I don't see how this means they don't get points. If I understand you correctly, the 5 scrubs play 3 games together and then the ringers come in for the other 7 games so only the min 10 games were played for the week. The scrubs still played 3 of the 10 games for the week, which is 30%.

Even if it did work as you described, the only people it would directly hurt are the other scrubs. It may irritate the ringers, and cause some irritation between the other scrubs and the ringers but the ringers still get the gold and did what they say they would do. It's beyond their control that someone else screwed up by leaving (after paying them the gold mind you)

This is way way too likely to get people that have earned their current rating. Playing the whole season and slowly working your way up the rankings as you figure out strategies and group compositions and teamwork. Perhaps needing to switch out a couple people for better alternatives. Your average rating could be stuck somewhere in the 1900 range since you only got to the super high ratings later in the season. In fact the only way to have a high average rating is by playing a high number of games at the highest ranking. If it takes you 100 games to work your way up to 2200 or whatever it can take a large number of games to get your average rating high enough.

If you progress linearly from 1500 to 2200, your game average rating would be 1850. Going slowly, say it took you 200 games to get there. You would have play a further 150 games to get a 2000 average rating. Once you're at a 2200 skill level, you may actually be better off to reform as a new team as you'd most likely rocket from 1500 to 2000 in pretty short order, and then it's much much easier to get your average rating up to 2000 as well.

I thought of a modification of the previous idea of personal ratings that would preserve the zero sum. Just put in a maximum point difference on the team.

So, say we set the maximum spread at 200 points. Given the previous example of a team with a 2200 rated player, the minimum would be set at 2000 rating. Anybody who teams up with him and has a rating below the minimum will temporarily be considered a 2000 rated player while they are teamed up together.

So, with the 2200 player & 1500 player, and the 2050 players ....

Team A - 2200 & 1500 (counts as 2000) - 2100 rating
Team B - 2050 & 2050 - 2050 rating

Team B wins

2200 player lost to a 2050 team - loses 13 points
1500 player (but counts as 2000) lost to a 2050 team - loses 7 points
2050 players beat a 2010 team - gains 10 points

it still serves the purpose of the individual ratings, but prevents potential abuses caused by large disparities of team member rankings. And is probably easier to maintain the zero sum system.

It is because that raid didn't directly impact other guilds raids in said zone.

I personally don't mind a top team selling spots on their existing team and taking the rating hit from letting the scrubs that are paying for the spot get their min fights in. My bigger concern is when a team feels the need to create 2-5 other farm teams to sell more spots in.

Selling a spot on your existing team is more akin to selling a raid spot. Rolling a second/third fourth or 5th team to sell more spots in the 1k point range is what wrecks the system and negatively impacts other teams.

/shrug, you're assuming that points, and standing mean something. Every team at the top has more than enough points, I get to outfit myself with an array of arena weapons I can use in town. As for the week to week standing, who cares, the end of season hasn't even been announced. I'd be happy if people queued up so I could get a game, I don't care how or why they're doing. We were losing 20 points a clip this weekend to a team that xferred onto our BG. They clearly weren't 200 rating or so below us, but they had to work their way up, but we didn't care about the rating we lost, it was nice that someone was just queueing up, we'll get our points back eventually, or we didn't deserve them in the first place.

The flipside of it, of course, is that when you play against that newly-purchased 2200 team, you're getting a shitload of free points that ordinarily don't come easy at those ranks.

And for this to truly have a significant affect on your arena experience you'd have both.

Please stop bringing up non-zero sum as a reason to not do individual rankings. It is trivial to multiply the winners gains by a factor that will retain the zero sum aspect of the system. In fact individual rankings are **more** zero sum then teams, because you don't have dead teams sucking up points or reformed teams inflating them.

I think that this is a particularly important aspect of Smurfing that needs must be noted, the lucky team(s) that get to play vs. the smurf team during its first few games each week can gain huge amounts of easy points. Additionally, while the pro players on the smurf team might put a hurting on a few ratings (or a significant one on a single team that chain-queues against the smurfers), that rating can be won back (rather easily, as XI notes: if you don't win the points back, from others, or from the smurfers, you didn't deserve them in the first place) after the smurfers log off.

I always chafed when my 2v2 went up against obvious smurfing teams back during S1, it pissed me off, sure, but then I realized that by dropping my team below our true rating, the smurfers gave my team a chance to regain true-rating+even higher as we could gain momentum from easy wins after our rating dump.

All the solutions being tossed around are just (in my humble opinion) variations on moral hazard: if we block smurfing, does this give teams more or less of an appetite for risky arena behavior?

Not really, you get natural rating inflation that hurts the system and ultimately you, in the end, worse. For instance, if most players play somewhat better down than up, with regard to their nominal ratings, then, as the pool of players increases, we will see natural rating inflation. Playing with a high rating way way above your nominal rating is suspicious at least, and makes the whole thing harder.

Same reason why 2700 used to be an all-time godly score in chess but now (post-1993ish) its just top 50 or 100 in the world. The consequences of unmitigated nominal inflation are pretty bad and I would certainly think outweigh a few easy victories here and there for legitimate teams up against smurfs.

Ultimately Blizzard would have to settle on an terrible unsavory conclusion:
1. Laissez-faire, let things continue without interference: Having PVP epics becomes trivial. Bad teams drown in purples. Insofar as PVP epics work as raid gear, content is trivialized because no one wants to fight for random boss drops when the cost of admission to the gravy train is so low. Walking around outside becomes the jungles of Viet Nam on PVP servers and battlegrounds and requires a good set of PVP gear to have a chance.

2. Artificial cap that adjusts for a flat -hundreds of points on every team: Devastates people that don't join early in the season, when gains are abundant and people can jump by leaps and bounds. As a player enters later and later in the season, if they are skilled, the delta to their natural rating increases steadily. Terrible situation for people on vacation or in a tight spot at whatever time the new season begins, or for new players. (as opposed to raiding, which doesn't significantly vary in terms of total time commitment to the "end" at any time you start up until the next expansion. A guild starting today in all blues in Karazhan with skilled players gets to Illidan at the same time an average guild that happened to start in February, within a small margin of error.)

3. Ever increasing scaling rewards prices (i.e. the zero-sum DKP solution Same problems as zero sum. A player getting in on the early end of things can rapidly become better than a well-established member, like the paladin that joins at zero and takes the EPIX TWO HANDER over the warrior, or whatever. However, unlike guilds which can establish grace periods of no loot or officer discretion or "guild bank tax DKP" or whatever solution they want, the arena is a constant unmanned mathematical system that allows "new recruits" from anyone at any time. This leads inevitably to a fucked up Nash equilibrium for the early half of the season, where no one wants to jump in and instead plotting their super-genius strategy as to "when its right" to dive in the water. This makes a huge portion of the nominal rating distribution pure luck based (on how lucky/successful you were at the time you broke equilibrium) instead of a skill-vs-skill grind that the arena (roughly) is at the high levels. In terms of "soft" effects, the early teams that just went right in fight the same small pool of suckers over and over and over again, which most people don't find fun.

Ultimately, I cannot think of any small consolation with letting things go as they have been that outweighs the long-term effects or later requisite adjustments.