The method of how points (divisions) are attributed to players and how this relates to ELO seems “funky” at best. It is quite unknown (at least to me) how this is calculated. For the love of god we can’t find a logic that seems consistently coherent for our group of friends.
However to paint a outline of the landscape this is what we found: an ELO is present for each game type: 1on1, 2on2, 3on3, and 4on4. We will only focus on ranked matches. The system appears to attempt to pull your division points towards your ELO. So if your ELO is 1.000 and your ranked points is 1.200 you will lose relatively a lot of points for a loss and gain relatively little for a win.
The above is “funky”: Why? Because there is an ELO per game type. A player can have 1.400 ELO in 2on2 and 1.000 ELO in 4on4. That is fine besides the fact that team division-points is a single figure. As a result: if the player just mentioned has 1.300 division rank points then playing 2on2 will for example give him 22 points for a win and take 20 for a loss. However: when this player switches to 4on4 it might only receive 5 points for a win and lose 40 for a loss. Why? Because there is a 300 point variance between his team division points and his 4on4 ELO, whereas this gap was not there when he played 2on2.
In conclusion: someone who plays mostly 4on4 and thus gathered quite some ranked points (and presumably ELO is following suit) may be severely discouraged from picking up 2on2 or 3on3 as his ELO there is still set to default (I presume 1.000?). Thus any win or loss this person gets in those game types might well increase his ELO for those game types but it will severely punish his ranked division points causing him to drop notable in the presented ranking. Note that we see a weird phenomena here: someone increases in ELO over say 50 games but will lose notably in ranked division points.
I am not sure if this is 100% a bug, but this also seems 90% sure not intentional…
A friend had 1310 division points for team games. His 2on2 ELO was 1110 (variance of 200). His division was well high up do to his 3on3 and 4on4 which he played notably more than 2on2.
We started playing 2on2 and played a total of 13 games of which 9 (69%) were won. His ELO went up to 1160 (+50) and his division points are now at 1290 ( -20). This is quite bizarre. With a near 70% win the variance of 200 between division and (2on2) ELO was only reduced to a variance of 130 and he lost division points because of this. So the mechanism as per previous post seems not only in place: it also seems incrementally slow in covering the variance cause the severe skew in points gained or lost to sustain over a severe amount of games.
@SavageEmpire566 Happy NY mate. I dislike tagging devs to gain attention but this is really weird. Is there public information on how this works? Or will this mechanic for points distribution be reconsidered for S4?
I’m not sure I know enough about this to speak intelligently. However, there is definitely a calculation done here that I have a good idea of. It is possible that over the 13 games, based on the team makeup of each match, that your wins didn’t offset your losses. So it is mathematically possible to lose points, in theory, while winning a majority of matches. It’s not ideal and hopefully doesn’t happen often, but it can as far as I am aware.
I think you could be right that this might be part of the problem as to why things seem “funky.” I’ll push this up the ladder with hopes of at least getting some more transparency here. Appreciate your thoughts on this @KiezelSteentjeh !