Jump to content

Justin S.

Members
  • Content Count

    20
  • Joined

  • Last visited

    Never
  1. Thanks for both Sturmovik. I recommend using yousendit.com in the future for file transfers (doesnt cap transfer rate).
  2. The major problems with the data above are from the turret test. The percentages for turret vs. upper hull hits while hull down, and the penetration chance against the turret are both off. Instead of 82% of hits landing on the turret when hull down, it is actually 75%. The penetration rate against the turret in this test is 62%, not 73%. These values completely change the outcome. Interestingly enough, a slightly more resistant, but still vulnerable turret makes going hull down less advantageous- this is due to the high hit percentages against hull down targets. Reading the PzIV turret debate, Treeburst155's conclusions seemed at odds with mine. He had tested about 500 hits; I wanted to find out if the discrepancy was an inaccurate sampling (not enough data) on my part. Same setup as before; this time recording 500 hits. Hit Distribution Against Exposed Panther, 500m: Lower Hull- 15% (75 hits, including track) Upper Hull- 46% (229 hits) Turret-------39% (194 hits, including gun) This deviates about 5% from my previous testing (100 hits- 20%, 41%, 39%). If we average this with Treeburst155's distribution (http://www.battlefront.com/discuss/ultimatebb.php?ubb=get_topic;f=30;t=001294;p=4) we get this (1012 hits): Lower Hull- 15% Upper Hull- 49% Turret-------36% I beleive the ratio is 50-35-15, as the turret distribution came out in nice fractions as well. I also conducted a seperate test to determine the hit distribution on hull down AFVs (again 500 hits): Hit Distribution Against Hull Down Pather, 500m: Upper Hull- 24% (121 hits) Turret-------76% (379 hits- 235 penetrations or 62% of hits) (Treeburst155 had almost identical numbers- 26% and 74%) Combined with Treeburst155's data it is clear that the 75/25, and 50/35/15 ratios for hit distribution, regardless of vehicle type , are very accurate estimations (good call Panzer76). With these relative "constants" we can construct some formulas. P= Penetration p= probability h= hit T= turret H= upper hull L= lower hull To find the probability of penetration against any AFV in CM: Exposed: Pp= hp[ .35(TPp) + .5(HPp) + .15(LPp)] Hull Down: Pp= hp[ .75(TPp) + .25 (HPp)] Now, this isnt as handy as it looks because you need very accurate data to fill in the probablities. However it becomes much more useful when you can cancel out factors because of "ones" or "zeros" (ie turret is always penetratable, or hull is proof). When we run the new, more accurate data this is what we get for our test: First Round Penetration Chance Against Panther, 500m: Exposed: .69[ .35(.62) + .5(0) + .15(.4)] = .191 or 19% Hull Down: (avg. hp of .46, not .47) .46[ .75(.62) + .25(0)] = .214 or 21% With all the variables, I believe this is too close to call. As the range increases (hit percentages agaisnt hull down really drop, turret is less vulnerable), the arguement for staying hull down is sound. But, as the range decreases, hit chances climb and turret vulnerablity rises. These two factors combined with the very high hit percentages against hull down targets make staying exposed the better choice in a knife fight. The next discussion Id like to see is why Battlefront went with such high hit percentages against hull down AFVs. To me (with the aid of some very though provoking and informative posts) it seems ridiculously high. [ September 15, 2004, 08:19 PM: Message edited by: Justin S. ]
  3. Some of the data and conclusions I posted are inaccurate. Although only a few values are off (and only 5-10% at that), they skew the overall percentages *immensely*. "Close" isnt quite good enough here, and I simply did not get enough data points. Later I will post the revisions and numbers that i have been working on. I tested over 1000 hits this week and combined with Treeburst155's data (over 500 hits) very accurate (i promise this time analysis can be made. My apologies for the inital inaccuracies (i plan to put a warning at the top of those posts), there has been a lot of data to sift through, and this has been a giant learning experience for me.
  4. *************EDIT************* The information, and conclusions below are inaccurate due to insufficient data. Please read later posts. There has been a lot of discussion about the PzIV and how the strong hull, but weak turret affects armor engagements. I believe the common perception is that to achieve maximum survivabilty, it is best to use that tank exposed to encourage hits upon the glacis. However, the obvious drawback to this tactic is a higher enemy hit percentage. I havent seen comparible data in reference to the Panther so i thought id run a few tests. I had said before that I wanted to test out all models of Panther; however it quickly became clear that this was not necessary. The only difference in turrets/mantlets is the shot trap, which is very rare and in CMBB unobservable, and on the hull, the armor flaw level (which didnt matter because at my chosen test ranges all upper hulls regardless of quality are impenetrable by the 85mm). Because of these limiting factors, and time, I only tested model G. Before i tested turret resistance i needed to find the hit distribution on exposed Panthers, and the percentage of penetration of all non-turret hits. Rexford has posted some great data on this (dispersion) and it is interesting to see how the data from CM deviates from the theoretical. Heres what I got with 100 frontal hits (1944 85L55 APBC) upon fully exposed Panther at 500m : Lower Hull- 15% (6 penetrations) Upper Hull- 41% (0 penetrations) Turret- 39% (seperate test later) Tracks- 5% (3 Immobilizations) I find the number of turret hits surprisingly high. As rexford showed, that number should be closer to 25%; based on this and other observations i dont believe statistical error is the primary cause for the discrepancy. Now the turret tests... As detailed before, I set up 10, dug in, hull down (behind a rise as well) Panthers, 1000 meters from 10 (veteran) T-34/85 (1944 late). After a few tests it was clear that the kill chance was very low and so i moved the range down to 500 meters (at 1000m less than 20% of hits where penetrating). As stated in the "Theoretical Hit Rates Against Hull Down Targets" thread, CMBB does not model a true 'hull down'. I found this out as well when the Panthers where receiving upper hull hits even when only the upper half of the turret was visible! In fact, nearly 20% of the hits upon a "hull down" Panther where on the glacis... To account for this I counted them seperately as well as together- one to see pure turret resistance, combined to see actual game hull down performance- they have a significantly different effect on survival rate. Results of 200 ( :eek: took a while) frontal hits on hull down Panthers; 500m: Upper Hull 18% (36 hits; 0 pen.) Turret 82% (164 hits; 120 pen. or 73%) Total Hull Down Penetration** rate: 60% (120 total penetrations/200 hits) **Over half of the penetrations where partial; this shows the accuracy of rexfords impact angle distribution chart. This also shows that hull down in CMBB is neither simply "tracks down", or only the turret exposed. First Round Hit Percentages Against Panther At 500m: Hull Down- 47% stated (game calc) - 50% tested Exposed- 69% stated " - 75% tested (only 100 hits so not as accurate) After the third round, the effects of hull down are greatly reduced. By the fourth round, the spread between exposed and hull down percentages are likely less than 10%. First Round Penetration Chance Against Panther At 500m: Hull down: 28% (60% pen. of 47% chance to hit) Exposed: 37% (73% of 39%[turret hits]+6%[lower hull]+3% [immob.], of 69% chance to hit) Although, this data is only good for 500 meters, i think it clearly shows that in the case of the Panther, the advantage of a lower hit probability while hull down, outweights the danger of higher penetration chance per turret hit. This is also due in part to CMBB's modeling of hull down and dispersion; the frequent upper hull hits raise the effectiveness of hull down in AFV's with superior glacis protection, and a higher rate of turret hits increases the vulnerbility of exposure. At longer ranges this advantage only increases. These factors combined with the wide threat angle of an exposed tank vs. hull down are more reasons to adopt a more realistic tactical approach. NOTE: I tested 200 rounds against the turret because for the first 50 I noticed some oddities (a string of 15 ricochets in a row) that i thought would skew the set. After 150 it was regular, and I did 200 for good measure- I think it is a accurate sample group. EDIT: Terrible at formatting. And I added a better conclusion [ September 15, 2004, 02:00 PM: Message edited by: Justin S. ]
  5. It definitely does. There were several tests done, particularly for late Pz IV (80mm front hull armour but just 50mm front turret). After some extensive testing against 76mm T-34s (which can punch through 50mm turret but not through 80mm hull armour of Pz IVs) it turned out that Pz IVs should not go to hull down positions when facing T-34/76s. In other words - tank with vulnerable turrets and invulnerable hull armour should avoid hull-down positions since the kill probability of the hull-down position: - reduced overall hit probability but every shot hits the weak turret is greater than the kill probability in the non-hull-down position: - greater overall hit probability but many of the shots hit the tough front hull armour instead of the weak turret. </font>
  6. Glider, Since youre interested Ill go ahead and run all models of Panthers. I will only use 1944 (late) T-34/85's. First run conditions will be dug in, and hull down behind a small rise (this is because even dug in there are frequent glacis hits which i want to avoid). Starting range is 1000m. I will record first, second, third, and fourth round+ accuracy by the T-34/85's with veteran crew (regular will be too low i think). In addition i will record, of those hits, their effects. Next range will be 500m- same test. After all the data is compiled i should get a curve of hit percentage and penetration percentage of those hits. All models of Panther will be run through this set up. The next test will be an immobilized group of Panthers with full frontal exposure (i realize one could make it complicated by not exposing the lower hull, but im more interested in survivability when fully exposed). All this exposure talk sounds dirty ... So on and so forth. Ill just start a new thread on the Panther while im on a roll...
  7. Id still like to test the survival rate of Panther, hull exposed and hull down, against Russian 85mm. Should I just post it here, or start a new thread? More importantly are any of you interested in seeing the data, or is this just for my own benefit?
  8. First I like to clear up some errors. Going back and reading through the threads on curved armor it appears that all of them tossing around the "30 deg." idea are prior to patch 1.03. Also, i apologize to rexford. He is quite right when he says he never advocated using the 30 deg average- He was posting in the same threads but never gave it his approval. The majority of talk about 30 deg. came from JasonC, whom i typically think is fairly reliable- To his credit this was pre-1.03 and things may have changed greatly, i dont know. However, even if things changed in 1.03, the common perception sure didnt. If you go back and review even this thread, and especially the "Seeking final word on mantlets" thread, I think youll find why i felt the need to test out the effects of curved armor, especially the Hetzer. Redwolf: Soddball: Soddball also mentioned how rare it is to see the TD's being played in CMBB; i would assume some of this can be attributed to the common idea that they are still 'flawed'. Im qouting these guys simply to show that before we did these tests, we had very different ideas about the effects of curved armor and the effectiveness of the TD's, specifically Hetzer. Id also like to summarize some of the points mentioned in this thread that really seem to stand out. Glider really hit it on the head when he reminded me of the color bars in the unit screen. This clearly shows that each AFV's curved armor is modeled differntly. I about smacked myself He also said this Which i think makes a key point (even though its been said many times); The algorythmic models are far more complex than the limited values it shows to the player. This particular example makes me more confident about more "accurate" AFV engagements, and shows what a great job the Dev team has done. Of course rexfords distribution chart really helps one get a better picture of how rounded mantlets work. A big thanks to everyone; I think we have settled the issue of curved armor (for me at least), even if you grogs already knew .
  9. Well, i just spent an hour running the test with over several hundred hits... the results where very surprising and frankly, quite disappointing. I set up the test as i mentioned above, and added the rough terrain to make the ammoless Hetzers stay put. The parameters where as follows: July 1944, Combined theaters, Mid Day, Dry, Cool, Still. Ten Sherman 75L38 firing APCBC down range at Ten Hetzers (all regular crews). All Shermans where 999 to 1000 meters from the Hetzers at 1st go- this is the stated range that the Shermans would penetrate 60mm of armor at 30 deg. from vertical. After watching two full turns of literally NOTHING but upper hull ricochets, i moved the range to 950mm for all the Shermans. Again, same thing. The occasional lower hull hit would cause internal flaking- that was it. I stopped tallying on the paper at this point because out of 57 upper hull hits, 54 where ricochets, the other 3 being gun or track hits. I continued this method all the way down to 100 meters. NO, repeat, NO penetrations occured against the front upper hull of the Hetzers. In CMBB the Hetzer is impenetrable by the 75L38 at *ALL* ranges on the upper hull. Conclusions: As Glider said, the 'curved' declaration may be diferent for each AFV. If this is the case, then the hypothesis for an all- inclusive average generated angle is out the window. Clearly, it is not 30 degrees in the case of the Hetzer. Note that i am assuming that the ricochet property is directly related to angle of impact and is not a seperate varible than can be adjusted. Im curious to see what Soddball's data for CMAK shows (i just realized i had been typing 'sodball' this entire time- sorry). Thanks for everyones input and I have the saved games and scenario set-up if anyone wants a second look. Next up: Id like to test out the Panther's mantlet/turret against late 1944 T-34/85 and see if going hull down is *statistically* advantageous in regards to reducing penetrations. Two main factors will be covered- Hit percentages against Panther (hull down vs exposed), contrasted with percentages of penetration. In other words, does the reduced enemy hit percentage of going hull down in a AFV with a superior glacis produce a higher survival rate than the same tank exposed?
  10. Soddball and I are going to test to see how curved armor affects the general survivability and protection levels of the german tank destroyers in both CMBB and CMAK. I would also like to determine if the *estimate* (by JasonC and others) of the random angle generated by curved armor (30 deg. roughly) is accurate. To do this I will create a flat, open ground map of ten "firing lanes" sperated by a LOS block- Nine ridge lines. In those lanes I will place a Hetzer (60mm front plate 'curved') behind no cover, at range opposite of it, and at the same elevation and aspect angle (to avoid oblique angles confounding data) an allied tank (both will be bogged and only the allied tank will have ammunition- i will only be counting upper hull hits- i chose not to place Hetzer hull down as this isnt a turret/mantlet test, and i didnt want low hit percentages being a problem). To enable a direct comparison with Soddball's experiment (and to try and find a noticable change from CMBB to CMAK) I will initially use a 75mm Sherman (i intend to also test Sherman 76, using varying rounds, and also the T34/85). I will be sure to note date and round type for each test and AVF type. Using the values in the unit screen for which ever AP type it is carrying at the momment, I will place the Sherman at a distance where its maximum penetration value is *equal* to the protection level of 60mm armor sloped at 30 deg. This is easily done because the gun data menu gives values for that angle of impact and i would graph it parabolicaly (with a best fit equation on a TI-83+) to find a more precise intersection point- the point where the effective armor resistance and penetration values are nearly identical. The reason for this is that if 30 deg is in fact the average randomly generated angle, then i would get a rough distribution curve that was measurable by counting the type of hits. If 30 deg is average i should get a significant number of 'partial penetrations', because the armor protection level would be very close to the rounds penetration ability. If the estimate is accurate it should also produce similar numbers that fail to penetrate as those that fully penetrate. Obviously we cant determine by how much the rounds succeed or fail by, but this is a more general test and Ill leave the higher theory and grog work to Lorin or Tittles. I plan to record at least 100 upper hull hits on the Hetzers (from each AFV and ammo type) at the stated range. I will record ricochets (to see if JasonC's 1/6 figure is accurate) round failures (because i wont have the data for by how much ill lump them together- same for full pen.), full penetrations, and partial penetrations. Hopefully at 10 Hetzers per test, this should go reasonably fast. After I get the data from my tests i will post it here in a makeshift, easy to read, 'chart' Combined with Soddballs data not only will we be total armor nerds, we might have established a fairly accurate proof of the average random angle generated by curved armor in CMBB and CMAK. Stand by for the numbers... (it might take a couple of days) Soddball- It might be good for you to conduct the same type of experiment so that we have comparible data. Let me know if you think ive got something wrong, or you disagree. Also any other members are more than welcome to offer suggestions and criticism to testing methods. EDITED to correctly spell Soddball.. 4 times 2nd EDIT: rexford in fact did not put forth the 30 deg. estimate, so i removed his name from that association. [ September 10, 2004, 07:50 AM: Message edited by: Justin S. ]
  11. OK, Ill be sure and do that. To keep this thread more or less on-topic im going to start a new one for our tests of curved armor.
  12. Using search, i found a multitude of threads regarding 'curved' armor, and TD's, most from 2003. However it gets a little tricky because some posts where prior to patch 1.03, and the values for Soviet 85mm rounds did change in that patch. There where also a few threads on specifically the Hetzer (vs. T-34/85). There appears to be some confusion because the numbers change dramatically between games and patches. After reading through posts by JasonC and rexford it seemed to be agreed upon that 'curved' armor generated, on average (rough, not statistical) a 30 deg. slope effect in CMBB. It also recieves a chance to ricochet (roughly 1/6 says JasonC). As mentioned the biggest drawback to this modeling would be the so called "sweet spot"- Where the random angle is very small and acts like vertical plate. Soddy- I cant comment on the possiblity of a change regarding 'curved' effects in CMAK. I dont own it and havent kept up on its modifications. However, i dont think using 75mm Shermans is a good test. According to the data i have the 75L38 penetrates about 83mm at vertical at 500 meters. Say at 700m that figure is around 80mm. Against the 60mm front plate of the Hetzer at range, that gun doesnt have a high probability of penetrating even if lower than average (30 deg approx.) angles are generated. Getting back to my question of why they moved to curved armor, it appears it was a game balancing move. In CMBO I believe the hetzer front hull was rated at 60mm at 50-55 deg or thereabouts. From what ive read it was dificult to kill, especially so when hull down or at angles which aided its slope effect. Dropping that down to 30 deg. on average (if that estimate is correct) would indeed make the Hetzer and other TD/ much more vulnerable (especially so with the chance of much lower values). EDIT: Just to make it clear, i was talking in very general terms in regards to penetration and protection levels. I didnt include round type, year, etc. which can alter figures considerably. [ September 08, 2004, 03:25 PM: Message edited by: Justin S. ]
  13. Well, then that leads me to another question (probably debated and answered a million times). Why do they use 'curved' for the upper hulls of the TD's? Was this due to an enigine limitation or a game balancing decision (ie they where too tough in CMBO)?
×
×
  • Create New...