we simulate limitations due to the force response of the stick and rudder at high airspeed. I watched your video. Have you read this and this posts? Yes, you have, otherwise you would not make this video. So, what another result did you expect? Do you think this 'issue' does really matter in a real dogfight, unlike in this test? You are welcome with any proven sources you are sure of. I would say that ~20 hits in average (in this particular case) is a most likely result according to the tests. See a distribution of probability curves. Are you serious? We are not NASA, we do not pretend to simulate a "flight to Mars" or how every bolt or nut works. This is a real-time game on your PC with many objects at the same time, and it still could not been accelerated any faster than x2 (sometimes) while players complain that x16 is a 'damn' limit... Maybe one day we will simulate an animal glue, dude, but not this time. I am going to show a chart with 90 deg results a little later. If a bullet hits the wing from dead six, then the probability of hitting the spar is pretty the same for a thick or thin wing. That's all for now, I come back to work...
  5. Guys, please do not confuse the interpretation of the test results and do not make a mockery of probability theory! The test condition was: "how many HITS are required", but not "how many bullets should be fired". We do NOT speak here about someone's aiming ability or about the probability of hitting the wing when the airplane is under atack. The case when the airplane get hits from the "dead six" is quite informative. Just because almost every hit at wing from this direction leads to damage of the spar(s). This is not very accurate because there is a wings incidence angle, but this case is still pretty close to the most "worse" of situation anyway.
  6. Let's think about it and just leave the DM outside of this topic for a few minutes. I suggest doing a mental experiment. Suppose there are 100 young guys who have never flown any flight simulator. They receive a ground training and a "flight" training on the FC with instructors, exactly the same way real students in real life would get. No one is allowed to do any of "fatal mistakes" or even to be involved into an "accident" during this training. The first question is: how many of those guys will complete this training with no accidents? But, let's assume that all of these 100 guys will become "pilots" successfully. Then my next question is: after they start fighting against each other in the virtual sky, what are the most frequent reasons of their "death" would you expect? How many guys will "die" because of "lost of control"? Even taking into account that the bet is not their life! Do you follow my thought? You, guys, have THOUSANDS of hours of "airtime". You have no fear to die. You did crazy stuff in the virtual sky plenty of times, which a real WW1 pilot would never has done in the real sky. How could you compare your experience with real WW1 pilots? Do you still really expect that the most frequent reason of crashes in the GAME would be the same as in the real life, such as "lost of control"? How often you guys crash in the game because of losing control nowadays? Do you think that crash because of the "pilot's death" or "engine damage" or "fire" or "airframe damage" should be even less frequent ? By the way, I'm not sure I have balls (sorry for my English) to loop a damaged airplane (after it got hits) at few hundreds feet above the ground, remembering that this airplane is designed using "sticks and ropes"... Maybe I'm not a bold pilot. Are you? (Just a joke)
  7. Hi guys! Here is some analysis of data I have collected in this topic, and data I got from an examination of the current DM. At first, I would like to say thank you everyone who took part in this poll! At least 40 FC-users seem to be interested in this subject (unfortunately not hundreds as I would hope but anyway a significant amount of players!). Here is a sheet with final results of your voting, guys: Based on this data we can easily draw a chart of distribution of your perception (blue curve) and expectations (green curve) regarding the number of hits required to break a particular part of Camel's wing. I remind, that the initial question in this poll was about an outer one third part of the top right wing of the Sopwith Camel, which flies in the level flight, and get hits from the "dead six" from a close distance. On this chart you can see how many users (vertical axis) think that a particular number of hits (horizontal axis) required for breaking the wing (all pics below are clickable): After uniting some close-to-each-other answers (using the very well known formula for calculation the "weighted average") and transfering from the "amount of users" to the "distribution of probability", we can see such result: Well, it looks like there are at least two noticeable groups of you (players) who have similar perceptions (how you evaluate the current DM), and at least two groups of you who have similar expectations (what you want to have eventually). I can see here that most of you think that right now only approximately five hits are enough to break the wing in this particular situation. Some of you, however, think that the number of hits is something around 15. I also see that some of you expect that this number of hits should be about 25, while others think it should be ~50. This is pretty interesting data to think about, isn't it? It becomes even more interesting after I add here the curve that mirrors how the current DM really works. As I told you recently, I have a special dev-tool that allows me to perform thousands and thousands tests very quickly, instead of firing at an airplane in the game. Using this tool in each test I can measure the number of hits required to break a particular airframe part of the particular airplane, taking into account the direction of shooting (in 3D space) and the particular type of ammo. In the past there was an issue with this tool: the airframe was tested unloaded (meaning the zero-gravity conditions). Thus, this tool measured the number of hits taking into account only the self-strength of the airframe. But here is a good news: last week I improved this tool. Now we can perform this test for the airplane in flight, at given altitude, airspeed, and G-load (in the level-turn other than 1G), while the airplane remains balanced by control surfaces. If the airplane can not be balanced at given airspeed and G-load because of the power deficiency, than we simulate a hard-break turn, meaning that the airplane reaches this G-load at this airspeed while it decelerates during the level turn. Finally... Using this upgraded tool last week I got a distribution of probability of the number of hits exactly for this particular situation, I asked you in this poll. Which is: how many hits of 7.92 mm bullets fired from the "dead six" from a close distance is required to break the outer one third part of the top right wing of the Sopwith Camel, which is in the level flight? Here is a gray curve, as a result of 10'000 tests, performed for the Camel in the level flight (1G) at altitude 500 meters MSL, at 180 km/h: This is how the current DM works. Wow... looks like the current DM is something between what you do feel and what you do expect from the game. It is very interesting, that most of you feel the current things differently. Especially taking into account that everybody can perform their own tests in-game to figure out real numbers (and some of you did! - and gave the most relevant answers, by the way, as I noticed). Yes, nobody has a tool like I have, and you could not conduct thousands of tests. But at least it is possible to do a few test like this one (this is my in-game test from one of the 4.005 Betas). Some may notice that it looks like players’ perception is based on dogfights, where G-load is a factor as no one flies at 1G in dogfight. Of course, in this situation, lesser number of hits are well expected. Although, firstly, my original question in this poll was about the level flight. So, I hope answers were also about this condition, rather than about turns at high G's. Furthermore, with the new tool it became possible to perform the test at any G-load! Thus, we can see how the distribution of probability of the number of hits changes with different G-load! I performed the same test for the Camel, but this time not only for one particular part of the wing, but for each part of its wings (10'000 tests for each part), then I calculated the average numbers. Here are the results for any "typical" part of Camel's wings on average, depending on G-load: This time on the vertical axis of this chart there are numbers of cases (tests) instead of probability. Let's dive deeper. Look at the amount of cases (tests) which ends with less than 10 hits required to break any “typical” part of wings. It is relatively low, even under a high G-load condition! For example, at 5G only in 787 out of 10'000 cases the part of the wing was broken after 5 or less hits. This is only 7.9% of all tests. Or, in other words, only in one out of ~13 cases the wing could be broken after 3-5 hits. The same way we can figure out that at 5G only in 2'191 out of 10'000 cases the part of the wing was broken after less than 10 hits (the previous cases included). In other words there is about 22% chance of such a situation, or it is one case out of five. There is also a statistics for others G's: Let's repeat it again: Only in one out of 13 cases the “typical” part of Camel's wings can be broken after 3-5 bullets hitting the SAME part of the wing (not whole wings!), fired from the "dead six", while Camel is doing a hard-break turn at 5G at 180 km/h. If the G-load is less than 5 - than more bullets are required. If one fires not exactly from the "dead six" (e.g. from the top) - than many bullets will "hit the air" and in the end more bullets are required. If one spreads bullets over the wings rather than place them at the same part of the wing - than more bullets are required. But somehow most of the interviewed players think that only about 5 hits are enough to break Camel’s wing in the level flight. Although it is possible only in one out of 169 situations as tests shows... This is really, really interesting, why players feel this way, and why you guys take so rare cases as a basis for your perception and conclusions. Maybe because we only memorize the most unusual cases? Maybe because we fly virtually and pull high G on these airplanes regardless of received damages, unlike a real pilot who has a real fear? I don't have an answer.. really. Anyway. I have finished the investigation of "what situation really is" and "how you perceive this". Moreover, now we have your expectations of "how it should be". Seems like we need to change the airframe durability. Or change the DM taking into account some other factors related to the process when bullet damages a spar, to make this damage not very hard. Perhaps... I asked about the Camel in that poll not accidentally. As I mentioned in other topic, I have already corrected the size of spars of all our WW1 airplanes in FC (this fix is coming soon). Some of the airplanes spars have been corrected a bit, while others have been corrected significantly. The durability of Camel remains almost the same. That's why I asked about this plane. Just to be aware what to expect from your expectations. I performed the test for all WW1 planes to compare their durability before and after I have corrected their spars size. This test was performed before I have updated the tool, so it was done under zero-gravity conditions. But it gives quite a good understanding of airplanes durability anyway: Any of your constructive comments and thoughts are very welcome. I open this topic for discussion, so you can post right here. What should we do? "That is the question!" (c)
  8. Closed for analysis finalisation. Feedback is coming soon.
  9. That's correct! Moreover, not only for the visual DM, but for the skin DM (and aerodynamics destortion) as well. I guess you guys would like to see all of your hits at enemy's wings, even if bullets "hit the air". The SPAD XIII.C1 in FC has 10 hit-boxes for the top wing, including 7 wing's sections, two ailerons, and one top fuel tank. If we want to put additional hit-boxes for the spars, we have to add 14 new hit-boxes: two for each section of the wing. Otherwise, if we add only two hit-boxes (one per spar), we won't be able to detect which exactly section of the spar (wing) should be broken, and also there will be a possibility to wrongly "hit" the spar in the area where the section of the wing has already been lost (this is not a critical issue since the wing is broken and the airplane goes down, but... still weird). By the way, to break a spar, firing, for example, from the 'dead six', you need to get a few hits at the same place of the spar, lengthwise. The minimum number of hits, placed like this, which are needed to break the unloaded spar simply depends on the spar's 'height'. What is the chance, that you put all of your hits at the same place of the spar, lengthwise? It depends on its length, obviously. Therefore, the average number of hits for breaking some section of the spar, also depends on its length. If the section is long, you need to make a good number of shots, untill at least few of them will reach the same spot. But sometimes it might happen occasionally faster. Or, instead, longer. This is how the probablity theory works. Certainly, if the spar is loaded (because of aerodynamic forces) then it needs less number of hits at the same place of the spar for breaking it. The most damaged place of the spar breaks the faster the higher G-load is.
  10. I have no excuses, except for the well-known fact what a programming is. It's the same as if you would be a writer and wrote a novel, but only one missing comma on the page 327 would devalue whole your work.
  11. Yes, I expect that some of you will perform this in-game test by yourself. I like it, it's very welcome. Sometimes our feeling does not match what we really have. Just because of our nature of perception. Thus, well performed test or an experiment could help much to figure out what is really going on. I'm not just asking you in that poll. I'm going to give you a feedback in a couple days, after the majority number of you are answered. I will write the result of our own tests (I use a special dev.tool for this, I will tell you later about it) and we will compare your and our expectations with what we have right now and where we go. Hold on, keep tuned!
  12. Thank you for your support.
  13. Hi everyone, I have a very specific question for you, guys. This is important. As we realized, many of you think that the number of bullets needed to break wings of WW1 airplanes is way too small in the current version of DM. Well, I've made an additional research on this issue. As I mentioned before, there is no clear and the only one "perfect" solution for the "absolutely ideal" DM, unless we simulate each corpuscle (which is not expected this year, obviously). Every our step is just an attempt to get closer to the reality. There are too many factors that make sense and difference, and we can not be 100% sure that our DM is absolutely correct. Anyway, this DM is based on the calculations, blueprints, charts, researches, and experiments that have been conducted. On the other hand, we see expectations from you, which differ from the results of our calculations. It's hard to say how far we are from the reality or how close to it. But we want you to be happy and to believe what you see on the monitor (VR is better, btw). Before we make any "final" decision (final for now, as we need to move further) and apply final adjustments on the DM, I would like to understand your expectations better. Just a quick check, to be on the same page. My question is very simple. Let's imagine the situation, that you are in the Fokker Dr.I in the level flight, having enemy Sopwith Camel in front of you at a distance about 100 meters. You are firing at him from the "dead six" position. You can clearly see that all (or most) of your 7.92 mm bullets hit the top right wing in the outer 1/3 part of this wing (i.e. the area between the tip of the top right wing and 1/3 of the right wing span counting from its tip). But from this distance you can not see at what place exactly you hit that part of the wing. For you it looks like you "spread" bullets over that particular part of wing. What is your expectation: how many HITS (on average) at this outer 1/3 part of the top right wing of the Sopwith Camel are needed, in your opinion, to break this part of wing off? Please, write here only one your answer, as a number. Nothing more. No explanation, no discussions, nothing. Just one number that you consider as a correct answer "on average". Any other answers will be removed. UPD: And please, after that number also write the 2nd number in parenthesis: what do you think, how many bullets are required in game for this right now (on average as well). Example: X (Y) If you want to discuss this subject, please, feel free to do it at this topic. Thank you!
  14. P.S. By the way, I've noticed here a little discussion of a funny picture of a magic cat, after someone has quoted my post from the ru-forum. Sometimes Google translator makes me crying. I must say that this Google-translation of my post is fully bullshit (sorry for my English) and has no sense in confusing words. Maybe someone here, who knows Russian (emely?) will translate my words better than Google.
  15. Hi guys! I've read the last few pages very briefly and I would like to give a few quick answers: 1. We are aware that a significant part of FC players perceive the new DM as a step back to the RoF. We hear your voices and are working on this issue. 2. In the current DM, as in the previous one, the direction of the bullet's impact into the wing is taken into account. For example, if the bullet hits the wing from above or below, the chance of spars damage is significantly reduced. Thus, most likely, the bullet "hits the air" in this case. The probability of hitting the spar when hitting the wing depends on the direction in 3D-space. 3. We do not want to use detailed hit-boxes for each element of the airframe structure (such as spars, stringers, ribs, wires, rods, hinges, etc.) because of two major reasons: 3-a) we try to save CPU time, while increasing the number of hit-boxes leads to an increase in cross-checks in collision detection algorithms, as well as usage of memory for all layers of 3D-models in geometric progression; and 3-b) an increase in the internal detailing of 3D models increases the cost of producing detailed content (in terms of time and money) and pushes us out of business. Therefore, we use major hit boxes with combination of methods of probability theory to clarify the fact of getting into a particular major element of the airframe structure. 4. Yes, we are aware that a couple of years ago ED started developing a very detailed model of hit-boxes for some airplanes in DCS. We still think that this significantly affects performance and this decision is premature for the real-time simulation of more or less massive dogfight on typical PC. 5. We do not take into account the angle of hit the fabric. Every bullet makes a 'virtual hole' in the skin of size that depends just on the caliber. As you know we have only three levels of visual damages to the skin for each part of the airplane. This is a convention of the simulator at this stage. 6. When a bullet hits a spar (taking into account the probability of hitting it), wing loses its strength. The load that the wing can withstand depends on the degree of damage of its spars. 7. The extent of damage to the spar depends on its material and on the number and caliber of bullets that hit the spar 'mathematically' at the same place (based on probability theory). That means that size of spar makes sense. 8. During the check of this 'FC problem' we figured out that the data source we used for determine dimensions of spars for all our WW1 airplanes was unexpectedly not very accurate. Last week I checked all blueprints and charts that we have, and I must say there is quite a difference with data we used. Some of our WW1 airplanes have almost right size of spars (Pfalz f.e.), while others have them from 10% to 30% (such as Albatross, Halberstadt, etc.), or even more than 70% thinner (Fokker D.7). That's sad, and we will correct this as soon as we are ready with the new and well checked data. 9. But this is only one side of the problem. The other side is that we need to improve the airframe DM from AP bullets anyway. Work in progress...
×
×
  • Create New...