What could be happening in this odd (apparent) bug situation in Overwatch?

There’s already an Overwatch thread but this is kind of a separate topic from just the game itself so I hope it’s okay I started a new thread–I’m interested in hearing from people who may be able to offer educated guesses about bug finding in general as well as those interested in Overwatch specifically.

So basically, what’s going on is:

Blizzard has stated that at the end of a Competitive Overwatch match, each player is to receive or lose a certain amount of ranking based on win/loss and some other factors, and NONE of those factors is (according to Blizzard) the presence of a “leaver” on either team. (A “leaver” being someone who left the match early.)

Many, many people on their forums and on other related fora are reporting, instead, that they are finding that IF there is a leaver on either team, THEN each member of the winning team invariably receives a minimally tiny amount of ranking. Meanwhile, if there were no leavers, the winning team receives generally larger amounts of ranking.

Blizzard looked into it, said yes there was a bug, and then they sent out a patch to fix it.

Players reported that, even after the patch, exactly the same issue was occurring.

Blizzard said, in response, that all of their data indicates that players are receiving the correct amount of ranking, that the presence of a leaver has no effect.

Players continue to report that this is not correct.

I have, in fact, not seen EVEN A SINGLE report that it is working, even though it’s all over the forum so if someone who reads the forums had seen it working right, surely they would have piped up.

And it is in fact true in my own games (and I’m beginning to make recordings to document this) – the presence of a leaver leads inevitably to a minimal ranking gain if my team wins (no matter which team the leaver was on), and meanwhile, that minimal ranking gain seems never to happen so long as no one left the match.

So, I’m really puzzled by this. Blizzard seems to be saying their own internal data are contrary to what, apparently, the actual users are reporting as an exceptionless occurrence.

Where could this disconnect be coming from? Blizzard’s latest communication said they’re trying to figure out why there seems to be this disconnect between what they’re seeing and what players are reporting. They haven’t explained, though, exactly what they’re looking at when they refer to their “internal data,” and I’m frankly mystified. Surely they can just look up some games that had leavers in them and just look and see how much ranking was awarded to the winning team, and like, just hand-calculate how much they should have received! If that’s impossible, surely they can just “listen in” on ongoing matches til they discover some that fit the criteria and see what happens.

So what do you guys theorize about this? Where could the disconnect be coming from? What could Blizzard be looking at that is giving them the idea that everything is working correctly, and why are they not simply either finding the wrongly-handled matches and focusing on those or else explaining how they are in fact (contrary to what seems intuitive) working as they’ve stated things are intended to work?

The simplest and most likely explanation is confirmation bias: People are paying attention to the change in their rankings when there’s a leaver, and not when there isn’t.

In other words, don’t look for what’s throwing off Blizzard. Look for what’s throwing off the players.

I would normally think so, and did think so at first, but like I said: Literally not even one single person has piped up to say “well, it worked for me in this instance,” and in my own games (which again, I’m now recording to document) the rule does in fact hold, exceptionlessly.

The few times I can remember Blizzard being confused about sometimes resulted from them testing portions of the code in isolation rather than the entire system.

Though from what I see people tend to leave stompy games and stompy games usually mean one team was so favoured that the MMR gain is low. Reddit flips over non-existant issues way more often than Blizzard is wrong.

Blizzard says (about an hour ago) that they’ve finally located the problem and have put in a “probable solution” that will propogate through the servers over the next 24 hours.

Still really curious to know what happened!