Truth behind malak's event

"No_Try;c-1913945" wrote:

"Drazhar;c-1913864" wrote:
Releasing Malak before this round of championship would have simply screwed up the matchmaking, which is, from what we saw, most likely based on variables like spending: y/n, how much spent, on what spent (Malak, first DR event) and so on and so forth. It's definitely not based on GP or what toons you have, as we clearly saw from the generalised complaint. Releasing Malak before the GAC would have brought so much complaint from whales getting defeated by f2ps that got Malak the second way around that they decided not to release him. It would make them lose some revenue by people that could see this as a deterrent, indeed. They're just doing all they can to sneak in whales' heads the solid belief that spending is right and that it does well in the game. They're just waiting for whales to get an actual edge even on other Malak owners (g13 Malak, g13 in general, other premium stuff they will release shortly and so on).

Did we clearly see that? Didn't we clearly see from bots that every matchups also fits the ingame announced pertinent gp too?

Oh, does it? Matching a 2.8M GP player with a 3.5M one is just fair then, is it? Clearly based on GP, that's clear. These two GP will surely also have the same pertinent GP. Guaranteed.

"Drazhar;c-1914018" wrote:

"No_Try;c-1913945" wrote:

"Drazhar;c-1913864" wrote:
Releasing Malak before this round of championship would have simply screwed up the matchmaking, which is, from what we saw, most likely based on variables like spending: y/n, how much spent, on what spent (Malak, first DR event) and so on and so forth. It's definitely not based on GP or what toons you have, as we clearly saw from the generalised complaint. Releasing Malak before the GAC would have brought so much complaint from whales getting defeated by f2ps that got Malak the second way around that they decided not to release him. It would make them lose some revenue by people that could see this as a deterrent, indeed. They're just doing all they can to sneak in whales' heads the solid belief that spending is right and that it does well in the game. They're just waiting for whales to get an actual edge even on other Malak owners (g13 Malak, g13 in general, other premium stuff they will release shortly and so on).

Did we clearly see that? Didn't we clearly see from bots that every matchups also fits the ingame announced pertinent gp too?

Oh, does it? Matching a 2.8M GP player with a 3.5M one is just fair then, is it? Clearly based on GP, that's clear. These two GP will surely also have the same pertinent GP. Guaranteed.

I went up against a guy that has 600k more GP, 5 more zetas, 6 more G12, 3 more G12+ pieces, and 13 more +15 speed mods than me. I lost by only 6 banners. I didn't play as efficiently as I could of. So to make things clear, just because your opponent has that much of a GP, zeta, G12, G12+, speed mods difference than you, doesn't mean you should immediately crawl into the fetal position and complain. As long as you try, and play smart, you can win.

"Drazhar;c-1914018" wrote:

"No_Try;c-1913945" wrote:

"Drazhar;c-1913864" wrote:
Releasing Malak before this round of championship would have simply screwed up the matchmaking, which is, from what we saw, most likely based on variables like spending: y/n, how much spent, on what spent (Malak, first DR event) and so on and so forth. It's definitely not based on GP or what toons you have, as we clearly saw from the generalised complaint. Releasing Malak before the GAC would have brought so much complaint from whales getting defeated by f2ps that got Malak the second way around that they decided not to release him. It would make them lose some revenue by people that could see this as a deterrent, indeed. They're just doing all they can to sneak in whales' heads the solid belief that spending is right and that it does well in the game. They're just waiting for whales to get an actual edge even on other Malak owners (g13 Malak, g13 in general, other premium stuff they will release shortly and so on).

Did we clearly see that? Didn't we clearly see from bots that every matchups also fits the ingame announced pertinent gp too?

Oh, does it? Matching a 2.8M GP player with a 3.5M one is just fair then, is it? Clearly based on GP, that's clear. These two GP will surely also have the same pertinent GP. Guaranteed.

It can be fair or unfair, depends on other parameters, on principle alone it's very fair if their important parameters matches. And I'm not even claiming the new MM is good whatsoever. Just wearing the tin foil hat regarding your crazy -theories- based on -facts-.

"Drazhar;c-1914018" wrote:
Oh, does it? Matching a 2.8M GP player with a 3.5M one is just fair then, is it? Clearly based on GP, that's clear. These two GP will surely also have the same pertinent GP. Guaranteed.

The first matchmaking algorithm made total GP vs total GP the primary indicator, whether or not ships were a part of the GA. This had much outcry because ships were unusable yet counted against you. The most common user suggestion was to exclude ship GP if ships aren't a part of the GA.

The second matchmaking algorithm made character GP vs character GP the primary indicator when ships were not a part of the GA. This had much outcry because it counted useless fluff at the bottom of the roster and encouraged practices like mod stripping. The most common user suggestion was to compare the top of the rosters based on how many teams you place to weed out all of the fluff.

Now the third algorithm DOES EXACTLY WHAT WE ASKED. It compares top of roster to top of roster. Why is matching a 2.8M GP player with a 3.5M player fair? Because the top of their rosters are similar and the bulk of that difference is probably a combination of ships, fluff at the bottom of the roster, mods stored on irrelevant toons, and random abandoned G7-8 toons.

And just like the last two algorithms had breakdowns, this one does too. But it's not 2.8M players facing 3.5M players. It's when you get up to that 3.5M+ range where that pertinent GP range is pretty much maxed and the difference becomes a ballooning roster of powerful toons below the cutoff. A valid criticism, but that will be iterated on in future takes on the matchmaking algorithm.

The matchmaking algorithm as it presently exists works as advertised. It's not that it doesn't care about GP. Rather, it interacts with GP in a sensible way that has its own issues but works better than comparing pure total GP to pure total GP.

"YaeVizsla;c-1914109" wrote:

"Drazhar;c-1914018" wrote:
Oh, does it? Matching a 2.8M GP player with a 3.5M one is just fair then, is it? Clearly based on GP, that's clear. These two GP will surely also have the same pertinent GP. Guaranteed.

The first matchmaking algorithm made total GP vs total GP the primary indicator, whether or not ships were a part of the GA. This had much outcry because ships were unusable yet counted against you. The most common user suggestion was to exclude ship GP if ships aren't a part of the GA.

The second matchmaking algorithm made character GP vs character GP the primary indicator when ships were not a part of the GA. This had much outcry because it counted useless fluff at the bottom of the roster and encouraged practices like mod stripping. The most common user suggestion was to compare the top of the rosters based on how many teams you place to weed out all of the fluff.

Now the third algorithm DOES EXACTLY WHAT WE ASKED. It compares top of roster to top of roster. Why is matching a 2.8M GP player with a 3.5M player fair? Because the top of their rosters are similar and the bulk of that difference is probably a combination of ships, fluff at the bottom of the roster, mods stored on irrelevant toons, and random abandoned G7-8 toons.

And just like the last two algorithms had breakdowns, this one does too. But it's not 2.8M players facing 3.5M players. It's when you get up to that 3.5M+ range where that pertinent GP range is pretty much maxed and the difference becomes a ballooning roster of powerful toons below the cutoff. A valid criticism, but that will be iterated on in future takes on the matchmaking algorithm.

The matchmaking algorithm as it presently exists works as advertised. It's not that it doesn't care about GP. Rather, it interacts with GP in a sensible way that has its own issues but works better than comparing pure total GP to pure total GP.

I think the most common suggestion on the first MM brainstorming period was to look at multiple parameters in comparison. Whether we can agree on what those parameters should or should not be is another matter. But this one like all other brute force approaches are bound to fail because of that simple reason, it doesn't consider any of the intricacies of the game. And GP itself as a faulty measure couples with that fact.

You guys are still waiting for Malaks 2nd return?

"DarjeloSalas;c-1913804" wrote:
Reading these threads is vastly more boring than the game for me.

Then why r u here?????

"YaeVizsla;c-1914299" wrote:

"No_Try;c-1914128" wrote:
I think the most common suggestion on the first MM brainstorming period was to look at multiple parameters in comparison. Whether we can agree on what those parameters should or should not be is another matter. But this one like all other brute force approaches are bound to fail because of that simple reason, it doesn't consider any of the intricacies of the game. And GP itself as a faulty measure couples with that fact.

"Multiple parameters" is not a suggestion. It is a lack of suggestion. The algorithm does look at multiple parameters already. But there's gotta be a weighting to them, and it's hard for us to assess on anything but the one principal parameter. And there always has to be a principal parameter. In this case, we're using pertinent GP.

What else does the matchmaking algorithm take into account? We do not know. And while you say pertinent GP is a poor measure, people cry foul because the most visible measures don't match up one-for-one, even things like number of zetas and number of G12s and number of +X speed mods (regardless of whether or where they're placed) which are also poor indicators in and of themselves, and become highly gameable if those are the principal matchmaking criteria. Or they point to, "Does my opponent have this one specific character I do not have?" as the principal indicator of whether or not a match is fair, which would be extremely gameable if it were made a hard line in the algorithm.

While GP is not a great measure, it's more difficult to game to such an extreme degree as some of the other suggested measures especially as the principal criteria.

There are improvements that can be made to the matchmaking algorithm and the GP formula. But it's been progressing along a reasonable course, and many of the other suggestions and non-suggestions would either be far more broken, far more gameable, or basically boil the mode to mirrors only.

Can you even be sure the algo is currently looking at any parameters beyond top X gp? I can give you concrete numbers what parameters to look at with what weighting and how much error thresholds if this was upto me, but it isn't. There doesn't need to be a principal parameters, an algo can simply check and match multiple parameters together with certain thresholds. i.e. consider a %20 threshold weighted in between zetas, g12, g13 and mod quality distribution (not the amount of speed of them), if need be this can be much more granulated with decreasing weighting to omegas and lower gear and the algo can balance loss in one parameter with gain on another (i.e. one player has less zetas but more g13s etc.)

Assuming that there really are different parameters applied behind the scenes right now, albeit softly (as said in one of the announcements I can't bother to find right now) they can simply dial the secondary weighting process up and that can also suffice too. And moving pertinent GP calculation beyond defense+offense slots can be nice, dunno where the sweet spot would be on that.

I'm all for keeping it as is and simply improving it, I don't think we are at a bad place at all right now, just a lacking one. If I did this from scratch, I'd simply revamp the GP tables all together to solve the problem once and for all, doesn't seem possible anymore as there are too many things attached to the initial abrupt GP calculation.

To reiterate for the last time, mods, g11-12-13,zetas and maybe omegas. No gp. Weighting doesn't need any use of gp at all. Current GP matching seem to have a very low threshold as pertinent GPs are almost a perfect match, that's not necessary at all as long as there's enough weight on other parameters and GP can only be used as a safety measure to avoid to far jumps. I bet if the above parameters nicely fitted between matches, people wouldn't care anymore if their opponents have +-250k gp.

And yeah, this was what the majority of what the talks consisted of when people were asking MM based on parameters-. The debate mainly relied on whether toon specifics should be involved, whether mods should be involved at all etc.

"Vinny_Vader_Vedi;c-1913891" wrote:
Searching for Malek: a swgoh documentary.
Part 1 - where is he?
Part 2- when will he return?
Part 3- dev comment needed- when will malak come back?
Part 4- no malak no fun!
Part 5 - ugh everyone has malak arena is stale! I hate it! Can we get a hard to get toon!?!

A new hard to get toon??? no way!

We just need counters that beat counters...

"No_Try;c-1914386" wrote:
And current Top X GP algorithm doesn't allow that exploit? You just listed a few toons that may possibly benefit from low gear, I don't wanna go case by case by basis but some of what you listed are not eligible as how you claim them to be. Anyway this wasn't the only thing I said, you seem to ignore rest of what I wrote. First of all I didn't say simply match g12s one to one but allow a difference threshold between all the parameters balancing them. And it's CG to decide if to take this much further and simply granulate everything, i.e. assign 6x value to g12 compared to g8 etc. (and not just 2x as the current gp table does).

What's exploitable most of all is purely visible parameters matched very closely. If CG just made what's involved in it vaguely knowable and if the outcome of matchmaking gave out a fair balance between the parameters that they said is involved...then it would be much harder to game the system...as that would simply correspond to making a better roster (as opposed to making a super lean roster for first algo or making an excelling top X while also developing secondaries for back up for the current algo(bear in mind g13 makes next to no pertinent GP difference while making the top X excellingly good)).

I did not ignore the rest of your post. I trimmed it down to the most pertinent part so we're not just dueling walls of text. It's a courtesy on old school forums.

And the thing about hard locking within X% of parameters is you create corner cases that can't be matched. The other day, someone posted on Reddit an opponent with a fairly developed roster that had multiple G12 toons... but absolutely zero zetas. You have to match everyone. But how do you match someone like that when you hard lock within +/-20% on zetas and G12s when you have 0 zetas and a lot of G12s? How many people do you think meet that criteria? Do you think it's enough to populate a bracket? Do you think it's enough to populate a bracket without having the exact same people fight each other time and again?