Author 
Topic 
Kazuo Ishiguro House of Marbles

Posted  2007.04.22 11:17:00  [ 31]
That's a bit pointless, seeing as it has the ability to fit 1 link module per level Note to self  do not fit evasive warfare link to Ragnarok in futureAnyway, corrections have been made and the new sheet is live (via my sig). 
Kua Immortal Garoun Investment Bank

Posted  2007.04.22 22:09:00  [ 32]
Originally by: Kazuo Ishiguro Note to self  do not fit evasive warfare link to Ragnarok in future
Gah! All those times I fitted the EWL to my Ragnarok. Oh the shame! /o\ Nice work, I'll try to look at this in depth when my brain is in the mood. 
Goumindong SniggWaffe

Posted  2007.04.22 23:41:00  [ 33]
Edited by: Goumindong on 22/04/2007 23:39:56 Originally by: Akita T
If anything, the turret "to hit" formula and "damage dealt" formula need a serious, long look and subsequent upgrade to a much more intuitive (one could even say "realistic") style... with the first change being to STOP using "absolute transversal velocity" and distance as base and start using "relative angular velocity" instead.
These are exactly the same thing. Transversal/distance == Angular Velocity You must travel 2pi*r distance to go around a circle in 1 second. If you go around a circle in 1 second your angular velocity is 2pi. 2pi*r/r = 2pi = Angular velocity. There is no difference. Halos give a similar boost for avoiding tracking as if you increased your speed by the inverse of the post halo sig divided by the prehalo sig. This is because ship signature is compared directly to gun resolution and then tracking/velocity is a multiplied by that number. I.E. a 90% reduction in signature radius is the same as going 10 times faster. So the relative tracking difference between two calcs when only the sig changes is the ratio of the signatures. 
Akita T Caldari Navy Volunteer Task Force

Posted  2007.04.23 00:32:00  [ 34]
EVE "transversal" / distance = "absolute" angular velocity. If ship A has an angular velocity of 0.5 rad/sec around me, it's the same for him as if I had a 0.5 rad/sec angular velocity, regardless who is actually circling around, rotating its own ship in the process, and who's just sitting there.
Now, the "proper" way to do this would have been relative to the ship axis, so not an "absolute" angular (derived from transversal), but a relative angular (relative to shooting ship). You could have scenarios in which both ships have a 0 relative angular (both same agility/speed, both orbiting eachother at same distance), you could have a 0 relative angular on one but a high relative angular on the other (on the slow one that's not moving/rotating), or you could have high relative angular on both (flyby scenario). 
Goumindong SniggWaffe

Posted  2007.04.24 18:53:00  [ 35]
Originally by: Akita T EVE "transversal" / distance = "absolute" angular velocity. If ship A has an angular velocity of 0.5 rad/sec around me, it's the same for him as if I had a 0.5 rad/sec angular velocity, regardless who is actually circling around, rotating its own ship in the process, and who's just sitting there.
Now, the "proper" way to do this would have been relative to the ship axis, so not an "absolute" angular (derived from transversal), but a relative angular (relative to shooting ship). You could have scenarios in which both ships have a 0 relative angular (both same agility/speed, both orbiting eachother at same distance), you could have a 0 relative angular on one but a high relative angular on the other (on the slow one that's not moving/rotating), or you could have high relative angular on both (flyby scenario).
Yes, but doing what you want it stupid and would ruin turret balance. And yes, the angular velocities produces are relative to the shooting ship as if it were a fixed point, this is how you calculate angular velocities. You want to add in ship orientation, which is a bad idea, and completly unrelated. 
Kazuo Ishiguro House of Marbles

Posted  2007.04.24 21:19:00  [ 36]
Edited by: Kazuo Ishiguro on 24/04/2007 21:15:05 Please don't derail this thread by discussing suggestions for changing the way that turrets work.
On topic: there are still (valid but pessimistic) doubts among some people as to whether the relation I've found is valid when tracking and sig radius are affecting the chance to hit as well as falloff, so I'd like to do one more test with the coercer, this time against a moving target.
If someone is prepared to fly out to Isinokka and steal a few bookmarks from me, I propose the following test (I don't have a presence on the test server and I only have one account):
Me: Coercer 8x Gatling Pulse Laser I (Gamma S) ColdGas afterburner (less cap) 3x Heat Sink II 1x CPR
Target: Blackbird (chosen for having 6 mid slots)
Tracking disruptor x2 Medium shield booster Boost amp x2 Medium cap booster with 200s
2x Expanded Cargohold (more 200s)
Orbit range: ~ 4000m, within optimal of 4312m (my skills).
Orbit speed  whatever we can keep roughly constant within optimal. I understand that transversal is not calculated vs. stationary targets, so I think the best course of action is to have the target move in a straight line at the lowest possible speed. If I'm wrong about this, we'll work something out there & then.
The low resistances on the target mean that more damage is done per shot, meaning more significant figures  better data, almost as good as shooting at structure. I've tested both ships' cap usage with QF and they seem to hold (good cap skills suggested for the BB pilot). Large shield extenders are not ideal for this test because they significantly increase a cruiser's sig radius.
With that Coercer fit it takes 5 minutes to gather about 1500 shots  my usual quantity for one run at a constant hit chance.
If required, I can provide the cap booster 200s (and even loan a BB)  I'll at least make sure there are plenty available in the system Wink
I'm online in the evenings between 19:00 and 23:00 eve time, though I can probably be more flexible at the weekends. If there are two pilots who could do this in their own time without requiring my presence, that'd work, but this is a bit fiddly and I want to be certain that everything is done correctly. 
Gariuys Evil Strangers Inc. 
Posted  2007.04.24 22:01:00  [ 37]
Originally by: dalman Edited by: dalman on 21/04/2007 23:44:25 As this topic came up again, which is good, I decided to post what I've always thought since it was found that quality was affected by hit chance. http://myeve.eveonline.com/ingameboard.asp?a=topic&threadID=257236&page=3#84
Looking at your numbers here, it looks like I am correct
Originally by: Kazuo Ishiguro http://dl.evefiles.com/media/corp/KzIg/turret_data.zip
So thanks for that test. I'm expecting some small errors in your test, as you're doing it with transversal. That makes it hard to get an exact test environment (easier to know that the game will actually have your calculated hit chance in a test in falloffrange with 2 stationary ships/objects).
Your work on "average" damage dealt gives the same result as my theory.
Your test series is better than the "lacking ones" I based my thoughts on... There's no hits occuring that shouldn't be there. Two thumbs up.
*edit* fixed link
Good thing that finally soemone could be arsed to come up with some good testresults... miss those old discussions though, we known too much these days. 
Akita T Caldari Navy Volunteer Task Force

Posted  2007.04.24 22:05:00  [ 38]
Originally by: Kazuo Ishiguro I understand that transversal is not calculated vs. stationary targets
Anchored, not stationary. Small but vital difference 
Kazuo Ishiguro House of Marbles

Posted  2007.04.24 22:19:00  [ 39]
Edited by: Kazuo Ishiguro on 24/04/2007 22:16:12 Hmm  my overview says that an unanchored container in space has no velocity of any kind  I suppose you mean any anchorable object rather than one that's actually anchored. In short, I still need another player to help me with this one...
//edit: do I have a volunteer? 
Kazuo Ishiguro House of Marbles

Posted  2007.04.26 19:30:00  [ 40]
This discussion has continued for the time being on the Scrapheap Challenge forums: http://scrapheapchallenge.com/viewtopic.php?t=4845&start=15&postdays=0&postorder=asc&highlight=There's still some work to do, on two fronts. One is to run a test in which chance to hit is affected by tracking, and the other is to see if we can work out the point at which turrets stop getting wrecking hits. So far, down to 10% chance to hit (with perfect tracking), it seems that 1% of shots fired will be wrecking hits, but various (hitherto undocumented) reports suggest that eventually no shots will hit at all. 
dalman TunDraGon Pandemic Legion 
Posted  2007.04.27 22:41:00  [ 41]
Edited by: dalman on 27/04/2007 22:49:16Yea, it's a good thing you've spent 3 pages arguing to get to basicly exactly what I had posted already Though I guess you've established through your test series that: 1. wreckings are broken, there is no second roll done  it's 1% of shots, not 1% of hits  despite what the devs said about changing that. 2. that the chance for wrecking actually isn't 1%. By the looks of it, the random number is picked in [0.50, +0.50] or in [0.00, 1.00] (it really doesn't matter, I guess the later is used, I'm not so into the normal random generators used). Anyways, that would mean the chance for a wrecking is 1/101 and not 1/100 if the random number is generated with 2 decimals. Or perhaps it is generated with a ****load of decimals, then that would not be the case, but that would also just cause extra load on the server. Although, with a random number with a ****load of decimals, it should be extremely easy to fix the broken wrecking hits. Just assign the roll R >= 1(hitchance/100) yields a wrecking hit. Done and dusted, wrecking would be 1% of all hits. And no, there's no way in **** that I'm registering for membership in the minmatar forum brigade. 
Kazuo Ishiguro House of Marbles

Posted  2007.04.28 10:02:00  [ 42]
Edited by: Kazuo Ishiguro on 28/04/2007 10:05:01Yep, your old post is very similar to what Goumindong came up with, except that the % of shots that are wrecking hits is capped by chance to hit: Set Rand(0,1) = x Set Hit Chance = z \\ calculated in the usual way involving falloff and tracking
If {x<0.01} Then {Dmg mod = 3} \\This is our crop conditional
If not {x<0.01} Then {Dmg mod = x + 0.49} \\ This is our normal conditional
If {x<z} Then {deal damage to enemy ship = Dmg mod x base damage}
If not {x<z} Then {Generate generic "you missed" message} The tests threw up a 50% minimum and a 149% maximum fraction of normal damage (under perfect conditions), so we are now agreed that this is the most likely way that it could work. The resulting estimate for expected shot damage is slightly more accurate, too: Expected damage per shot = normal damage * [min(chance to hit, 1%)*3 + max(0,chance to hit  1%)*(0.99+chance to hit)/2] Case closed, I think. I'm doing one more test to confirm that tracking has the same effect on damage distribution as falloff (via hit chance), but it seems more like a formality now. 
dalman TunDraGon Pandemic Legion 
Posted  2007.04.28 11:23:00  [ 43]
I don't agree with your/goumin's formula/solution. You've come some way, but I think it's more like my mechanics! This is based on the fact that we have (ofc there's supposed to be some "=" in here):
miss< 0 < close miss < 0.25 < glancing miss < 0.5 < scratch < 0.625 < light hit < 0.75 < hit < 1.0 < well aimed < exellent < 1.49
My assumption that the game generate (hitchance(0,1) + rand(0,1)  0.5)/2 suits perfectly with the above.
Looking at your first big series, it fits perfectly. Goumin's code leaves nothing that sorts out all the missed shots in the 3 categories of misses. 
Kazuo Ishiguro House of Marbles

Posted  2007.04.28 12:36:00  [ 44]
Agreed, but (forgive me if I'm mistaken) don't our respective ideas lead to the same estimate for average damage? 
dalman TunDraGon Pandemic Legion 
Posted  2007.04.28 14:26:00  [ 45]
Yes, they do.
Or, it's basicly impossible for us to establish exactly what the intervals are through testing, making it equally impossible to get the right avarage formula for certain. But the formulas that's been suggested already are good enough... 
Kazuo Ishiguro House of Marbles

Posted  2007.04.28 14:52:00  [ 46]
True; the total damage in the 14k run at 100% chance to hit was within 0.085% of what the final estimate predicted. We have to stop somewhere 
Akita T Caldari Navy Volunteer Task Force

Posted  2007.05.03 10:54:00  [ 47]
No idea where this came from originally, but I recall somebody suggesting a "slightly reasonable" alternative to extreme longrange fighting with blasters. At those ranges, tracking would have no effect on hit chance whatsoever, leaving only faloff (and to a much smaller degree optimal range) influencing hit chance. A large neutron blaster has a base faloff of 10km, say 12km with reasonably obtainable skills. At extreme ranges of say 200+km (and a negligible optimal due to use of antimatter), your hit chance will be something akin to 10^(80), or in other words not bloody likely to ever hit. With insanely maxedout skills and all possible falloff bonuses (L5 trajectory analysis, Zainou ZGC1000 implant, two Hybrid ambit T2 and one T1 hybrid ambit), your falloff would reach a whooping 21735m. To even stand a chance of getting close to 1% on your hits (taking the most "advantage" out of the wrecking shot "loophole"), you'd need to be closer as 70km... and then again, you'd barely get 3% of expected DPS (i.e. you'd need over 30 ships at that range to deal the same damage as a close range one). In other words, practically useless. Just in case somebody gets the desperate idea of trying it out, I'd thought save them the trouble of even trying 