Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[umiman]

Actually, I was referring to my own experiences testing with the full AD runepage. You were probably here, when I was telling everyone how awesome it was and how people should stop using ArPen.

From my experience, level 1 damage right now, from highest to lowest would be something like:
Sivir > Ashe > MF > Ezreal > Graves > The Rest

And in endgame, level 18 with IE, BT, and PD, it would be something like:
Vayne > Twitch > MF > Kog > Trist > The Rest

I haven't actually done any math (other than asking my friend to stand still while I whack him with a full AD combo), but that's my metric. For endgame, I go by who has better steroids.

[penguinofhonor]

Yeah, you seem right. I just wanted an excuse to mathcraft a bit. Everyone's going on full AD pages now, by the way.

[umiman]

Heh, people should listen to me more.  Then everyone here would always be ahead of the curve~

If you're curious, my AD page is AD red, yellow and quints with crit blues. I get about 3% crit. I figure I'd rather have that crit than the extra few AD since no one ever has any crit at level 1. 3% chance is infinitely better than 0%. Has helped me out quite a lot actually... which is a bit odd.

I found that statistically, I can actually rely on the 3% crit chance to proc on my opponent. I'm not sure why, but it does. I figure it has something to do with the algorithm Riot currently uses to even out crit chance.

Or maybe I'm just imagining things. But it's been more than three times now that I've gone into a do-or-die duel at bot lane and *boom* random crit. The odds of that happening should be 3 out of 100. Happening repeatedly? That's some bullshit luck right there.


Edit: Is it actually bullshit luck? I'm kinda bad at probability, but let's see. Correct me if I'm wrong okay.

If there are 5 autoattacks. 3% chance each to crit. It's a total of 15% chance total that I'd get one crit. In other words... In every early level fight I had a 15% chance of a guaranteed win. Huh... that's kinda good isn't it? I'd take those odds for a 3AD trade.

[penguinofhonor]

Edit: Is it actually bullshit luck? I'm kinda bad at probability, but let's see. Correct me if I'm wrong okay.

If there are 5 autoattacks. 3% chance each to crit. It's a total of 15% chance total that I'd get one crit. In other words... In every early level fight I had a 15% chance of a guaranteed win. Huh... that's kinda good isn't it? I'd take those odds for a 3AD trade.

You're wrong, sorry. Assuming that Riot's no-crit-streaks algorithm isn't going to let you get multiple crits on a 3% chance in five autoattacks (because I've got no clue how it actually works), your chances of getting a crit in early fights are .03*.97*.97*.97*.97, or 2.66%. Add maybe like .1% to that for the incredibly smaller chance of getting multiple crits in the five autoattacks. You can't add the percents together, you've go to do some multiplication.

Now, it's true that lucky crits are going to give you a chance of a decent advantage in early fights. I typically don't like the "any chance at all is better than no chance" argument because I really think +3 AD is going to get you more consistent damage over the course of the game than 3% crit chance, but I have to recognize the power of the occasional early crit when harassing. It might not be better over the course of the lane, but when concentrated into a single burst it gives you some pretty good (if temporary) tactical advantage.

[umiman]

Multiplying percentage together is used for determining percentage chance of something occuring in a row. In other words, the chance of getting 6, 6 times in a row is 1/6*1/6*1/6*1/6 etc.

This is the percentage chance of something happening with 6 rolls of the dice. So addition is used. 1/6 + 1/6 + 1/6 + 1/6 etc.

In other words, the odds that I get a crit if I autoattack increases the more often I autoattack. It doesn't decrease since that would make no logical sense. All I need is ONE crit to win a fight. I don't need 5 in a row.


Edit: You made me doubt myself, so I checked Wikipedia.

Independant probability:
"for example, if two coins are flipped the chance of both being heads is "

Mutually exclusive:
"For example, the chance of rolling a 1 or 2 on a six-sided die is "

[lemon10]

No, your wrong, it does decrease sequentially.

Lets say you have a 3% chance of critting, and you attack 100 times, does that mean you have a 300% of critting at least once? Of course not, the chance of close to 100% but not actually at or above it.

But penquin is getting his math wrongcalculating something different as well (using somewhat bizarre math) the chance of at  crit in 5 hits (at 3%) is 14.1%, not 2.66%

EDIT: I just realized what penguin was doing, calculating the chance for getting exactly one crit, not calculating for at least one crit.
EDIT2:

Or maybe I'm just imagining things. But it's been more than three times now that I've gone into a do-or-die duel at bot lane and *boom* random crit. The odds of that happening should be 3 out of 100. Happening repeatedly? That's some bullshit luck right there.

I think whats happening is that when you don't get the crit you don't notice it (since your chance is so low), but whenever you do get it you file it in your memory and go wow I got really lucky.
Or alternately, you just do get crazy lucky.

[umiman]

Mrm, so what's the formula by the way?

[Sirian]

Mrm, so what's the formula by the way?

It's (1-(0.97 ^5))*100 = 14.12659743% to hit at least one crit in 5 attacks

The logic behind it is : what is the chance of NOT critting after 5 times ? well, you'd have to be under the 0.97 the first time, then against the second time (0.97*0.97), then the third time ... and the 5th time you have (0.97*0.97*0.97*0.97*0.97) chance not to crit. Then you do 1- that number, and you get the chance to crit. (and multiply by 100 to get a percentage)

[umiman]

Yeaah... it's been just under a decade since I did actual probability calculations. I really can't remember anything about it.

Still, 14% chance of winning an early level fight that could easily determine the rest of the game. It's basically insurance (since why would I be fighting to the death anyway if I thought I was going to lose?).

I think that's good for the price of blues. I never liked MR on ranged AD anyway. Didn't click with my philosophy of positioning myself away from all threats.

[Sirian]

Ah also, you're right to say that

[..]it's been more than three times now [..]That's some bullshit luck right there.

The chance of it happening 3 times in a row are (0.141(..)^3)*100 = 0,2819114%, and only 0,0398245% (~1:2511) chance for it to happen 4 times in a row. Well, IF you're autoattacking exactly 5 times each fight.

edit : corrected

[Jopax]

Ziggs sounds quite fun and awesome, might get him after his free week

[Cthulhu]

Oh jeez, I can't believe I'm doing this.  I'm playing.  Not pvp yet.  I'll probably play some smurf games to get reacquainted if I'm gonna do that.

I'm confused though.  The current most up-to-date malzahar guide has Archangel's Staff as a core item.  This scary and disturbing.  Is it true or is this a bad guide?

[Errol]

It is most definitely not true. Do not fall for the trap.

Go for a couple of Dorans into a Deathcap. At least that's the caster cookie cutter build last time I checked.

[Cthulhu]

That's waht I thought but I wasn't sure if it had changed at some point.

21/0/9 mages are as hilarious as I'd assumed when I first saw the new tree.

[penguinofhonor]

Yeah, okay, my math was wrong, but I still sort of think your guys' is too. Shouldn't the probability of at least one crit be this:

(0.03*0.97*0.97*0.97*0.97)+(0.03*0.03*0.97*0.97*0.97)+(0.03*0.03*0.03*0.97*0.97)+(0.03*0.03*0.03*0.03*0.97)+(0.03*0.03*0.03*0.03*0.03)

As in, the chance of one crit, plus the chance of two crits, plus the chance of three crits, plus the chance of four crits, plus the chance of five crits. I think this should be significantly less than 14%. Google is telling me this is about 2.7% chance total, meaning I was right in my prediction that the chance of getting multiple crits was basically negligible compared to the chance of exactly one.