Willard Phule, on 08 August 2014 - 06:28 AM, said:
Faulty information gathered is still faulty. Using Elo for ANYTHING is just ridiculous.
I'm sure you have a readily implementable system to use to replace Elo?
Willard Phule, on 08 August 2014 - 06:28 AM, said:
Players are given an artificially inflated Elo score to put them in the mid-range as soon as they finish their first 25 matches. They did not earn that Elo score by working their way up, it was handed to them. Using Elo for anything at all is, in itself, skewed.
I'm pretty sure you argued the exact opposite in a previous thread... Which do you believe?
Willard Phule, on 08 August 2014 - 06:28 AM, said:
Sure, the average Elo score between IS and Clan was comparable.....but, so is the Elo score between someone that just finished their first 25 rounds and someone that's got 2000 matches under his belt and still happens to be in the mid-range. They are NOT equal in skill. The guy with 2000 matches actually earned that Elo score and has an accurate representation. The "Elite Veteran" of 25 matches....not so much.
Which is why they probably did more than 1 Clan v. IS test, yes?
Willard Phule, on 08 August 2014 - 06:28 AM, said:
IF YOU CANNOT REPLICATE THE RESULTS OF A TEST, THEN THE THEORY IS FLAWED. I don't know of any other way to explain that. No, we do not know the results of the first test. We do, however, know the ultimate result of the second one....Paul's nerf bat. So, using simple deduction, we can safely speculate that the first test did NOT end up as 90/10. If it had, we'd have already felt the nerf bat. Again, I can't make that any simpler.
Actually, it just means you haven't done enough experiments yet. 2 experiments can't lead you to say, "This is false". The fact that the nerf bat is coming out after the second test would seem to indicate that the second test mirrored the results of the first test. Apparently 2 experiments were enough to convince PGI that serious changes need to be made between Clan and IS Tech.
You claim the result of the second test included Clan v. Clan which altered the win/loss percentages even though the express purpose of the test was to see how each Tech type did against each other. If PGI was so efficient that they would hand out the nerf bat after the first experiment, why would they be inefficient as to not separating IS v. Clan results and Clan v. Clan results? You can't have it both ways to suit two differing views.
Willard Phule, on 08 August 2014 - 06:37 AM, said:
Learn to differentiate between quote bubbles.
Edited by Be Rough With Me Plz, 08 August 2014 - 06:47 AM.