This Forum has been archivedVisit the new Forums
This is something I wouldn't believe, if I had not seen and tested myself.
Distance effects on Spell Hit Chance
I was trying to find out the movement speed of shamans lightning bolt, by firing it against a Expert's Training Dummy in Orgrimmar. I know exactly three distances: 30, 25 and 0 meters. I am not going here too much into details. My Shaman is Level 18 and the Dummy is 60, so most of the bolts are expected to miss. Indeed, 5 of 159 made damage at 30 meters, 5 of 215 at 25 meters, and, this is the most amazing: 13 of 142 at 0 meters, e.g. standing on the Dummy.
I don't know if the statistic matches the equations, but I guess that the first 2 do, and the third not.
Up to now, I haven't found any other reports about this fact in the internet.
I imagine, that for a same area around the Dummy, where you are able to do melee attacks, the Spell Hit Chance rises. So at 5 meters the statistic is the expected and explained in wow-wiki.
I want to make more measurements that include melee and range attacks, but I would really like to konw, if somebody knows what is happening here.
30 meters distance: 159 data Lightning Bolt Hits : 5 in 4,16 secs in average. Stddev= 0,04 secs Max and Min Values: 4,21 and 4,12 secs Lightning Bolt Miss : 154 in 4,58 secs in average. Stddev= 0,16 secs Max and Min Values: 5,33 and 4,28 secs
25 meters distance: 215 data Lightning Bolt Hits : 5 in 3,87 secs in average. Stddev= 0,11 secs Max and Min Values: 4,01 and 3,69 secs Lightning Bolt Miss : 210 in 4,29 secs in average. Stddev= 0,13 secs Max and Min Values: 4,56 and 4,00 secs
0 meters distance: 142 data Lightning Bolt Hits : 13 in 2,52 secs in average. Stddev= 0,13 secs Max and Min Values: 2,82 and 2,36 secs Lightning Bolt Miss : 129 in 3,01 secs in average. Stddev= 0,14 secs Max and Min Values: 3,34 and 2,58 secs
Analysis of data
The time needed to hit is always lower that the time to miss. This cannot be explained by statistical fluctuations, since the difference of about 0,45 seconds is three times the standard deviation. I tried to imagine that the bolt leaves a circle around the Dummy if it misses. Using elemental geometry and considering a certain angle by which the bolt is fired, most of the data should be around the maximum. But this is not the case. The median is actually nearer to the minimum than to the maximum. If the difference between maximum and minimum would give a hint to the programs main cycle time, this would be about 0,8 seconds, incredibly long. I haven't got a clue how else to visualize it.
programs main cycle
First of all, I don't know how exact my distance to the dummy really is. Considering the Standard deviation, which is almost constant for all distances, one or two steps toward the enemy make no difference. I do not believe that the fluctuations of about 0,15 secs arise from latency, which is about 350 msecs in my case, because I get server times, not local computer times. So, the server calculates how mucht time the lightning bolt needs to reach his destiny. This fluctuations may be originated by fluctuations of the main cycle of the program on the server. The program is very likely to be event oriented, as most of real time games do. So, if there are a lot of events to process in one main cycle, and the time shown ist the one by which the event is processed, this time may vary indeed. Nevertheless, a main cycle time of this magnitude seems to me to be very big. On the other side, there is really a big amount of events to process in a very short time, considering just the movement of chars.
It is interesting to notice, that for distance 0, the Bolt may reach his target even in a shorter time that the expected 2,5 seconds shown for it in the spellbook. I have no talents or glyphs implemented for this char now. So, if the programs main cycle time is i.e. 0,3 msecs, the event timer is set in the program to be 2,2 seconds and not the shown 2,5 seconds; so, the main cycle is early enogh to process the event in time. --SebOeh (talk) 23:13, April 24, 2010 (UTC)ColocoloSebOeh (talk) 23:13, April 24, 2010 (UTC)