Do you have an Arena 5-man team? - Page 2

LuckyAC · 02/07/07, 2:04 PM

Winning 20/25 should get you a higher rating than 10/10 given no other information. The average person should be paired with people of equal strength, and thus have about a 50% win %, so all that matters is the "games above .500". However, if you are one of the very top teams, you will tend to play people considerably lower, just because of the available pool, and 10/10 will probably be better (it will be as long as the average opponent is between 125 and 150 points lower, assuming same weightings as the FIDE system).

Either of these extreme results are only possible during the first week (when you may be very underrated and should play as much as possible until you reach your true strength) or if you are one of the very top teams in the system and thus are forced to be paired down (in which case there's no advantage to playing more, unless your skill/gear actually improves).

Fenrus · 02/07/07, 2:06 PM

Interesting. Of course what people should do and what the will do are often 2 different things. It's going to be interesting to see how things end up. After a while I'm sure there will be serious teams who will begin to understand the way the system really works. Those teams will play until they reach their equilibrium point (their "true strength" as you say) each season and rearrange their player makeups at the start of each new season.

Like all aspects of the game, there will be hardcore arena players and casuals. Casual arena players probably won't concern themselves with the finer points of the system and will only be interested in aquiring a specific piece or pieces of arena gear. They'll be happy playing as many games or as little games as the feel like as long as they are able to attain a rating that gets them enough points for that one piece of gear at the end of the season.

Ghostz · 02/07/07, 3:45 PM

To answer the OPs question:

We don't have any teams set up yet, but I'm sure we'll have at least two 5v5 teams ready to go for the first season. They haven't been assembled yet, but the interest is definitely there. I think everyone's just too caught up with all the available content that they're not spending any time thinking that far (yes, I realize its not really "far") ahead.

Gauss · 02/07/07, 5:18 PM

Meh the only kind of PvP that matters to me is ganking. As a mage main I wouldn't be caught dead in a lv 70 arena group due to the massive stam on everyone and my unwillingness to go some trinket trash spec to actually get a few kills. Perhaps some guildies are thinking of doing groups, but 5v5 arena still seems too sterile for me.

Beliandra · 02/07/07, 6:00 PM

Originally Posted by Fenrus

Like all aspects of the game, there will be hardcore arena players and casuals. Casual arena players probably won't concern themselves with the finer points of the system and will only be interested in aquiring a specific piece or pieces of arena gear. They'll be happy playing as many games or as little games as the feel like as long as they are able to attain a rating that gets them enough points for that one piece of gear at the end of the season.

For sure, I think you'll find a number of teams playing a LOT more than 10 games, simply because they're not concentrating on minimaxing the system, they're having fun, and the queues are short and the games plentiful. A talented arena team should be able to tear up huge numbers of casual teams in the first week before the ratings spread out and they stop being matched against them.

• sadris · 02/07/07, 6:29 PM

If there were 2 teams, A and B, which are rated 30th and 31st. They both played their 10 games and won against teams 20-29. Are you saying there is no inherent advantage for either team (A or B) to play an 11th game, even if they *may* end up playing team 19?

Dakous · 02/07/07, 7:24 PM

Originally Posted by sadris

If there were 2 teams, A and B, which are rated 30th and 31st. They both played their 10 games and won against teams 20-29. Are you saying there is no inherent advantage for either team (A or B) to play an 11th game, even if they *may* end up playing team 19?

Rating is not ranking. Multiple teams can share the same rating (at least, they can in Elo, which I understand to be the basis of the system). So, teams A and B, rated 1500 both, win 10 games, and magically the math works out on the net ratings for each team they both faced to result in them ending up with a 1600 rating. Yes, there's an advantage in playing game 11, winning, and thus having a 1605 rating (a number generated from complete randomness, since the ratings involved are only relevent relatively).

And yes, if some team is rated 2100 and the next highest team is 2000, that first team is indeed first place. But your choice in example suggested a confusion on ratings versus placings. Beating the #1 team when everyone is at 1500 and they had 1510 isn't as significant as beating them when you're at 1500 and they're at 2100. Ratings, ratings, ratings. It's like you're in the business of television.

The idea here, though, is that as you advance in rating, you will eventually face more difficult teams. That means your odds of winning go down. Eventually, most teams' recent match history will be approximately 50% wins, 50% losses - you're in your skill bracket. Or, more simply, you win a match, so the system finds you a tougher team, that you lose again. You just lost, so the system finds you an easier team. You win that one, so the system finds you a tougher team...

At that point, whether you played 10 matches or 30 matches, any change in position is essentially rounding error.

• Sservis · 02/07/07, 7:33 PM

Originally Posted by Igni

Originally Posted by Fenrus

In the event that a team does exceptionally well in their first 10 games and wins say 8/10 or 9/10 of them, then yes, they obviously won't be motivated to play any more and risk lowering their rating.

For your first example, the results will be the opposite of what you've said for teams who understand the Elo system.

Fenrus is right in a teams equilibrium behavior.

There's a concept of "true strength" inherent to the Elo system. You've reached that point when you're winning half of your games and your losing half of your games. So the team who has won 8 or 9 of 10 games should keep playing because they haven't yet reached their true score yet. They can continue playing games and when they start to lose as many as they win, only then should they stop.

If only every 10 game sample perfectly reflected expected performance. Just ask any NFL fan (.500 teams over 16 games have a SD of 2 games, which means the 2 SD window of 95% is between 12-4 and 4-12).

Ex.
A month into the 3 month season let's pretend that the rankings have mostly stabilized, (If they haven't, then what value is the "ranking" as information about skill)

Your team logs on and plays their 10 games for the week,

Here are the following outcomes possible and the likelyhood (assuming each game was truly a 50/50 proposition)

Wins %
10   0.1%
 9   1.0%
 8   4.4%
 7  11.7%
 6  20.5%
 5  24.6%
 4  20.5%
 3  11.7%
 2   4.4%
 1   1.0%
 0   0.1%

Your team now has a choice, play more games, or stop.

If you're willing to assume that Blizzard's matching algorithms and the rating system will do a good job of giving you 50/50 games (if they won't, then the whole rewards for good ratings seems even more pointless). I've put the system out a month to allow some settling of rankings as initially it will be purely random and tier N will own tier N-1 until things are sorted out.

The more that you're willing to trust the rankings and matchup algorithm, the more that you should decide to stop whenever you win more than half your games.

Why?

Because you've outperformed expectation, and in every rating system that I've heard of, that's good for your rating, and throwing more 50/50 matches into the mix doesn't help things generally (if you get win it does, but you lose 50% of the time)

Note: If you're not willing to trust the rankings and matchup algorithm, you may as well enter a lottery for cheap epics since the system isn't measuring skill accurately.

But if you're at a 90% win rate and you stop, you're leaving points on the table. When you queue up, hopefully you'll get paired up against a stronger opponent, you beat them and then your score rises. If you queue up and get opponents who beat you, then you've begun to approach your true score.

Your point about continuing to play after winning 90% comes down to statements that "Our teams results this week come from the rankings not being sorted out yet, not from luck." and "By continuing to play we will be matched up against better and better opponents, and outperform the result we would have received if we had stopped at 9-1 (but we still won't do better than 10-0")

In the early part of the season, I'm more than willing to agree with the first statement. I'd attach significant doubt to it by weeks 3 and 4 though. If they can figure it out in 13 weeks, they'd better be able to put teams in the right ranking region by weeks 3 or 4.

I'm not sure I can comment on the second statement. There likely isn't that much room between 9-1 and 10-0 they've both in the "vastly out performed expectations" category. How much room there is between the two in the implementation is a detail we have yet to know.

===============

Other thoughts on the Arena ranking system?

I think 10 is far too few games. That level of investment in a system is borderline not playing at all. This is my personal opinion, but I wouldn't be surprised if Blizzard's configuration parameters were changed at some point.

Given that it seems optimal to stop playing when you're ahead after your 10 games (after ratings have semi stabilized), it's also likely a very good idea to get your 10 games in early in the "week". Why? Teams know that when they've had a good run in their 10 games, it's best to stop playing (they trust the matching system to be working). This means that late in the week the only teams that are still playing are those that lost early and are looking to improve and teams that started later in the week. The sample of teams you might play no longer includes those teams who already won most of their games but were rated near you to start the week. These are exactly the teams that you hope to match up against if you've lost your early games and want to prove your early results were off, but these teams are not available late in the week.

ex. Would you prefer to play team A or team B, both ranked the same as you at the start of the week, but team A has gone 8-2 and B has gone 15-20? Assuming the rankings mean something after weeks of data and one weeks games mean less, I'd want to play team A. Late in the week, team A will no longer be playing in the Arenas (Why would they want to go 5-5 and go from 8-2 to 13-7?) Given this, I'd expect whatever timeframe is selected as the start of the "week" as the most important time to be doing Arena PvP.

The real data point that's important won't be available until the system is live. How close were the matches? If they are close (could have gone either way based on some factors), then the system is working and you should quit when ahead (system working, results luck). If they aren't close, then they're still matching up Dick and Jane in greens with bored raiders in purples and oranges, and I'm not sure what the right call is here. (system not working, do what?)

Keline · 02/08/07, 9:38 PM

I'm not 70 yet. I must get to 70 before Arena starts ohmygod.... must not sleep!

RK · 02/13/07, 3:28 AM

I have an arena 2 man team (warlock/shaman), and we're not likely to recruit a 3rd (if we do, we'll just start a 3-man as well most likely). Being the guy who only comes in for the bare minimum 30% will suck. Being the guy left out so that guy C can come in for his 30% will suck.

Speaking of 2-man teams, I struggle to think of what could beat warlock/warlock.... and I'm also a bit worried that any matchup with two locks is going to come down to who manages to fear the other one first, unfortunately (fear should really break instantly if a certain amount of hp is lost during it, because losing 6k HP in one fear and STILL having three or 4 dots ticking away at you is arglshfd;kgjbnk.jbskjfvbfjvb sorry, too much eye of the storm vs teams with 6-9 warlocks). Still, in 2 vs 2 I figure there's going to be lots of other "two friends of whatever classes" teams to have fun against. I don't expect to be server #1.

I think Igni has it absolutely right in terms of # of games to play, though. If you win less than half and you're genuinely unlucky (ie. you're not simply playing in a rating group too high for you or you're not drunk and fatfingering everything), keep playing. If you win most of your games, sit and keep your winnings. This is less true in the first few weeks and more true once the rating groups are more correct.

Also, gearing up from outside sources in between weeks (ie. raids) should prove to be a decent advantage, since from one week to the next you'll get stronger without this being reflected in your ratings yet, giving you a leg up over other people in your rating group who didn't get more gear.

Dricen · 02/13/07, 9:08 AM

Holy Paladin (me)
Feral Druid in partly healing gear
2 x Rogues
MM Hunter

The combination is brutal. While the druid is feral, he focuses on hybrid stats, making him able to support heal, and if he is under attack a change to Bear Form lets him survive almost any beatdown, and if ignored and "unemployed", then i have heard a rumor about Mangle, they say it hurts alot :P

this team lacks offensive dispellers, what happens if you run into a :
-PS priest
-AP mage
-Warrior
-Paladin
-Druid

Can't dispel neither Fort, Motw, hell PS or AP , so yea... while u have 2x rogues + a drood beating on that priest spamming PoM taking less than 60-100 dmg a hit and tanking them np.. , their AP mages/drood and warrior obliterate the rest of your team.. Not to mention you'll lose all your buffs, while they keep theirs.. that's a HUGE advantage in all aspects (armor, HP, blessings etc..)

When creating a 5man arena team you need to make sure you can at least cover the basics of support, those are offensive/defensive dispelling and healing

As for me, no i don't have an arena team yet, still on the push to 70 (new server), people don't seem to be making teams prematurely until the system starts.

Oh and yes, what others said is true, playing just 10 games and winning them will be the standard behavior for a lot of teams, but playing 10+ is what will make teams rank higher and higher in the ladder, however.. it has risk accompanied with it of course, and that is losing.

It's a risk vs rewards system, playing 10 and winning all 10 would be the epitome of "playing it safe with low risk" , the problem with that would be a mediocre ranking in the ladder, whereas playing 10+ = playing it riskier, but having a higher placing. In the end, the winners of the season will be the ones at the top of the ladder, but.. that won't prevent the people that played it safe from still getting some purple gear.

Casuals and conventional teams, will go the "stay at 10" route, the more hardcore and dedicated team will go beyond it.

Ghostz · 02/13/07, 9:20 AM

-PS priest
-AP mage
-Warrior
-Paladin
-Druid

The chances of actually being able to dispel AP or PS against a group like that are next to none. Everyone in the group will have:

Shadow, Fort, Int, Mark, Spirit and a Blessing. This is on top of their self-buffs (fearward, inner fire for priests).
I doubt anything but a shaman is going to be able to spam dispel enough on a target to take AP or PS off. As a priest usually being the focus of most of their dps, its pretty much impossible to keep yourself and your group up while offensively dispelling people with 8+ buffs. Also, I'd like to mention that mass dispel is really quite shitty. :\

Dricen · 02/13/07, 9:24 AM

indeed, but still, a group with a shaman could have no problem taking care of that, and not to mention, its not always about dispelling everything on everyoen , but at least essentials, for instance, if you're running the assist train on the priest, the other team's priest can spam dispel on taht 1 target and he WILL take that PS off, same for purge spam.

So yeah.. like you said, you'll most likely wont purge their whole team, but at least the person you have the train on, but if you can't even do taht, then it screws your assist train pretty badly and it can cost you a match.

Dricen · 02/14/07, 10:20 AM

I haven't seen another thread , or anyone mention this , but... this is new...

http://www.wow-europe.com/en/info/ba...t-season01.xml

that announcement is huge imo... brings WoW to a whole new level of competitive play, and will most likely up the sales numbers due to more people attracted by this new feature.

The fact that arenas would have seasons and would be competitive was old news, but this whole "World Championsip" brings it to a whole new level imo.

Narwe · 02/14/07, 2:57 PM

Originally Posted by Dricen

The fact that arenas would have seasons and would be competitive was old news, but this whole "World Championsip" brings it to a whole new level imo.

If I were an old-school MMORPG fan, I'd be pretty terrified at that news.

Cynic · 02/14/07, 3:10 PM

Originally Posted by Narwe

Originally Posted by Dricen

The fact that arenas would have seasons and would be competitive was old news, but this whole "World Championsip" brings it to a whole new level imo.

If I were an old-school MMORPG fan, I'd be pretty terrified at that news.

Would you care to be more specific? I had a grand time in EQ when we had "Best of the Best" guild competitions on the test server. Going up against other people you knew from different servers was a blast.

Igni · 02/14/07, 3:11 PM

Originally Posted by Narwe

If I were an old-school MMORPG fan, I'd be pretty terrified at that news.

In what regard? The new blood it will bring in?

Originally Posted by Dricen

will most likely up the sales numbers due to more people attracted by this new feature.

It helps explains why so much love has been heaped on the PvPers with TBC. Could you imagine the profit to be had if some fraction of the people who still play Starcraft signed up for WoW? It would also explain why they introduced the resilience statistic in an effort to prevent raiders from translating their gear advantage wholesale to PvP.

Nicketzsche · 02/14/07, 4:38 PM

What really sparks my interest about bringing competitive PvP to this level is the possibility of seeing some truly creative/innovative/adaptive/skilled players who do not necessarily get to the point they're at by spending >5 hours per day playing the game, grinding, farming, and putting up with the politics/time commitment of playing in a large, successful PvE guild.

The only reason I can think of as to why Warcraft does not have double/triple the player base is because it lacked a way for truly gifted players to show what they were capable of (outside of occasional fraps, which were typically not in an arranged 5v5 format where teamplay and creativity were necessary). Even when the occasional 5v5 was showcased, it involved only players from one realm.

Merely opening PvP in this way is more exciting to me than the expansion itself.

I really hope Blizzard finds someone to broadcast these tournaments... reminds me of the Southpark episode, everyone I know was glued to their tube for that, and I don't doubt the same would be true for a global 5v5 tournament.

Narwe · 02/14/07, 4:48 PM

Originally Posted by Cynic

Originally Posted by Narwe

Originally Posted by Dricen

The fact that arenas would have seasons and would be competitive was old news, but this whole "World Championsip" brings it to a whole new level imo.

If I were an old-school MMORPG fan, I'd be pretty terrified at that news.

Would you care to be more specific?

Sorry, "Old school MMORPG fan" is not very descriptive of the kind of player I mean to refer to. To be more specific, the kind of player that enjoyed the genre for its fantasy and community elements - the 'RPG' in MMORPG - may not be so excited about the potential impact that this level of competitive play might have on the genre/industry as a whole. WoW is an extremely entertaining game that does an excellent job at appealing to a wide market. Given this success, there is a reasonable concern that development of focused, niche MMORPGs that appeal to a less testosterone-fueled (re: more thoughtful) market may be hindered by lack of funding and confidence.

This is not an anti-PvP post. I am a hardcore PvP fanatic and actually find the PvP in WoW to be rather childish (though the combat mechanics are fantastic). I'm a fan of PvP environments like those in Lineage or Eve. In fact, it is the popularization of WoW's inconsequential, gear-driven, non social, non world-impacting variety of PvP that concerns me the most. In the context of WoW, the arena should be quite fun and I look forward to participating. In the context of the genre as a whole, instanced PvP disturbs me.

Nicketzsche · 02/14/07, 4:54 PM

I definitely miss recognizing players of the opposing faction on my server, but I suppose this is a case where I can't have it both ways (much, much more PvP with the same community feel).