Aircraft Maneuverability (Final? Adding PR Spitfire)

Aircraft Maneuverability (Final? Adding PR Spitfire) (Full Version)

All Forums >> [Current Games From Matrix.] >> [World War II] >> War In The Pacific - Struggle Against Japan 1941 - 1945 >> Scenario Design

Message

el cid again -> Aircraft Maneuverability (Final? Adding PR Spitfire) (2/25/2006 12:47:28 PM)
UPDATE: The "proposed" ratings are now adopted for testing. For reasons unclear version 0.05 updated Japanese but not Allied planes - so this corrects that. The PR version of the Mosquito was revised backward in time so it is available at the beginning of 1944 - it is the first high altitude PR Mosquito. This data set is intended to give better ratings to biplanes and high performance twin engine fighters - and basically does. Examination of the database for aircraft indicated two patterns were used: Neglecting the divisor, these were either Speed/10 or Speed/10 + ROC/500. I felt that maneuverability was more than just a function of speed, so I adopted the second pattern. There was also a divisor - approximately the number of engines - except it was modified for special cases (4 engined divided by 8 vice 4; 2 engined fighters divided by 1 vice 2). I felt a more strict formula was needed so I made it number of engines period. This produced a generally fine result - slightly higher in some cases where the ROC factor had not been applied. Since we wish air air combat less dangerous, and since the primary factor in air air combat (plane wise) is maneuverability, I then divided this entire formula by 2: [(Speed/10) + (ROC/500)] /2. At one point I posted a proposal to use (Speed/20) + (ROC/500) but drew no comments. I will now show that this does for Allied aircraft as an alternate to the current data: Designation Type Present Proposed Speed ROC RN CARRIER AIRCRAFT Albacore Torpedo-Bomber 9 10 161 750 Avenger I/II Torpedo-Bomber 15 16 271 1430 Barracuda Torpedo-Bomber 12 13 228 840 Corsair III/IV Fighter 23 26 405 2750 Firefly I Fighter-Bomber 18 20 319 2000 Fulmar Fighter 15 16 280 1200 Hellcat II Fighter 22 25 380 2980 Martlet I/II/III Fighter 18 20 321 2030 Seafire I/II Fighter 20 26 335 4600 Seafire III/XV Fighter 20 27 352 4800 Sea Hurricane Fighter 18 20 315 2000 Swordfish II/III Torpedo-Bomber 15 8 139 500 Wildcat/Martlet IV+ Fighter 18 20 320 1760 USN CARRIER AIRCRAFT 0 0 0 F2A-3 Buffalo Fighter 18 21 321 2290 F4F-3 Wildcat Fighter 19 21 333 2030 F4F-4 Wildcat Fighter 18 20 318 1950 F4U-1D Corsair Fighter 23 26 405 2750 F4U-2 Corsair (NF) Night Fighter 23 26 410 2800 F4U-4 Corsair Fighter 22 30 446 3870 F6F-5 Hellcat Fighter 22 25 380 2980 F6F-5N Hellcat [NF] Night Fighter 22 24 375 2800 F7F Tigercat Fighter-Bomber 12 28 435 3000 F8F-1 Bearcat Fighter 25 29 423 3800 SB2C Helldiver Dive-Bomber 16 18 281 1800 US LAND AIRCRAFT 0 0 0 A-20B Boston Level-Bomber 10 11 350 2000 A-20G Havoc Level-Bomber 9 10 339 1408 A-24 Dauntless Dive-Bomber 12 14 217 1428 A-26B Invader Level-Bomber 9 10 355 1235 A-36A Apache Fighter-Bomber 19 23 390 1852 B-17D Fortress Level-Bomber 4 4 291 1333 B-17E/F Fortress Level-Bomber 4 5 325 900 B-17G Fortress Level-Bomber 4 4 302 541 B-18A Bolo/B-23 Level-Bomber 6 6 215 1010 B-24D Liberator Level-Bomber 4 4 303 909 B-24J Liberator Level-Bomber 4 4 300 900 B-25C/D Mitchell Level-Bomber 8 8 284 909 B-25H Mitchell Level-Bomber 7 8 275 789 B-25J Mitchell Level-Bomber 7 8 275 790 B-26A Marauder Level-Bomber 8 9 310 1250 B-26G Marauder Level-Bomber 8 8 283 1000 B-29 Superfortress Level-Bomber 5 5 358 526 C-46A/R5C Commando Transport 7 8 269 1300 C-47/C-53 Skytrain Transport 6 7 229 1130 C-54A Skymaster Transport 4 4 275 676 C-60A Lodestar Transport 7 8 266 1515 C-87/LB-30 Liberator Transport 4 4 306 1000 2xCG-4 Glider & C-47 Transport 4 4 150 720 F-4 Lightning Recon 10 12 350 2778 F4U-1 Corsair Fighter 23 27 417 2890 F-5C Lightning Recon 12 13 420 2500 F-6A Mustang Recon 22 24 400 1852 F-6D Mustang Recon 25 29 437 3500 OS2U-3 Kingfisher Float-Plane 8 9 164 344 P-26A Fighter 14 16 234 2360 P-35A Hawk Fighter 16 19 281 2439 P-36A Mohawk Fighter 19 22 302 3400 P-38G Lightning Fighter-Bomber 11 14 400 3777 P-38J Lightning Fighter-Bomber 12 14 414 3500 P-38L Lightning Fighter-Bomber 13 15 440 4178 P-39D Airacobra Fighter 19 22 335 2632 P-39Q Airacobra Fighter 22 26 386 3333 P-40B Tomahawk Fighter 20 22 342 2560 P-40E Warhawk Fighter 20 22 362 2083 P-40N Warhawk Fighter 20 22 352 2100 P-43A Lancer Fighter 20 23 356 2500 P-47C Thunderbolt Fighter 24 27 433 2640 P-47D Thunderbolt Fighter 25 28 428 3120 P-47N Thunderbolt Fighter 26 28 459 2700 P-400 Airacobra Fighter 20 22 360 2100 P-51B Mustang Fighter 25 28 442 2770 P-51D Mustang Fighter 24 27 437 2400 P-61A Black Widow Night Fighter 10 11 366 1667 P-63A Kingcobra Fighter 24 27 410 3425 P-66 Vanguard Fighter 20 22 340 2520 P-70A Havoc (NF) Night Fighter 9 10 342 1408 P-80A Shooting Star Fighter 16 18 558 4167 PBM Mariner Patrol 5 5 198 410 PBN Nomad/GST Patrol 5 6 200 990 PBY-5 Catalina Patrol 5 6 200 990 PB2Y-3 Coronado Patrol 3 3 213 570 PB4Y-1 Liberator Patrol 4 4 279 830 PBJ-1J Mitchell Patrol 7 8 275 790 PV-1 Ventura Patrol 9 10 312 2230 PV-1 Ventura (NF) Night Fighter 9 10 312 2230 R4D-5 Skytrain Transport 6 7 227 940 R5D-1 Skymaster Transport 4 4 275 676 SOC/SO3C Seagull Float-Plane 9 10 165 880 0 0 0 UK LAND AIRCRAFT 0 0 0 Beaufighter VIF (NF) Night Fighter 9 10 333 1923 Beaufighter Mk X Fighter-Bomber 8 9 303 1429 Beaufort VII Torpedo-Bomber 7 8 265 1200 Blenheim I Level-Bomber 8 8 285 1304 Blenheim IF (NF) Night Fighter 8 8 285 1300 Blenheim IV Level-Bomber 7 8 260 1500 Boomerang II Fighter 18 21 305 2940 Brewster 339D Fighter 18 21 321 2290 Buffalo Fighter 17 20 292 2570 Catalina I Patrol 5 6 200 990 CIXV-W Float-Plane 8 9 143 863 Corsair I Fighter 24 27 417 2890 CW-21B Demon Fighter 19 22 315 3280 Dakota I (C-47) Transport 6 7 229 1130 Do 24K-2 Patrol 5 5 186 597 Empire Patrol 3 5 322 950 F.K.51 Recon 8 9 146 1025 Hawk 75A Fighter 18 20 302 2500 Hudson IV Fighter-Bomber 8 9 284 1587 Hurricane I Fighter 18 21 316 2380 Hurricane IIc Fighter 19 22 339 2400 Hurricane IV Fighter 17 19 286 2200 Kittyhawk I Fighter 20 22 362 2083 Kittyhawk III Fighter 20 22 350 2000 Lancer Fighter 20 23 356 2500 Lancaster VII(FE) Level-Bomber 4 4 287 481 Liberator VI Level-Bomber 4 4 270 500 Lockheed 212 Transport 6 7 225 1360 Lysander I Recon 12 14 219 1350 Martin 139 (B-10) Level-Bomber 6 7 200 1860 Mitchell III Level-Bomber 7 8 275 790 Mohawk IV Fighter 19 22 323 2800 Mosquito B.XVI Level-Bomber 11 12 415 2000 Mosquito FB.VI Fighter-Bomber 11 12 380 2000 Mosquito PR.34 Recon 12 12 418 2000 Spitfire VB Fighter 21 24 374 2667 Spitfire VIII Fighter 23 26 408 2857 Spitfire XIVE Fighter 25 28 448 2857 Sunderland III Patrol 3 3 213 840 Tempest V Fighter 26 31 435 4700 T-IVa Torpedo-Bomber 9 10 161 1000 Thunderbolt II Fighter 23 25 427 1739 Vengeance I Dive-Bomber 15 17 279 1327 Ventura V Patrol 9 10 296 2230 Vildebeest IV Torpedo-Bomber 9 9 156 840 Walrus Float-Plane 8 9 135 1050 Wellington IC Level-Bomber 6 7 235 1120 Wirraway CA-1 Fighter-Bomber 12 14 205 1950 Spitfire VB Fighter 21 27 422 2857 USSR LAND AIRCRAFT 0 0 0 I-153c Fighter 17 20 275 3280 I-16 Type 4 Fighter 16 17 283 1583 I-16 Type 24 Fighter 19 23 304 4100 KOR-1 Float-Plane 10 11 171 1025 IL-2 Shturmovik Dive-Bomber 14 15 251 1367 IL-2M Shturmovik Dive-Bomber 14 15 255 1262 IL-4c Level-Bomber 8 9 267 1833 LaGG-3 Fighter 20 23 348 2780 La-5FN Fighter 24 27 403 3565 La-7 Fighter 24 28 413 3644 Li-2VP (C-47) Transport 5 6 188 927 MiG-3 Fighter 23 26 398 2928 Pe-2 Dive-Bomber 10 11 336 2343 Pe-2R Recon 10 11 340 2400 R-12 [Yak-4] Recon 10 11 345 2479 SB-2 Level-Bomber 7 8 255 1900 SB-2C Level-Bomber 8 9 280 1900 Tu-2S Level-Bomber 9 10 340 1726 Yak-1 Fighter-Bomber 22 25 360 3644 Yak-3 Fighter 24 28 407 4000 Yak-9D Fighter 22 24 374 2877 Yak-9UF Recon 23 26 395 2900 0 0 0 USN CARRIER AIRCRAFT Dive-Bomber 0 0 0 SB2U-2 Vindicator Dive-Bomber 14 15 242 1500 SBD-3 Dauntless Dive-Bomber 12 14 217 1428 SBD-5 Dauntless Dive-Bomber 13 15 245 1190 TBD Devastator Torpedo-Bomber 11 12 206 720 TBF/TBM-1 Avenger Torpedo-Bomber 15 16 271 1430 TBF/TBM-3 Avenger Torpedo-Bomber 16 18 278 2060

witpqs -> RE: Aircraft Maneuverability (New Proposal) (2/25/2006 7:19:53 PM)
quote: ORIGINAL: el cid again The question is this: to what extent is getting into firing position a function of speed and to what extent to things other than speed? Well, this is the $64,000 question. It depends on tactics - dogfighting versus slashing attacks. The way I understand real life in WWII is this, provided you use the appropriate tactics. If speed is similar and you have an advantage in maneuverability, you win (statistically, and all else being equal). If you have a speed advantage, you 1) survive even if at a maneuver disadvantage, 2) win if maneuver is good enough compared to opposition, 3) get some hits/kills if maneuver isn't good enough compared to opposision. Note that the above includes anybody getting some kills and suffering some kills whether superior of inferior in any category. It just outlines the statistical edge. As for how best to do it under the current code, I guess testing the proposed data is the best way to know. Look at results under controlled conditions and see what they look like.

el cid again -> RE: Aircraft Maneuverability (New Proposal) (2/25/2006 9:51:33 PM)
The comments above (and below) are well written. However, they actually miss something almost always missed on this board: a 90% determinant of air air combat victory is surprise - both offensive and defensive. That is, if you see the enemy first, and want to escape, you succeed 9 times in 10. If you see the enemy first, and want to attack, you succeed 9 times in 10. Neither speed nor maneuverability has any impact on these outcomes. Because an intelligent pilot who sees the enemy first maneuvers so that he is NEVER seen. Second to this, early in the war, the Japanese had a tactic which worked even against an enemy with superior performance. Until we understood this tactic (the translation of its name is "turning in maneuver") there was no defense - even if we actually had a better plane in maneuverability AND speed. It was observed in use by a student by the famous Genda, who asked to follow the student in a simulated dogfight. Once he understood this tactic, he taught it, and, unusually, it was adopted by JAAF as well as JNAF. This tactic may be one reason for the (time limited) "zero advantage" rule - although it should not be limited to that one plane type. There are probably other things like this - it is not always a case that statistical performance in either speed or maneuverabily is decisive. But we cannot model such things in the aircraft data itself - so this is just a comment. quote: Well, this is the $64,000 question. It depends on tactics - dogfighting versus slashing attacks. The way I understand real life in WWII is this, provided you use the appropriate tactics. If speed is similar and you have an advantage in maneuverability, you win (statistically, and all else being equal). If you have a speed advantage, you 1) survive even if at a maneuver disadvantage, 2) win if maneuver is good enough compared to opposition, 3) get some hits/kills if maneuver isn't good enough compared to opposision. Note that the above includes anybody getting some kills and suffering some kills whether superior of inferior in any category. It just outlines the statistical edge. As for how best to do it under the current code, I guess testing the proposed data is the best way to know. Look at results under controlled conditions and see what they look like.

witpqs -> RE: Aircraft Maneuverability (New Proposal) (2/25/2006 10:38:11 PM)
re surprise: True. That's why tests would have to be controlled and 'all other things being equal'. Ideally this would result in surprise evening out over the run of the tests, even though individually results might be lopsided due to surprise. A potential problem is this. WITP has some anti-cheat features built in. One is designed to prevent the IJ player from just rerunning a turn until he obtains acceptable results. It basically stores a random seed earlier, so if you rerun the turn you get the same results. Can a test be designed to work around this feature so that randomness is obtained in each iteration of a test? re 'turning maneuver: This is exactly the kind of technique I was alluding to. When dogfighting, this technique heavily favors a plane with superior maneuverability. It does seem likely that the Zero Bonus is meant to account for the Allied learning process in this area. Beyond the ZB, another question is does the combat model account for the different tactics (DF vs. Slash) used by different sides? Or, is it simply built into how the a/c data is used in formulas? It would take Mike Wood or somebody with similar knowledge/access to say for certain.

akdreemer -> RE: Aircraft Maneuverability (With Corrected Data) (2/26/2006 12:10:41 AM)
Okay, clear up something for me, and excuse me if this has been covered already. First, how are the maneuverablility ratings used in the game? Are they used for attack or defence? I seem to recall a previous discussion that the ratings were for how difficult the plane is to hit? The key offensive in the air-to-air routine was speed? Second, although I understand the need for testing the new values, how might have it been possible that the games designers got this all of this so terribly wrong? Is a limited mathematical equation sufficient when there are other quantifiable factors such as roll rate, turning radius, and accelleration also have key roles? I find it interesting that the best USAAF fighter in the Pacific in terms of total numbers of enemy aircraft shot down was the P-38 with, if I am correct,had a loss ratio greater that 10-1. If so, then there has to be other factors than climb rate, number of engines, and max speed that effects the maneuverability rating? Thus it could have been that these factors were incorporated into the quantifying the rating and that this rating is not depended solely upon the data inherent in the game model.

el cid again -> RE: Aircraft Maneuverability (New Proposal) (2/26/2006 12:13:21 AM)
quote: does the combat model account for the different tactics (DF vs. Slash) used by different sides? It cannot. There is no mechanism in place to allow a player to specify which. And it clearly is not always hard coded that one side or the other must use one or the other. It is clear from the model that IF air combat is joined the tactical mode is DF.

el cid again -> Japanese planes (Revision 0.06) (2/26/2006 3:57:26 AM)
Designation Type Maneuverability Revised Speed ROC A5M4 Claude Fighter 16 19 270 2745 A7M2 Reppu Fighter 23 26 390 3218 A6M2 Zeke Fighter 19 22 331 2642 A6M3 Zeke Fighter 20 22 338 2640 A6M5 Zeke Fighter 20 23 351 2765 A6M7 Zeke Fighter-Bomber 20 24 345 3129 A6M8 Zeke Fighter-Bomber 21 24 356 2880 F1M2 Pete [FF] Float-Fighter 13 15 230 1709 A6M2-N Rufe [FF] Float-Fighter 16 18 270 2440 N1K1 Rex [FF] Float-Fighter 18 21 304 2965 J2M2 Jack Fighter 22 26 371 3493 N1K1-J George Fighter 21 23 362 2514 J7W1 Shinden Fighter 26 29 466 2693 Q1W1 Lorna Dive Bomber 4 4 143 751 Ki-44IIIN Tojo Fighter 24 28 394 4021 Ki-76 Stella Level Bomber 6 7 110 820 D3A2 Val Dive Bomber 15 17 267 1696 D4Y2 Judy Dive Bomber 20 22 360 2139 B4Y1 Jean Torpedo Bomber 10 10 173 703 B5N2 Kate Torpedo Bomber 13 14 235 1283 B6N2 Jill Torpedo Bomber 19 20 344 1577 B7A2 Grace Recon 20 21 352 1900 C6N1 Myrt Night Fighter 21 24 379 2400 C6N1-S Myrt [NF] Level Bomber 21 12 379 2414 G3M2 Nell Level Bomber 6 7 232 1084 G4M1 Betty Level Bomber 7 8 266 1275 G4M2m22 Betty Level Bomber 7 8 272 863 G4M2e Betty (w Okha) Level Bomber 7 7 252 800 P1Y1 Frances Level Bomber 10 5 339 2315 G8N1 Rita Dive Bomber 5 11 368 1494 C5M2 Babs Recon 18 20 303 2480 J1N1-C Irving Recon 9 11 329 2335 H6K4 Mavis Patrol 3 3 212 1174 H8K2 Emily Patrol 4 4 290 1607 H6K2-L Mavis [Xpt] Transport 3 3 207 1174 H8K2-L Emily [Xpt] Transport 4 4 261 1237 L1N1/AT-2 Thora Transport 6 7 223 1483 L2D2/DC-3 Tabby Transport 6 6 220 800 L3Y1 Tina Transport 6 6 216 1007 G5N2-L Liz Transport 4 4 261 1082 E8N2 Dave Float-plane 11 12 186 1509 E7K2 Alf Float-plane 10 11 171 1082 E13A1 Jake Float-plane 13 15 234 1617 E16A1 Paul Float-plane 16 18 273 2109 E14Y1 Glen Float-plane 9 10 153 968 M6A1 Seiran Float-plane 16 17 295 1089 2xKu-8 Gander & Tug Transport 4 4 139 745 Ki-27 Nate Fighter 18 21 292 3057 Ki-43-I Oscar Fighter 18 21 307 2982 Ki-43-II Oscar Fighter 19 22 329 2821 Ki-44IIa Tojo Fighter 23 26 376 3832 Ki-45 KAIa Nick Fighter-Bomber 10 11 339 2610 Ki-45 KAIb Nick Fighter-Bomber 10 11 337 2585 Ki-45 KAIc Nick [NF] Night Fighter 10 11 336 2679 Ki-61-I Tony Fighter 19 22 319 2975 Ki-61-II Tony Fighter 22 24 379 2732 Ki-83 Stan Fighter 13 14 437 3280 Ki-84 Frank Fighter 22 25 392 2780 Ki-100 Tony Fighter 21 23 360 2734 Ki-102b Randy Fighter-Bomber 10 12 360 2750 Ki-32/30 Mary/Ann Level Bomber 15 16 262 1503 Ki-51 Sonia Dive Bomber 15 17 264 1653 Ki-48-I Lily Level Bomber 8 9 298 1823 Ki-48-II Lily Dive Bomber 9 10 314 1923 Ki-21-II Sally Level Bomber 8 9 302 1490 Ki-49 Helen Level Bomber 8 9 306 1250 Ki-67 Peggy Level Bomber 9 10 333 1361 Ki-36 Ida Level Bomber 12 14 216 1548 Ki-15 II Babs Recon 18 21 316 2380 Ki-46-II Dinah Recon 10 11 375 1460 Ki-46-III Dinah Recon 12 13 450 1300 J1N1-S Irving [NF] Night Fighter 9 10 315 1713 Ki-34 Thora (Ki-59) Transport 6 7 223 1483 Ki-56/LO Thalia Transport 7 8 249 1483 Ki-57/MC-21 Topsy Transport 8 8 290 1040

JeffroK -> RE: Japanese planes (2/26/2006 7:06:27 AM)
There are some real anomolies Ki-83 Only 4 prototypes made, twin engined in the style of the Tigercat or Hornet. But rates 14 which is worse than many bombers Ki-45 rates as 11 I think your formula is tough on twin engined fighters, as above and for P-38 & Mosquito. The D3A2 Val and A5M Claude rate at 17 & 19. Both had reputation of great manouverbility. Does the manouerability rating actually mean "Combat Ability"?

el cid again -> RE: Aircraft Maneuverability (With Corrected Data) (2/26/2006 11:40:03 AM)
quote: First, how are the maneuverablility ratings used in the game? Are they used for attack or defence? I seem to recall a previous discussion that the ratings were for how difficult the plane is to hit? The key offensive in the air-to-air routine was speed? Previous discussion was based on conjecture - reasonable conjecture - but conjecture nonetheless. We no longer are working with conjecture: the code ignores speed in the direct sense. It is focused on using maneuverability. Now we don't actually have the algorithm - the decision to release such things has not been made. But it is very likely that the way it works (aside from modifiers of various kinds) is to look at the maneuverability rating of both planes in a single impulse of air-air combat. And the way to do that well is to subtract the defender's maneuverability from the attackers - producing a differential. The differential becomes, in some sense, the probability the attacker reaches a position from which it may shoot. IF it does, probably BOTH planes can shoot - IF the defender has guns that bear in the situation - and IN THAT CASE then "accuracy" is used (we know it is really "hit probability" and it is the square root of ROF - which for this purpose is a reasonable concept). So the sequence is something like this: Do you reach firing position (roll on differential of maneuverability, with many modifiers)? IF YES, Do you hit (roll on accuracy, with modifiers)? ALSO IF YES, AND IF defender has guns that bear in this situation, roll for his shot. IF YES, what effect does your hit have (roll on effect, with modifiers, including armor as one of them). SAME for any defender hit.

el cid again -> RE: Aircraft Maneuverability (With Corrected Data) (2/26/2006 11:48:09 AM)
quote: Second, although I understand the need for testing the new values, how might have it been possible that the games designers got this all of this so terribly wrong? Is a limited mathematical equation sufficient when there are other quantifiable factors such as roll rate, turning radius, and accelleration also have key roles? First, I do not think "the designers got this so terribly wrong." I am actually impressed that they came up with such a simple system that works so well. I usually do much more complex calculations to get very similar results. Second, yes, it appears a limited mathmetical equation may be reasonably sufficient. The reason is that many factors involved in maneuverability are well modeled by ROC. As a general rule, a plane with really lousy ROC is not going to turn on a dime, for example. I never dared propose such a limited set of factors to model air combat - but I am very impressed with how well this works. It is in the same league as Fletcher Pratt's naval mineatures sytem - were visual estimation well approximated real naval shot estimates. That system correctly predicted the fate of the Graf Spee - but no one believed it UNTIL it happened almost the same way in real life! Really good designs ARE simple - and this may be a case in point. Third, what Matrix "got wrong" - and systemically gets wrong - is its combination of lack of documentation (which means that moders cannot know how to define things properly - critical things like maneuverability) and lack of testing (which means that even if they got all the database right and the algorithms in the right league, they never figured out the constant K needed to put the outcomes in the actual right range). Both these "errors" are deliberate, and related to cost. Before getting too upset about it know that a software company that does not get product out the door generating income before it runs out of money does not survive - much less thrive and generate more software. The crude rule of software is "get something that kinda sorta works out there generating income" - and Matrix is good at that - so it survives. Without doing WHATEVER it takes to make this happen, we would have NO product.

el cid again -> RE: Aircraft Maneuverability (With Corrected Data) (2/26/2006 12:02:26 PM)
quote: I find it interesting that the best USAAF fighter in the Pacific in terms of total numbers of enemy aircraft shot down was the P-38 with, if I am correct,had a loss ratio greater that 10-1. If so, then there has to be other factors than climb rate, number of engines, and max speed that effects the maneuverability rating? Thus it could have been that these factors were incorporated into the quantifying the rating and that this rating is not depended solely upon the data inherent in the game model. First, let me say I will try to address your concerns re P-38 ratings - and am working on that. Second, let me say that you should not be too impressed with simple numbers. The Matrix system is a SERIES of calculations - and the model is cleverly done: Your P-38 has something many other planes do not: PUNCH. That means, when it gets to shoot, better not be the target! All P-38s have a gun punch of 15 - which is sustained. The later two models in my set ALSO have rockets - value 240 - and that based on square root of warhead (because energy dissipates by the inverse square law). These are one shot weapons - but God help ANY sort of target when they go off! Ship, plane, ground target - ALL are in deep trouble if hit by these. Your P-38 has something else not many other planes have: RANGE. Multi engine planes have that advantage. The amazing A6M2 has a range of 32. But the P-38 has a range of 38! [If there is debate about wether a Zero pilot can stay awake long enough to reach 32 hexes, there should be even more debate about a P-38 pilot reaching 38.] But range matters. It means that P-38 can engage aircraft that have no fighter escort at all. In the Pacific range is a major factor - and it is the reason the P-38 looks so good victory wise. The Me-109 scored a lot of kills - but it was nothing like the best fighter plane of Germany - much less the world. The Val scored more ship sinkings than any other aircraft in history - but it is hardly the most impressive in performance or weapons. There is more to a good combat record than just statistical performance. Finally, allow me to say that the P-38 rating may yet improve - IF someone can give me the data the system needs to do its job right. I do not have Initial ROC data - so I am using a different ROC figure which is less than the IROC - it might be called 'average ROC to a defined altitude.' Get me the IROC and I will plug it in. [At the service ceiling, ROC is by definition 100 feet per minute = 30 meters per minute - that is the DEFINITION of service ceiling - you are at it when that is your ROC. So by definition all planes are equal in ROC at service ceiling. You cannot use that ROC as an indicator of maneuverability. It is the opposite end of the spectrum that is the best indicator: the Initial ROC. But this data is not always in the reference books. Instead, they give us "minutes to abcd feet" - which is a ROC - but not the one we really want. Sometimes we get IROC - sometimes more average ROCs - and so we do not have the option of using all one or all the other - which we really need to get the data right. Get me P-38 IROC and it should look a little better.]

el cid again -> Compare P-38 with Ki-45 (2/26/2006 12:07:18 PM)
The "Japanese P-38" is the Ki-45. Like the P-38 it is represented by three different marks - and could be represented by more. Like the P-38 it has real cannon - and therefore punch. Like the P-38 it has more range than most single engine fighters. Now compare the ratings - present and proposed - for both types. The P-38 is not only better in the present rating, when we double the impact of ROC, it gains more than the Ki-45 does. This implies that it really IS more maneuverable, and that the new rating system better reflects its advantage over a similar aircraft (in configuration and mission terms). Now compare the durability ratings of the aircraft. All Ki-45s are rated at 7. All P-38s are rated at 9. The chance of being damaged is significantly less for the P-38 - wether operational attrition - AAA attrition - or as a result of air-air combat. Now compare the ranges. They start the same - but as models are introduced the P-38 range increases greatly - and the Ki-45 range declines every time. Now compare the operational ceilings. The Ki-45 never equals the P-38, model for model. Over time the Ki-45 ceiling declines while the P-38 increases significantly. This means the P-38 will enjoy ever increasing altitude advantages over the enemy. In order to understand an aircrafts values in the model, you must consider ALL the ratings - range, maneuverability, speed, firepower, bomb load, service ceiling, etc.

el cid again -> RE: Japanese planes (2/26/2006 12:22:51 PM)
quote: There are some real anomolies Ki-83 Only 4 prototypes made, twin engined in the style of the Tigercat or Hornet. But rates 14 which is worse than many bombers Ki-45 rates as 11 I think your formula is tough on twin engined fighters, as above and for P-38 & Mosquito. The D3A2 Val and A5M Claude rate at 17 & 19. Both had reputation of great manouverbility. Does the manouerability rating actually mean "Combat Ability"? No it does not. Instead it means "the relative ability to achieve a firing position." Combat ability ALSO involves OTHER things - for example - IF you achieve firing position - how much punch does your shot have? Twin engine fighters generally are well armed. Also, IF you are hit, how durable are you? Twin engine fighters ALWAYS have an advantage here, and are more likely to survive. Also, how far away can you actually fight? Twin engine fighters usually have an advantage in this. Also, how much bombs/rockets can you carry to a target? Again, two engine planes normally have an advantage in this. It is the sum of these (and other) factors which determine combat utility. Note also I did not invent this system. I am just insuring the system is uniformly applied to all aircraft. Actually - the original system penalized 4 engine planes more than I do - and I think without cause. In this system 4 engine planes almost all have a maneuverability of 4 - a few have 3 - one has 2 - and a couple have 5. I see no need to cut these values in half because we divide by 8 as was originally done. IF you want more maneuverability for high performance single and twin engine planes, we should adopt the revised formula - which doubles the non-speed portion of the rating. The 4 engine planes remain anchored - but the 2 engine ones improve - and the really good ones you mention improve more. The single engine types improve if they were good - but not if they were not.

CobraAus -> RE: Aircraft Maneuverability (New Proposal) (2/26/2006 8:35:11 PM)
quote: A potential problem is this. WITP has some anti-cheat features built in. One is designed to prevent the IJ player from just rerunning a turn until he obtains acceptable results. It basically stores a random seed earlier, so if you rerun the turn you get the same results. Can a test be designed to work around this feature so that randomness is obtained in each iteration of a test? This is true when testing you will get the same result in all combat results all the time I have not tested this but found that before running a test change something - like pick a ship not involved in test and change its orders - the random seed changes then run test this seems to work for both sides Cobra Aus

witpqs -> RE: Aircraft Maneuverability (New Proposal) (2/26/2006 8:51:20 PM)
CobraAus - thanks!

witpqs -> RE: Aircraft Maneuverability (New Proposal) (2/26/2006 8:53:14 PM)
Sid, regarding air to air rockets - WITP does not have air to air ammo limits. Apparently it was taken out of the model in UV. Can air to air rockets be made to work realistically with unlmited ammo?

worr -> RE: Aircraft Maneuverability (Japan added/Updated) (2/26/2006 10:16:21 PM)
quote: ORIGINAL: el cid again Examination of the database for aircraft indicated two patterns were used: Neglecting the divisor, these were either Speed/10 or Speed/10 + ROC/500. I doubt it.

Daniel Oskar -> RE: Aircraft Maneuverability (With Corrected Data) (2/26/2006 11:34:19 PM)
The following rate of climb data is provided for the P-38. Source is the "Pilots Flight Operating Instructions for Army Model P-38. P-38H Standard Day 54"Hg/3000rpm to 20,000' 16100lbs 155 KIAS 3500 fpm Sea Level to 5000' 16100lbs 152 KIAS 3400 fpm 5000' to 10000' 16100lbs 149 KIAS 3200 fpm 10000' to 15000' 16100lbs 141 KIAS 2500 fpm 15000' to 25000' 16100lbs 131 KIAS 1200 fpm 25000' to 35000' 19500lbs 155 KIAS 2500 fpm Sea Level to 5000' 19500lbs 152 KIAS 2300 fpm 5000 to 10000' 19500lbs 149 KIAS 2200 fpm 10000' to 15000' 19500lbs 141 KIAS 1600 fpm 15000' to 25000' 19500lbs 131 KIAS 400 fpm 25000' to 35000'

Hoplosternum -> RE: Aircraft Maneuverability (With Corrected Data) (2/27/2006 3:53:23 PM)
quote: ORIGINAL: el cid again quote: First, how are the maneuverablility ratings used in the game? Are they used for attack or defence? I seem to recall a previous discussion that the ratings were for how difficult the plane is to hit? The key offensive in the air-to-air routine was speed? Previous discussion was based on conjecture - reasonable conjecture - but conjecture nonetheless. We no longer are working with conjecture: the code ignores speed in the direct sense. It is focused on using maneuverability. Now we don't actually have the algorithm - the decision to release such things has not been made. But it is very likely that the way it works (aside from modifiers of various kinds) is to look at the maneuverability rating of both planes in a single impulse of air-air combat. And the way to do that well is to subtract the defender's maneuverability from the attackers - producing a differential. The differential becomes, in some sense, the probability the attacker reaches a position from which it may shoot. IF it does, probably BOTH planes can shoot - IF the defender has guns that bear in the situation - and IN THAT CASE then "accuracy" is used (we know it is really "hit probability" and it is the square root of ROF - which for this purpose is a reasonable concept). So the sequence is something like this: Do you reach firing position (roll on differential of maneuverability, with many modifiers)? IF YES, Do you hit (roll on accuracy, with modifiers)? ALSO IF YES, AND IF defender has guns that bear in this situation, roll for his shot. IF YES, what effect does your hit have (roll on effect, with modifiers, including armor as one of them). SAME for any defender hit. El Cid where are you getting this from? The only tests I have seen are not supporting this view at all. The Zero bonus (which boosts Zero manouverability) was shown to have NO impact on the number of kills the Zeros scored, but did reduce their casualties. This suggests very strongly that the manouver values (perhaps relative manouver values and perhaps not) act as a saving throw / avoid damage or hit chance when a plane is attacked and do not help at all with attacking. From my own experience of the game and the general feeling on the boards I believe that speed and pilot experience seem to be the chief determinants of attack success (with firepower vs durability / armour then seeing if you get a damage or a kill). You appear to now be doubleing up the speed advantage by making it a chief determinant of manouverability. I am sure it is important anyway from the stats. Fast, late war planes tend to have good manouver values. If this is your intention fine, but I am not sure why exactly you want to do this. When you say it is very likely it works like.... that all sounds very plausible. But the testing done in the zero bonus thread does not support you at all. Manouver was shown to have no impact at all on kills (in a single air combat - obviously preserving good pilots longer has a long term boost). Why do you believe your ideas in this case to be correct? Have the developers said something?

el cid again -> RE: Aircraft Maneuverability (New Proposal) (2/27/2006 6:31:06 PM)
quote: Sid, regarding air to air rockets - WITP does not have air to air ammo limits. Apparently it was taken out of the model in UV. Can air to air rockets be made to work realistically with unlmited ammo? Yes and no. In a fundamental sense, ALL air air weapons are potentially wrong IF the weapon in question WOULD HAVE run out of ammunition. This needs to be addressed - and simply so - by simply reinstating the code that was written to do it (it is almost certainly still there - just remarked out or branched around). This takes no programming talent at all - and we probably will get it done fairly soon. What I did with rockets was define them as a single shot weapon! This is done by the accuracy setting - no other weapon has a setting of 1 - but all air-air rockets do. Accuracy is really square root of ROF - and sq root of 1 = 1 - so this is the correct value. So far testing indicates rockets are about as effective as they should be. Mostly they are used on ships - and of course that is a one pass deal - so it is not going to be an issue. The real problem with them is historical: they are an Allied weapon (a tiny Japanese rocket notwithstanding) - and by the later war period many planes begin to show up with them. It is just more nails in the coffin. I am not really worried about rockets on bombers - but those on fighters may turn out to be as much a problem in an uber-air battle as guns are. I am working on Joe's principle: get the data right and we will fix the code presently. Probably sooner than you expect.

el cid again -> RE: Aircraft Maneuverability (With Corrected Data) (2/27/2006 6:36:13 PM)
quote: The following rate of climb data is provided for the P-38. Source is the "Pilots Flight Operating Instructions for Army Model P-38. P-38H Standard Day 54"Hg/3000rpm to 20,000' The only problem is that we are using P-38G, J and L models. The first value you give in your list is the initial ROC - the one nearest the ground. And it does help. But we need that value for all the models, and as it happens your manual is NONE of the models we need. It is a thousand feet per minute higher than we are using though - it might matter.

el cid again -> RE: Aircraft Maneuverability (Japan added/Updated) (2/27/2006 6:39:14 PM)
quote: quote: ORIGINAL: el cid again Examination of the database for aircraft indicated two patterns were used: Neglecting the divisor, these were either Speed/10 or Speed/10 + ROC/500. I doubt it. I wonder why? When a max speed = 350 and the maneuverability = 35, I see a formula max speed / 10 why don't you? I am trained to reverse engineer codes and formulas - and this was a trivial exercise for me. Particularly since our resident mathmetician (Joe) advised me to begin with speed as a possibilty.

el cid again -> RE: Aircraft Maneuverability (With Corrected Data) (2/27/2006 6:50:36 PM)
quote: El Cid where are you getting this from? The only tests I have seen are not supporting this view at all. The Zero bonus (which boosts Zero manouverability) was shown to have NO impact on the number of kills the Zeros scored, but did reduce their casualties. This suggests very strongly that the manouver values (perhaps relative manouver values and perhaps not) act as a saving throw / avoid damage or hit chance when a plane is attacked and do not help at all with attacking. From my own experience of the game and the general feeling on the boards I believe that speed and pilot experience seem to be the chief determinants of attack success (with firepower vs durability / armour then seeing if you get a damage or a kill). You appear to now be doubleing up the speed advantage by making it a chief determinant of manouverability. I am sure it is important anyway from the stats. Fast, late war planes tend to have good manouver values. If this is your intention fine, but I am not sure why exactly you want to do this. When you say it is very likely it works like.... that all sounds very plausible. But the testing done in the zero bonus thread does not support you at all. Manouver was shown to have no impact at all on kills (in a single air combat - obviously preserving good pilots longer has a long term boost). Why do you believe your ideas in this case to be correct? Have the developers said something? I am not really guessing about what factor is used. Speed as such is NOT used in the air-air combat routine. Instead maneuverability is. However, after analysis of the maneuverability value, I find speed really is the dominant element in it anyway - disguised by a divisior for multi-engine bombers and transports. It is not wise to say more than "I am not making this up." I do think your concept that things like pilot experience matters is correct. I also think altitude, surprise and other things matter. I am only describing the heart of the routine - omitting all the modifiers and conditional jumps. For the purpose of this discussion, the only thing I am interested in is the technical factors of the aircraft which are used in the air combat routine. I don't need to consider any non-aircraft factor when rating a plane. I do need to understand how the values work in order to rate planes properly - and to insure that all the planes are done to the same standard (which, clearly, they have not been, which seems typical of this game - NO database is consistently done).

worr -> RE: Aircraft Maneuverability (Japan added/Updated) (2/27/2006 7:46:54 PM)
quote: ORIGINAL: el cid again I wonder why? When a max speed = 350 and the maneuverability = 35, I see a formula max speed / 10 why don't you? That certainly is possible, but not probable. There are aircraft that have a max speed over 400 and do not have a maneuverability of 40 plus.

witpqs -> RE: Aircraft Maneuverability (New Proposal) (2/27/2006 8:31:51 PM)
quote: ORIGINAL: el cid again .. by simply reinstating the code that was written to do it (it is almost certainly still there - just remarked out or branched around). This takes no programming talent at all - and we probably will get it done fairly soon. ... get the data right and we will fix the code presently. Probably sooner than you expect. It sounds like you know somthing the rest of us would like to know. Are you talking? quote: What I did with rockets was define them as a single shot weapon! This is done by the accuracy setting - no other weapon has a setting of 1 - but all air-air rockets do. Accuracy is really square root of ROF - and sq root of 1 = 1 - so this is the correct value. I take it this will make the rockets single shot per plane fired at, but that it will still be able to fire at multiple planes if so engaged in combat (one after another, the way the model works).

el cid again -> RE: Aircraft Maneuverability (Japan added/Updated) (2/28/2006 2:01:59 AM)
quote: That certainly is possible, but not probable. There are aircraft that have a max speed over 400 and do not have a maneuverability of 40 plus. The key to decoding the algorithm was to add in the divisor for multi-engine planes. There is NO case where a SINGLE ENGINE plane EVER had a maneuverability less than 10% of its speed. Many cases have more. This is the simple case: the divisor is one for a single engine plane. Turns out it was ALSO one for a twin engine fighter or night fighter. But it was two for a twin engine bomber or transport - and interestingly apparently 8 for a four engine plane of any sort.

el cid again -> RE: Aircraft Maneuverability (New Proposal) (2/28/2006 2:08:56 AM)
quote: It sounds like you know somthing the rest of us would like to know. Are you talking? I have been told to shut up.

el cid again -> RE: Aircraft Maneuverability (New Proposal) (2/28/2006 2:15:05 AM)
quote: I take it this will make the rockets single shot per plane fired at, but that it will still be able to fire at multiple planes if so engaged in combat (one after another, the way the model works Maybe. It won't be the SAME plane - the rockets have the broadside of a heavy cruiser! In fact, if there WAS a "target rich environment" (say a bomber formation densely packed) it would be hard NOT to get several kills per salvo! Another factor is "are the rockets used?" That is, when given multiple weapons, the code does not always elect to use all of them (see submarines - I have torpedoes and midgets - and fore and aft versions of torpedoes and kaiten - but I never see ALL used at the same time). However, I agree with you - IF there is a multi-round air combat, AND IF the rockets are chosen for use by the "code" pilot, it may be we have somehow got "reloadable" rockets (which is perfectly possible - just not historical). It is as bad as the guns that never run out of ammunition. But I know of one Japanese pilot who engaged 18 F6Fs, shot down something like eight of them, and the rest ran - for fear he was NOT out of ammo! Clearly he fired short bursts. Maybe if we have such a case with rockets, they do not fire all at once? If you have a solution to this problem short of code changes, I am all ears too. For example, you might propose "put rockets on bombers, but not on fighters, until we fix the ammo problem." For the record, rockets so far have worked on ship targets. They have yet to be selected for use in air combat - it may only be they are chrome and do not work at all. This is a test mod - in its pre release test mode - time will tell.

witpqs -> RE: Aircraft Maneuverability (New Proposal) (2/28/2006 4:52:58 AM)
quote: ORIGINAL: el cid again If you have a solution to this problem short of code changes, I am all ears too. For example, you might propose "put rockets on bombers, but not on fighters, until we fix the ammo problem." I wish I did. Maybe the best thing is to see how they work and, if too powerful (unrealistic), then take tehm off the fighters. Without code changes I don'r see a better way. I suppose it's also possible that the code will only see rockets as an air to surface weapon (I do not know how you defined them - like a gun or like a bomb? - so I don't even know if this is actually a possibility).

el cid again -> RE: Aircraft Maneuverability (New Proposal) (2/28/2006 2:07:48 PM)
quote: I suppose it's also possible that the code will only see rockets as an air to surface weapon (I do not know how you defined them - like a gun or like a bomb? - so I don't even know if this is actually a possibility). My "rockets" are not like the "ss rocket" devices in the game. They are renamed aircraft guns! This so they should work in air air mode. I suppose if they are too powerful, we might reduce the salvo on fighters to only 2 - that way it gets 3 or 4 shots with correct values - which should cover most battles. Due to reduced durability ratings, I do not expect many planes to survive hits by heavy weapons!

Page: [1] 2 3 next > >>

Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI

0.9375