Ledge Statistics S04
" This is what happens when you combine excessive random number generation (items, spawns) with a small overall sample size (total EL events, average participation level). IGN: KoTao Last edited by KoTao#4717 on Oct 14, 2013, 12:01:48 AM
|
|
" I think Kotao's point was that of the 4 classes that can cleave effectively - Ranger-Duelist-Shadow-Templar- the differences amongst the top results come down to pure RNG of zone type. "When I close my eyes, I see this thing, a sign, I see this name in bright blue neon lights with a purple outline. And this name is so bright and so sharp that the sign - it just blows up because the name is so powerful... It says, "Diamond Supporter."
|
|
" It does have some effect, but there were 15568 participants in the result pool, of which only 432 were level 1. The sample size is large enough that aggregated random effects tend toward an expected value. Of the Cleave classes, the largest difference is between Ranger and Duelist. A winning Ranger had an average 20% more experience than a winning Duelist. That's statistically significant. |
|
|
cleave is not dominating this mod because of the way you overlevel each zone, which is huge for spells, and then there's the monster density/variety which makes it impossible to aoe with a melee skill. It's also safer to use a trap against lightning thorns or for permanent strafing if you get curses and mages early.
| |
|
witch op xDD
http://www.twitch.tv/rhoxxx
follow the stream :) |
|
" Another insightful comment! Thank you Rhox. twitch.tv/nugiyen (Live almost everyday from 15.00 +1 GMT)
follow @nugiyen on twitter www.youtube.com/nugiyen |
|
" Do you have rolled mob statistics? If you see EL rankings by participations there you will see some players have only 1 good rankings and the next race they died 1-2 times. Create statistics of users with a 100% survival rate if you believe it's skill and class based. On my first EL(was possibly turbo but I forgot already) I think I died twice because the second area was a knockbackfest. The second time I reached the level 13 or 15 area without resists and elemental weakness elementals with tons of elemental damage mobs. In short. Trigger Elementals no more charges. Coldsnap+ele weakness frozen for several seconds some mobs hit me= dead in less than 2 seconds. Another EL I met some flicker Discharger champs(I was forced to level up that area) with "gains power charge on hit" so I had to run away because I would certainly get shocked at some point and the next discharge would kill me. " You are ignorant if you believe all winning rangers use cleave. RoA PA FT is also a really effective Ranger build. You can be certain that most winning rangers used a bow to get expierence in harder zones. |
|
|
Actually, PolarisOrbit is right, statistically speaking, you have to assume - ceteris paribus - that on average all players across all classes have similar RNG. You cannot assume for example, that people who play shadow have worse RNG than people who play Marauder (bad example; let's say duelist instead). There simply is no causality, at least as far as we know.
When you look at probability theory, more specifically the Central Limit theorem, it clearly proves that these statistical extremes people are arguing for here are less likely. Put simply, when you roll a dice ten times, the minimal value is 10, the maximal value is 60, given the CLT however your statistical average will most likely be somewhere around 25-35 most of the time. That is, because no number has a higher probability of being rolled than any other. Granted the lower the sample size, the more unlikely this is to be true, but that is a seperate issue, personally I think the sample size is large enough. What's more likely to be the reason for why Rangers did so much better than all the other classes is the fact that some of the zones in Endless Ledge are simply impossibe as melee. I tried both, ranger and melee (duelist and templar with cleave), unlike in normal races, in Endless ledge you would sometimes hit a wall of shit as melee that is simply impossible to survive. Whereas with ranger you can kind of "cheese" your way through every zone. With that said, bow ranger is obviously alot worse than melee in normal races, so it's a moot point, because Endless Ledge was obviously never meant to be "competitive" or balanced in the same sense that signature races or Demi races are. edit: I can only tell you my personal experience, but in the one endless ledge in which I came 2nd overall, I went RoA Ranger, not cleave ranger. With templar I never even got close to such a good result, simply because melee in EL is a lottery, if you get a bad zone you are dead (or at the very least your kill speed is slowed down significantly). With RoA I got several bad zones but just off-screened all the mobs. #1 Victim of Murphy's Law. Last edited by SlixSC#6287 on Oct 14, 2013, 7:20:16 PM
|
|
|
Thanks SlixSC, that is a much better explanation than I could have written.
Honestly, I read Hilbert's post and decided I couldn't even reply to it because I had no idea where he was going with it.
Secret message for the Hilbert
The only reason I called them the "Cleave classes" was because of Morsexier's claim that those 4 were distinguished only by RNG. I do take some responsibility in that I never explicitly stated the conclusion that there had to be more to it than just Cleave for a difference of 20% experience. Is this not the same thing that the Hilbert was trying to prove?
|
|
" Don't worry about it, I think Hilbert adds alot of entertainment value to this sub-forum and sometimes I even agree with his arguments, but most of the time his arguments just seem.. incoherent.
Spoiler
I love you Hilbert
#1 Victim of Murphy's Law.
|
|
