Last edited on 16 October 2016 - 01:07 AM by Unwise
A very nice, mathematically sound conversion from elo to MMR. Enjoy.
Last edited on 21 August 2016 - 04:06 PM by Rocktah
This is for everyone who doesn't want to watch the video
This is not 100% accurate
Posted on 21 August 2016 - 04:29 PM
Justie wrote
This is not 100% accurate
Archy has confirmed that it's far more accurate than any other theory suggested right now, but without the actual numbers archy uses to skew things and more MMR theory, this is as close as anyone is going to get.
Posted on 21 August 2016 - 04:31 PM
Unwise wrote
Justie wrote...
Archy has confirmed that it's far more accurate than any other theory suggested right now, but without the actual numbers archy uses to skew things and more MMR theory, this is as close as anyone is going to get.
Posted on 21 August 2016 - 06:53 PM
Automatically DeletedPosted on 21 August 2016 - 09:22 PM
So near impossible to get out of Emerald in any ladder but buildPosted on 22 August 2016 - 12:38 AM
SletursiWantsElo wrote
No, this is for build only. The values are almost undoubtedly different for every ladder.
Last edited on 22 August 2016 - 01:08 AM by majerSNACK
this doesnt seem that accurate imo but more accurate than some I've seenbut if it was confirmed than I guess it's accurate
Posted on 22 August 2016 - 08:57 AM
feel like it should be a bit lower. I got off s12 1500 elo and I'm silver 3Posted on 22 August 2016 - 03:01 PM
I posted these ideas on the video comments but I guess I'll share them here too:This seems accurate, but only if we know that there are equal intervals between ranks. Otherwise, if MMR between ranks is on a curve, then dividing the difference between 1400 and 2700 gives only an average which isn't so meaningful. Was there any post by Archy or Gberry that confirmed that intervals between ranks are equal? Didn't see anything in either the S13 Patch Notes or the 4th Special Dev Log, but maybe it's a fair assumption anyways.
Also, I think that since S13 was so short and also unpopular, fewer people played ranked matches and the MMR pool was smaller than S12's ELO pool which may explain why people weren't as high ranked or why ranks weren't as spread out. For example, most Emerald V players were probably above 1976 in S12.
Posted on 22 August 2016 - 05:48 PM
this doesnt seem accurate because most silver 1's and gold 5's in season 13 were 1700-1800s in season 12 .Posted on 22 August 2016 - 08:39 PM
OuterArts wrote
IronicGolems wrote
Also, I think that since S13 was so short and also unpopular, fewer people played ranked matches and the MMR pool was smaller than S12's ELO pool which may explain why people weren't as high ranked or why ranks weren't as spread out. For example, most Emerald V players were probably above 1976 in S12.
Posted on 22 August 2016 - 09:45 PM
@Unwise, do you know if ranks are equally separated (ex. if it takes equal amounts of MMR gained to rank up from Gold 3 to Gold 2 and then from Gold 2 to Gold 1)?Posted on 22 August 2016 - 10:11 PM
I feel like it was easier to get elo because you could drain a lot of people unlike this season you actually fight people near your skill so it's harder to gain itPosted on 23 August 2016 - 06:42 AM
Unwise wrote
Justie wrote...
Archy has confirmed that it's far more accurate than any other theory suggested right now, but without the actual numbers archy uses to skew things and more MMR theory, this is as close as anyone is going to get.
Posted on 24 August 2016 - 03:56 PM
IronicGolems wrote
I'd assume yes, but I can't say I'm 100% sure.
Posted on 24 August 2016 - 04:03 PM
Unwise wrote
IronicGolems wrote...
I'd assume yes, but I can't say I'm 100% sure.
Posted on 24 August 2016 - 04:12 PM
People whine that they want ELO back, but I doubt they could even code a decent one ^.^Posted on 24 August 2016 - 05:21 PM
Unshift wrote
Unwise wrote...
Was thinking that, and I assume it was the same in S13 as well.
Posted on 03 September 2016 - 05:51 PM
Gold 3 1760? Nice joke