My argument was to just show the inaccuracy of stats as well as how simple it is to abuse compared to attempting to just win the match. Fine you want to be stubborn then I will just keep picking apart the issues for you.
Bad players spam ability's. This is a realistic statement right? Bad players don't manage MP well and constantly cast skills when they are off cd.
good ones save them for when they are useful. Your subjective Statistical system needs to tell the two apart. That requires the context of each game. Good players are expected to evade abilities, expected to use them at the right time. Newbies will burn their MP on minion waves and are trigger happy with spells the moment they see a window to land their combo.
2 newbies fighting will have super unpredictable stats.
Yes. A better player will generate stronger stats vs weaker players, but their stats drop down when vsing some one around their skill. This has to be accounted for. So again, your talking about a subjective system and that system fails if its not considering each and every individual match as a separate event because stats are only as valid as the context of the game that generates them.
If it cant tell a good player from a bad one, when stats are padded, and does not determine if something resulted in any impact on a win or loss what exactly makes it remotely usefull for grading skill?
I don't know how to help you if you don't understand this. Im running out of ways to try to get this point hammerd into your head here buddy.
iron players can get an S and Diamond players can get a C.
Your complaining about the gap of skill to me for a bronze player and a gold in this thread so should be clear mismatches Fk values of stats up every time theres not a close skill range.
There's no crystal ball that knows specifically how good some one is with 0 games
I really hope you don't think that an Iron S is the same as a Diamond S but that mastery System you talked about is still is giving both players an S.
Please, please think for a second and consider how stats are made and how freaking easy it is to just pad them if that's the single metric for how you climb with no consideration for the outcome of a match (Win or loss) to generate context.
Look, i do not think you are an idiot.So then your likely going to "well just make it react to wins or loses and change the # of points". Ok... so basically then you have to admit. T**he single best metric of skill is wins and losses in relation to the context made by who your playing vs?
**
Gee goly gosh that sounds... Like an MMR system! You can tack on stats, they may or may not accurately reflect any impact. You want to make a fancy system that tracks it accurately?
You cant. If stats are only in the context of each game they need something that accurately can judge stat impact exclusivly to the match. So Lacking the context of each game stats are generated by the constantly fluxuating player base. # of players using champions is something you can not predict and the values shift constantly. Its a failed idea. cant be used.
Also even if you did have static values and you want to start calculating stats by match ups? That looks like N! / ((N/2)! X 2^(N/2))
Sorry. Boards does not express math formulas very well.
To be fair... I might have this wrong. If so by all means anybody can correct me on how I am calculating my possible Pair combinations. What I have used for Binary yes or no odds before so I do honestly believe it to be accurate in this context. If not I will adjust to a better formula if it gets provided.
rounding off the .5 2^72 by the way is 4,722,366,482,869,645,213,696. That's called a Sextillion and I literally had to google it cuz I had no idea WTF it was. I tried to do the math but its numbers to large for the calculators I have access to Got Erros for days running N=145.
Assuming it was something like only 1 billion possible combinations and I had a good computer program that could calculate to 100% skill to statistical relevance (again literally impossible due to constant fluctuations in averages generated by player %) each matchup in exactly 1 second. Id only need to let the program run for... well... 31 years, 251 days, 13 hours, 34 minutes, 54.7843 seconds. If this is in the trillions it adds a little bit more time and just needs to run for something like 32,000 years.
So... Again. STATISTICAL SYSTEMS NO GOOD. Get it? NOT GONNA WORK. Have a fucking realistic expectation please.
Now your other system? Just use ranks?
how the hell do you think ranks are made?
I know 2 ways. 1. Population. 2 static MMR values that are literally inaccurate with high uncertainty.
Populations just a glorified leaderboard and its garbage for 5v5 nature. You can do it for 1v1... but well your grind to the top increases directly in relation to how many active players there are who started befor you. So... that's trash. You also sink if your not constantly playing so you better be grinding all the time. Also its RNG on if your able to climb because your not adding to who evers at the top of the game, they only stand to lose position.
unless its not position and you use a "points" system that's directly related to... winning... losing... and who your playing with and Vs?
Sounds like something I keep talking about here yes? Once again. MMR is the better option.