9 replies to this topic

[GeminiWolf]

I would love it if the process for selecting teams included a randomizing algorithm. In effect it would gather the 24 pilots for the match, then randomly place said pilots on either team. It would kill the sync drop.

[Bombast]

...Isn't that pretty much exactly what the MM does now?

[justcallme A S H]

Bombast, on 13 February 2019 - 09:04 PM, said:

...Isn't that pretty much exactly what the MM does now?

Yup. No idea what he's talking about with sync drop, it's random already... Sigh.

What it needs to do is SORT them after it finds 24. At least TRY to balance...

Edited by justcallme A S H, 13 February 2019 - 09:29 PM.

[RickySpanish]

You would think so, but actually their current code is:

if ( (index = m_players->indexOf("GeminiWolf")) >= 0)
m_teamOrganizer->RandomizeThenScrewPlayerAt(index);
else
m_teamOrganizer->Randomize();

[Prototelis]

lawl, sync drops aren't even really a thing in SQ. And even when its a thing it isn't a big deal.

[Variant1]

What mm needs to do is distribute average match score between 2 teams instead of randomly throwing them together, that how alot of stomps occur in this game. I also find that mm doesnt distribute loadouts evenly, sometimes there will be a team of high ppfl up against a team with regular builds ex: ppc/gauss, vs med laser mix builds

[GeminiWolf]

RickySpanish, on 13 February 2019 - 09:32 PM, said:

You would think so, but actually their current code is:

if ( (index = m_players->indexOf("GeminiWolf")) >= 0)
m_teamOrganizer->RandomizeThenScrewPlayerAt(index);
else
m_teamOrganizer->Randomize();

[Prototelis]

You don't want a matchmaker that only uses one metric.

A matchmaker that only relies on AMS would suck. Theres a lot of stuff going on at the higher end of the AMS scale, it can be farmed. Not everyone up there is there because they're good at the game.

[Shredmetal]

I started trying something a few months back, and blocked all the bad players I meet. Usually, the MM seems to split blocked players fairly evenly. If it doesn't, the side with the fewer blocked players usually wins. There might be a hidden average match score team splitting that we're not being told about.

[Curccu]

GeminiWolf, on 13 February 2019 - 08:57 PM, said:

I would love it if the process for selecting teams included a randomizing algorithm. In effect it would gather the 24 pilots for the match, then randomly place said pilots on either team. It would kill the sync drop.

Isn't that pretty much how it works?

You can sync drop with your friends and might get into same match and if you do it's pretty much 50/50 that they are in same team or enemy team. I'd even claim that most often they are in enemy team.