100 replies to this topic

[StillRadioactive]

Matchmaker only works on one match at a time. This means that, when there's a mismatch between the number of attackers and defenders, extra attackers are meaningless.

Let's just take an extreme case as an example... say there are 48 defenders who came on at random times, and 600 attackers who all came on at once.

Out of those 600 attackers, matchmaker picks 12. It then holds that team for 10 minutes waiting for defenders to show up. The other 588 attackers are in the "ghost queue," so they'll see their timer counting up and up and up, with a lobby full of "PILOTS RECEIVING ORDERS."

Attackers playing: 0. Attackers queued: 12. Attackers in ghost queue: 588

When matchmaker can make a 12-man team of defenders, it launches the match and then finds another group of 12 attackers.

Attackers playing: 12. Attackers queued: 12. Attackers in ghost queue: 576

It holds this state for up to ten minutes, and during that time it find another group of 12 defenders and launches the match, then queues up another 12 attackers.

Attackers playing: 24. Attackers queued: 12. Attackers in ghost queue: 564

It holds this state for up to ten more minutes, and during that time it finds a third group of defenders and launches the match, then queues up 12 more attackers.

Attackers playing: 36. Attackers queued: 12. Attackers in ghost queue: 552

During this third match, the first match finishes and all its players go back into the queue. It then launches the fourth group of attackers against this recycled first group of defenders, while the first group of attackers go to the end of the line.

Attackers playing: 36. Attackers queued: 12. Attackers in ghost queue: 552.

Then the fourth group of defenders is formed and matchmaker launches a match for them.

Attackers playing: 48. Attackers queued: 12. Attackers in queue: 540.

This cycle continues perpetually, with the number of attackers actually PLAYING limited to the number of defenders, so long as there are defenders on the planet.

This means that players from more populated factions end up staring at the "PILOTS RECEIVING ORDERS" screen for upwards of 30 minutes during peak times, waiting for defenders to come available.

Auto-wins only happen when there are no defenders available. This happens during times of low player population, or during the ceasefire when there is one more team queued for one side than for the other... and usually, the planet isn't close enough that the one auto-win doesn't decide its fate.

I cannot speak for the other populous factions, but I can speak from my on personal experience as someone who assists in coordinating Davion's peak-hour operations. I know of only two times that this has happened for Davion, out of the 84 planet-days that have occurred so far (not counting planets that are protected by diplomatic arrangement). One attack and one defense.

So yes, they do rarely happen during peak hours... but when Russ said that their impact on actual planetary control is minimal, he was telling the truth.

Edited by StillRadioactive, 25 December 2014 - 06:28 PM.

[jay35]

You may be technically correct, but that does not explain how a planet would go from 1 point in the defender's favor to +5 in the attacker's favor over the course of less than 30 minutes, during which the one populated match that we know for sure did happen (because we were in it), the defenders won.

That means somehow, magically, 6 matches occurred and were all "won" by the attackers. Not only is that statistically improbable, but they would have all had to start and end within the timeframe of the match the defenders actually played (and won!).

The only way that's possible, given your scenario, is if ghost wins are happening simultaneously to fully-populated drops (or there are arbitrary score adjustments being made behind the scenes which is rather silly to even posit). You can claim ghost wins (aka defenderless drops) weren't happening during peak hours, but the actual user experience during the first week and a half of CW told a very different story. I can't even recall the number of times we had that happen to us while fielding multiple 12-mans defending key contested planets. One 12-man would pop out of a match that we won, only to find a significant swing in favor of the attackers, yet all FWLM teams reporting in after matches did not report results lining up with what was being shown on the planet's status. This is what leads to reasonable people suspecting that such auto-wins were occurring.

The other reason people suspected it was happening, even during peak hours, is because there were times we were dropped on planets without defenders despite being prominent planets to defend. It was reasonable to assume that if it was happening to us, it was happening to others as well.

It also seems logical that ghost wins, which take anywhere from 2 to at most 5 minutes to accomplish, could easily stack up in quantity during the timeframe of one fully-populated match, thus proving a very reasonable explanation for the wild swings in planet status.

All that said, has it gotten better since Russ started tweaking things? Yep. Has it gotten better since they started providing Attacker/Defender population info next to the planets list? Yep.

So it is much better than it was, but there were some strange things happening during the first week or so of CW. Those scenarios may have been mitigated, but let's not pretend that how it works today is exactly how ti worked at that time, too, when so many of us experienced some very odd occurrences firsthand.

Edited by jay35, 25 December 2014 - 10:46 AM.

[LegoPirate]

if theres 10 matches happening (not unlikely on high activity battles), winning 1 match is nice but if you lose 6/9 of the others theres your deficit

[StillRadioactive]

Alright, let me go ahead and explain what needs to happen to trigger the maximum number of auto-wins.

36 attackers, 0 defenders.

Matchmaker finds a group of 12 attackers and pulls them from the ghost queue into the actual queue. It then waits ten minutes while it searches for an opponent.

Attackers playing: 0. Attackers queued: 12. Attackers in ghost queue: 24

At the end of 10 minutes, it launches an auto-win and queues the next group.

Attackers playing: 12. Attackers queued: 12. Attackers in ghost queue: 12.

Then the first group finishes its auto-win and re-queues.

Attackers playing: 0. Attackers queued: 12. Attackers in ghost queue: 24.

Then the second group launches, and it queues the third group. At which point you go back to the previous step.

The maximum rate of change through turret drops is 6 points per hour. If a planet has seen a rapid shift, like the 6-point shift you described, it's because there are actual populated drops that are going all in the favor of the same side.

Remember that, as of the last time we got statistics on group vs solo play, solo play represented 84% of the matches played... so yeah, your group might have won, but there are a LOT of pugs on each planet during peak hours, and they probably all lost... especially if they hit enemy 12-mans who are motivated to turn a planet in the last few hours of the attack window.

Macro-level fact and micro-level experience can differ without negating either one. This is a case where what is happening on the macro-level (hundreds of players over dozens of drops) doesn't match what's happening on the micro-level (your 12-man winning a game).

Edited by StillRadioactive, 25 December 2014 - 06:29 PM.

[Lord Ikka]

If MM works the way you say StillRadioactive, it is the most inefficient one I've heard of. One match at a time makes no sense whatsoever in a game with a decent population.

[Driftwoood]

I see what you're saying...

To crunch the numbers a little further though, I'd say you need at least 4 units (48 defenders) to prevent any ghost drops in a BEST CASE scenario...

Assuming instant drops for defenders, you are probably looking at a little over 30 min per match (lets say 1 min in lobby, 1 min in drop ship, ~23-28 min game, maybe 1 min load time at start and end of match, and say 2-3 min between matches for getting unit to ready up, make any skll or mech lab changes, etc...) so somwhere around 28-34 min...

BEST case: these matches are evenly spaced out, so one starts every 7-8 min... so that means NO ghost drops should take place...

WORST case: All 4 matches start at the same time, so there are no defenders ready for new match for 28-34 min... so in that time, attackers get to ghost drop 2-3 matches...

REALITY case is probably closer to BEST case, so I'd guess maybe 1 ghost drop in every 15-20 matches?


Now with only 36 defenders:

BEST case: matches evenly space out every 9-11 minutes... so you probably get 1 ghost drop for every 6 matches with defenders?

REALITY case: 1 ghost drop for every 4-5 matches with defenders?


And then with 35 defenders (or less), which is not uncommon for Liao at the moment in the hours before ceasefire (hopefully this will change with time):

BEST CASE: 11 guys are doing nothing, only 2 defending teams evenly spaces out 15 min apart... you get 1 ghost drop for 1 defender game...

REALITY CASE: probably 5 ghost drops for every 4 defender matches...


Agree or am I missing something?

Edited by Driftwoood, 25 December 2014 - 01:46 PM.

[Kalendra RA]

I think you're all a little bit confused over how the queue is set up. To use OP's scenario, an indefinite surplus of attackers on a world, against 48 defenders, would queue up 12 attackers and start a 10 minute countdown: if there are already 12 defenders on the planet that can fit together to form a group, it immediately starts a one minute countdown during which time everyone involved in the match can revise their drop decks and choose a mech; once the match starts that lobby is cleared from the queue and the next teams are built, and so on, meaning all 48 defenders could be in a match in as little as four minutes, after which for every ten minutes there are fewer than 12 defenders in queue an attacker would get an autowin. So let's say there's time for about two attacker groups to get in an autowin if all possible defenders are currently in a match, and the number of attackers is sufficiently high as to never be a limiting factor.

The official numbers for autowins during peak hours are about 2-3%, so it does happen, just rarely.

[Driftwoood]

SirPseudonymous, on 25 December 2014 - 04:48 PM, said:

I think you're all a little bit confused over how the queue is set up. To use OP's scenario, an indefinite surplus of attackers on a world, against 48 defenders, would queue up 12 attackers and start a 10 minute countdown: if there are already 12 defenders on the planet that can fit together to form a group, it immediately starts a one minute countdown during which time everyone involved in the match can revise their drop decks and choose a mech; once the match starts that lobby is cleared from the queue and the next teams are built, and so on, meaning all 48 defenders could be in a match in as little as four minutes, after which for every ten minutes there are fewer than 12 defenders in queue an attacker would get an autowin. So let's say there's time for about two attacker groups to get in an autowin if all possible defenders are currently in a match, and the number of attackers is sufficiently high as to never be a limiting factor.

The official numbers for autowins during peak hours are about 2-3%, so it does happen, just rarely.

No confusion... what you describe is the WORST case scenario for 48 defenders vs. 60+ attackers... that being: that all 48 defenders arrive at the exact same moment, and kickoff 4 matches all at the same time... thereby, any surplus attackers get to ghost drop every 10 minutes while those 48 defenders are in a match... (approx 30 min)

The more REALISTIC case, is that with 48 defenders, 1 match ends / new match starts every 7 or 8 minutes, and since the countdown for ghost drop is 10 min, ghost drops never (or rarely) happen...

But again, like I pointed out in my post earlier: you need at least 48 defenders to avoid any ghost drops...

36-47 defenders means probably 1 ghost drop in every 6 games...

24-35 defenders probably mean 1 ghost drop in every 2 games...

Less than 24... 3 ghost drops in every 4 games?

[StillRadioactive]

Yeah, and the official statement 2-3% autowins during peak hours lines up not only with what is reported to me by the boots on the ground and my own personal experience, but it also lines up with the number of planets which are decided by a last-minute ghost win (2 planet-days out of 84 = 2.38%).

That said, defenders don't ever show up en masse, they always just trickle in... this makes ghost wins extremely rare during peak hours even when - in Liao's case - the total number of defenders is around 45-50.

Also, I'm going to edit the original post to change the term "in limbo" to "ghost queue..."

Because ghost queue goes with ghost planets, ghost drops and ghost heat.

[R E T I]

Candy coat it however you like but one then i know for sure davions are the kings of zerg rushing and match stalling

Zedekiah316, on 25 December 2014 - 07:24 PM, said:

Candy coat it however you like but one thing i know for sure davions are the kings of zerg rushing and match stalling

[Driftwoood]

So tonight for example, Liao has been pressing the attack at Maladar (or trying anyway)...

2 attacking units have been pressing 3 defending units, and winning mostly... but needle barely moves becuase of likely ghost drops by 3rd defending unit...

Now, with last hour before cease fire, we have 4 defending units... so likely around 4 ghost drops by the 2 extra defending teams while the 2 attacking teams engage the other 2 defenders...

So I expect Maladar to be back at 0% by cease fire...

BOTTOM LINE: I have trouble believing ghost drops did not make the difference tonight and didn't make up 15%-30% of matches in the last couple hours...

Granted, tonight is Christmas... hoping things change after holidays...

Edited by Driftwoood, 25 December 2014 - 08:39 PM.

[Scoops Kerensky]

The credibility of gains/losses in this initial version of CW will always be in doubt until ghost drops are revised as a mechanic and the majority of the playerbase feels like the results on the map can be trusted.

You need to quit wrapping your ego around these wins because the insecurity of davion posters flooding every subforum with arguments about how, "ghost drops aren't a big factor and guys, GUYS our wins count we really ARE the best right?" look pathetic right now.

Y'all aren't to blame, its bad design, its a beta, take a deep breath and take whatever the map looks like right now with a grain of salt. It doesn't matter right now.

[Roadbeer]

Lord Ikka, on 25 December 2014 - 01:31 PM, said:

If MM works the way you say StillRadioactive, it is the most inefficient one I've heard of. One match at a time makes no sense whatsoever in a game with a decent population.

It's not one match at a time, it is 1 match being created per minute ASSUMING that both attacker and defender have 12 players queued up.

There is the MAXIMUM of 15 battles going on simultaneously, staggered in 1 minute intervals in their creation. It's actually kind of efficient as it churns the queue faster than waiting on 15 battles that were CREATED simultaneously, which would freeze the queue until one of those battles ended.

This way, the maximum wait time, in the worst edge case scenario is 15 minutes before another match is created.

It's explained better here.

http://www.comstarch...xplains-queues/

[Shimmering Sword]

Zedekiah316, on 25 December 2014 - 07:25 PM, said:

davions are the kings of zerg rushing

May I introduce you to one such unit called the BWC, they just so happen to reside in Marik space, and are/were one of the most notorious light spamming units. I wont complain though, because this is a game, they can use whatever tactics are necesarry to win.

Socop, on 25 December 2014 - 10:34 PM, said:

insecurity of davion posters flooding every subforum with arguments about how, "ghost drops aren't a big factor and guys, GUYS our wins count we really ARE the best right?" look pathetic right now.

We're just here trying to explain the mechanics of the system, and how it debunks the mystical "Davion wins everything on ghost caps alone" theories that are everywhere. We are responding to other faction claims, these are not proactive posts. Pathetic are the people who can't handle the fact that their faction might be underpopulated AND lose more matches than they win. I envy the factions who are underpopulated, you don't have to deal with long wait times.

[Roadbeer]

Having been on the receiving end of more Davion attention than any other faction. I can tell you as a certifiable fact, at least along our border, that while ghost wins CAN happen, when the queues are full (15 zones occupied) it is statistically insignificant and usually wiped by the resolution of the next drop in that zone.

If you're getting turret wins against you during the Hot Zone (the 3 hours prior to Cease Fire), then you really need to evaluate where your forces are deployed. You only need to have 96 pilots in the queue to prevent a ghost drop if the planet is at 50% (attackers cannot ghost drop a zone they already occupy, they can only defend), and that number goes down the more territory they take.

[Scoops Kerensky]

Shimmering Sword, on 27 December 2014 - 12:35 AM, said:

We're just here trying to explain the mechanics of the system, and how it debunks the mystical "Davion wins everything on ghost caps alone" theories that are everywhere. We are responding to other faction claims, these are not proactive posts. Pathetic are the people who can't handle the fact that their faction might be underpopulated AND lose more matches than they win. I envy the factions who are underpopulated, you don't have to deal with long wait times.

You're trying to argue (without the hard data to back it up - we have only Russ's 3% tweet with no explanation or data to place it in context - is it 3% of all planets total? Per planet? Is that an evenly distributed number? How many of those are in the last 2 hours versus the other 22?) that davion's gains are legitimate and should be accepted as legitimate by the whole community with less than a month's testing of a beta game mode, and that your overwhelming population advantage has no real impact.

Considering PGI's reputation with information and the constant gloating on nearly every subforum of CW that davion is the greatest ever, all us plebs can eat it - no, you're not responding. You're a bunch of people power tripping over arbitrary lines on a map in a free to play mmo and being hostile to anyone suggesting there might be flaws in a BETA mode they just released. Breathe. Calm down, and let them revise ghost drops so the majority of us feel like the game map is credible. Trying to force a conclusion to the issue is making you all look insecure, so quit insisting you have the answers when you don't have the hard data available to back it up either way.

[StillRadioactive]

Socop, on 27 December 2014 - 03:24 AM, said:

You're trying to argue (without the hard data to back it up - we have only Russ's 3% tweet with no explanation or data to place it in context - is it 3% of all planets total? Per planet? Is that an evenly distributed number? How many of those are in the last 2 hours versus the other 22?) that davion's gains are legitimate and should be accepted as legitimate by the whole community with less than a month's testing of a beta game mode, and that your overwhelming population advantage has no real impact.

Considering PGI's reputation with information and the constant gloating on nearly every subforum of CW that davion is the greatest ever, all us plebs can eat it - no, you're not responding. You're a bunch of people power tripping over arbitrary lines on a map in a free to play mmo and being hostile to anyone suggesting there might be flaws in a BETA mode they just released. Breathe. Calm down, and let them revise ghost drops so the majority of us feel like the game map is credible. Trying to force a conclusion to the issue is making you all look insecure, so quit insisting you have the answers when you don't have the hard data available to back it up either way.

You keep harping about this overwhelming population advantage, completely ignoring two basic facts:

Kurita + Liao + Marik = a substantially larger population than Davion, and Davion is still winning.

Kurita = a slightly larger population than Ghost Bear, and Ghost Bear is still winning.

Kurita = a substantially larger population than Smoke Jaguar, and Smoke Jaguar is still winning.

Davion may be the single largest faction, but we only make up ~15% of the game. Population alone doesn't determine success... if it did, then Smoke Jaguar would have never left Turtle Bay.

[Scoops Kerensky]

StillRadioactive, on 27 December 2014 - 04:27 PM, said:

You keep harping about this overwhelming population advantage, completely ignoring two basic facts:

Kurita + Liao + Marik = a substantially larger population than Davion, and Davion is still winning.

Kurita = a slightly larger population than Ghost Bear, and Ghost Bear is still winning.

Kurita = a substantially larger population than Smoke Jaguar, and Smoke Jaguar is still winning.

Davion may be the single largest faction, but we only make up ~15% of the game. Population alone doesn't determine success... if it did, then Smoke Jaguar would have never left Turtle Bay.

Again you can't help but try to make this personal, as if my points had some relevance unique to just Kurita or Davion alone, or you try to make a more ridiculous claim that somehow Kurita/Marik/Liao have bigger numbers at prime time (citation needed) or that this is going to matter because these three groups can queue together (they can't, except for defending a kurita planet from clan attack).

Do numbers affect the game significantly or do they not? Because you seem awfully quick to insist that they do when it would make davion look like the outgunned hero defying all odds, and just as quick to insist they don't when people point out your unusually rapid progress on every front bordering you.

More than once Davion in this thread and others have tried to push the line that you're all somehow just that much better organized and better at playing than literally the entire game playerbase by using teamspeak to coordinate (something all the factions fighting you do as well) and everyone is just jealous of your skills. Funny then that you can't stop squawking about numbers or making personal attacks anytime people suggest that the mechanics of a beta might not be perfect yet.

[StillRadioactive]

I haven't made a personal attack at all. I'm pointing out relevant examples of population imbalance where the larger faction is losing.

You keep claiming that numbers alone are causing Davion to win on 3 different fronts.

You refuse to acknowledge that Kurita outnumbers Smoke Jaguar and is losing on that front.

You can't have it both ways. Either numbers are everything, or they're not.

[Scoops Kerensky]

StillRadioactive, on 27 December 2014 - 05:52 PM, said:

I haven't made a personal attack at all. I'm pointing out relevant examples of population imbalance where the larger faction is losing.

You keep claiming that numbers alone are causing Davion to win on 3 different fronts.

You refuse to acknowledge that Kurita outnumbers Smoke Jaguar and is losing on that front.

You can't have it both ways. Either numbers are everything, or they're not.

If you'd like to point out where I said the numbers are the only factor I'd love to see it. My beef is the insistence by certain folks that it not only doesn't matter, the current mechanics mean there's no way it could.

If you want an explanation for the smoke jaguars situation, being the 4th largest faction is still quite a bit smaller after the first 3 that the majority of the game comes for, and it's playerbase spread out along 4 (5 for most people who don't read these forums) different fronts rather than 3, and it's the holidays so more than a few folks like me aren't playing right now. Not to mention on top of it clan attacked worlds are giant pug queues that anyone can join, and more often than not people join them because they *don't* care about CW and just want to play invasion mode regardless of wins/losses.

Numbers aren't a binary everything/nothing line either. It could just be that their an X factor that's going to tip people into winning more often than losing statistically. None of us can know unless PGI decides to release hard information that we can all look at. That's my point.

EDIT: I forgot to mention this point regarding Kurita specifically so I apologize. I know personally from attending HC meetings that a lot of House Kurita is based in non North American timezones. Quite a few of our most popular units are only active in European time zones can do little to influence CW at the moment to their frustration.

Edited by Socop, 27 December 2014 - 07:00 PM.