

Mech Stats Results!
#21
Posted 08 August 2017 - 07:09 AM
#22
Posted 08 August 2017 - 01:36 PM
mogs01gt, on 08 August 2017 - 06:31 AM, said:
You are not wrong. I just didn't feel like combining them manually, its not something I know how to easily do in Excel. The data is available on the Google Sheet link in my OP. By all means, please sort the data by chassis.
Edited by VirtualRiot, 08 August 2017 - 01:36 PM.
#23
Posted 08 August 2017 - 02:23 PM
BLACKJACK BJ-1DC 171 64 98 0.65 57 133 0.43 34,348 104,513 19:21:01
Edit; same for the Cyclops ADC, most seem to be from me.
CYCLOPS CP-11-A-DC 205 114 88 1.30 216 148 1.46 100,226 205,602 24855 days 03:14:07
And yes, that 24855 days is a stats page bug.... and I've played a few matches in it since I sent the report... so ignore those few extra matches over the spreadsheet.
Edit 2; And the Mauler 2P... I would have though more piloted that, but 200 of that 250 is from me... and I haven't played that in many months because I sold it.
Edited by Athom83, 08 August 2017 - 02:31 PM.
#24
Posted 08 August 2017 - 03:23 PM
Athom83, on 08 August 2017 - 02:23 PM, said:
BLACKJACK BJ-1DC 171 64 98 0.65 57 133 0.43 34,348 104,513 19:21:01
Edit; same for the Cyclops ADC, most seem to be from me.
CYCLOPS CP-11-A-DC 205 114 88 1.30 216 148 1.46 100,226 205,602 24855 days 03:14:07
Man, I forgot about this. Had I submit my stats, though, Blackjack playtime would have been skewed with 2176 matches played in the chassis, and 384 in the BJ-1DC specifically.
#25
Posted 08 August 2017 - 04:25 PM
At best this is not representative sample at worst shits bugged yo.
Edited by davoodoo, 08 August 2017 - 04:31 PM.
#26
Posted 08 August 2017 - 04:51 PM
davoodoo, on 08 August 2017 - 04:25 PM, said:
At best this is not representative sample at worst shits bugged yo.
It's the average KDR of people who submit results, not the average KDR of every account in the game. The more accounts you add, the closer to 1.00 you'll get.
#28
Posted 08 August 2017 - 07:19 PM
Athom83, on 08 August 2017 - 05:42 PM, said:
Which should still give us 1.0...
Yeonne Greene, on 08 August 2017 - 04:51 PM, said:
It's the average KDR of people who submit results, not the average KDR of every account in the game. The more accounts you add, the closer to 1.00 you'll get.
Yeah i know, but then that gives unrepresentative sample.
#29
Posted 08 August 2017 - 07:40 PM
davoodoo, on 08 August 2017 - 07:19 PM, said:
Yes, and? OP never claimed it was perfectly representative. The limited info gives us limited insight, which is better than zero insight.
You aren't pointing out anything that hasn't been considered.
4 user(s) are reading this topic
0 members, 4 guests, 0 anonymous users