Monday, April 17, 2006

Comments on PvP System

Thoma points out some issues with my proposed system, and I've decided to make my responses into a full post, to expand upon my idea.

1. Skill isn't rewarded, just wins/losses. Being AFK is even more rewarded under your system then the current system. Since all honor gained is win/loss rather then a mix of win/loss/HKs then you just turn AV into a bot farm.


No, because a loss means your rating goes down. That's the difference between now and then. When you go afk, you basically have a 50% chance to go up and down, which becomes less than 50% because your team is effectively outnumbered.

To go afk and gain in rating implies you were on a succession of teams that consistantly won, even though they were a man down. That seems improbable, at the very least. If it was an organized team, the leader should recognize that you aren't contributing and boot you.

The reason botting works now is that there is no risk to just sitting and collecting HKs. In the current system, if you spend the entire match fishing, you will end up better than you started. Guaranteed.

2. The stake to the heart you propose world pvp is going to really marginalize the PVP servers. It just makes it into Gank fests that don't end due to diminishing returns.


Perhaps. It won't be a gank fest though. It will return world PvP to the same state as before the honor system (and to be honest, roughly the same state it is now, at least from my latest experiences on a PvP server). As well, each goal would only be achieveable once per week, hopefully minimizing the disruption.

The real gankfest was when we had the Honor system but no BGs. That was a terrible time. I was in the high 40s on a PvP server, within range of the 60s. It was supremely unfun. It's why I have a character on a PvE server now.

My system has no rewards for killing individual players, only for achieving goals, so people will only gank for the old reasons of boredom, opportunity, malice, and fun.

3. You assume that all servers have enough PVP to make seperating out "skill groups" worth while. And what happens to the Grand Marshals? Do they only face each other? Or do you have to wait for 40 of each groups high ranks get together for an AV?


Yeah, to be able to separate in skill implies sufficient numbers. This is probably not possible for most servers at the moment. However, Blizzard is working on cross-server BGs, and it would be a good solution then.

As to high ranks, the players closest in rating play each other. If there were only two teams, they would play each other, even if there was a great difference in average rating. However, the better team would only increase rating by a tiny amount if they won, and the losing team would only lose a tiny amount. However, if the weaker team won, there would be a larger gain/loss.

It's basically the same system--in principle--as the ELO chess rating system, or Microsoft's Trueskill™ system.

7 comments:

Thoma said...

Except that you don't need 40 people pulling thier weight in AV. In WSG a person botting can make a differance but in AV you can have 10 people botting and it doesn't matter. All you need is 10 people to spearhead and 10 following behind to win.

Given that on my server, there are accusations of people being "power leveled" in PVP ranks, it's not that hard to slip one person in to bot if you're a ten man pre-made knowing you have a 80% chance of facing a pug.

But you're still missing the point of this. It doesn't account for individual skill at all. It only counts group skill. If I kill 100 horde by myself but everyone else gets killed and we lose, how does that reflect on my skill?

And what happens if I disconnect? Do I gain credit if a win happens even if I was away long enough to be ported to IF? Or if I fake a disconnect because we're losing does it count against me? So what prevents me from never taking a loss by just disconnecting?

Problems your system still has:

1. Still doesn't measure individual skill.

2. Makes it harder for people to grind for gear unless they end up in a pre-made. In essance you've just turned the PVP system into the Raiding system.

3. Still doesn't fix world PvP.

4. Ignores AV. If you give AV a high honor value then losing makes it too much of a crapshoot to be worth while. Too small and it takes too long to be worthwhile.

5. Doesn't provide any benfits over the current system.

GSH said...

If you kill 100 horde and lose, why do you deserve to gain in rank? You lost. You were beaten. The other team outplayed you.

What about the person who played defence and successfully held her flag while you were off gallivanting around and racking up kills? Is the current system fair to her? In victory or loss, right now she will be worse off than you.

What about classes that cannot kill as well as others? Exactly how many paladins won the Test of Honor contest?

The fundamental difference is that I don't see kill counts as a true measure of skill. A skilled player is one who wins. I think that given two players, the one with a better win/loss record is probably the better player.

The most skilled PvP'er I have ever faced was a Horde druid named Goar. The man defended the Farm brilliantly, ran amazing flag caps, and generally did everything he possibly could to win. He never really had a high kill count in the games I played against him, and yet he was universally acknowledged as the one person you did not want to see facing you in a BG. And yet, this does not matter under the current system. All that truely mattered was how many hours a week he spent playing.

Also, the quality of opposition matters. If you're on a good team, beating pickup groups with a poor rating won't improve your rating. You need to beat good teams.

There is so much more to PvP than simply killing the opposition. Feints, false retreats, flanking manoeuvers, all the great tactics of warfare simply are not useful in the current "who has the most kills" mentality.

Also, disconnecting = loss. If you're good enough, you'll make the loss up.

1. I think it does measure individual skill if somewhat indirectly. If you are always on a losing team, maybe you should take a good, long look in the mirror.

2. I don't think that PvP should be a grind for gear. Rank should equal skill. Skill is the opposite of grind. You can always do PvE if you want to grind. Secondly, there is a grind built in with the factions. You can always grind for the faction rewards.

3. I've already acknowledged that World PvP is a flaw. But it's a flaw under the current system as well. I think that faction-wide goals is a pretty good motivator of proper play.

4. I'm not ignoring AV. I'm not sure I agree with this "you only need 20 people" idea. If you have 20 good people, and the other side has 30 good people, you should lose. Maybe people will stop playing the rush to enemy base and no defence style of AV, as there is now a penalty for losing.

5. Yes, it does. In the current system, the highest ranks go to those who put in the most time, rather than the most skilled. My system encourages people to play BGs properly, to defend, to heal, to do the boring but necessary things that lead to victory.

GSH said...

Also a rating system is more consistent. Your rating lasts your lifetime and does not change drastically. How long do people who make Grand Marshal stay at Rank 14?

Is it right that they drop rank right away? Shouldn't the Grand Marshal for a server, the best pvper for that faction, still be among the best a week or two later?

This whole "a new Grand Marshal every week" is kind of silly. The best player on the server should be Grand Marshal, until such time as another player proves she is better.

Thoma said...

Except that your system has no relation to skill of the individual. Just group skill. That one druid may have held that node...while the other four got rolled. For all of his skill, he lost honor. If one great player gets stuck in PUGs then he will never advance, no matter how good he is.

You post the theory the person with the most wins is the best player. That's like saying the Steelers have the best quarterback because they won the Superbowl. They don't by a long shot.

The honor rewards work like like this under your system:

"Sure you defended this node five times. You killed 10 horde. But I'm taking Honor away from you because the rest of the group sucked. Sorry, sucks to be you."

I will point this out again. You have zero metrics to view individual player skill. All you do is give plus honor if the team did well and take it away if the team didn't do well.

There is a reason that detailed stats are kept in team sports. So you can acknowage great players on bad teams. Trying to take individual ranking systems and applying them to a team ranking then assuming that you get an accurate individual ranking is foolish. Again with the teamsports analogy, there is a reason records reset every season.

GSH said...

"You post the theory the person with the most wins is the best player. That's like saying the Steelers have the best quarterback because they won the Superbowl. They don't by a long shot. "

No, the Superbowl is just one game in a single-elimination tournament. This is like me saying that Tom Brady (wins multiple Superbowls over the last few years) or Peyton Manning (wins most games last season) is the best quarterback in football. And that's a lot closer to reality.

The other problem with referencing sports is that people don't change teams in sports that much. In a BG, you could easily be on 10 different teams during a single week. If all of those 10 teams sucks, there's a good case that you suck as well.

Secondly, the current system doesn't do individual metrics well, either. If you are in a raid, you share honor evenly with the people around you. The only difference is that people on defence get less honor than people on offence.

Finally, these types of rating systems work! Read the Trueskill paper/site. It takes the uncertainty of exactly how much each person contributes to the win/loss and uses it in its system.

An analogy from the Trueskills FAQ:

"Suppose you have four objects (players), each having an unknown weight (skill). Suppose further that you have a balance scale (game) to measure weight (skill) but are always only allowed to put two objects on each side of the balance. If you always combine the same pair of objects, the only information you can get is which pair of objects is heavier. But if you recombine the players into different pairs you can find out about their individual skills. As a consequence, the TrueSkill ranking system will be able to find out about individual players' skills from team outcomes given that players not only play in one and the same team all the time but in varying team combinations."

Thoma said...

The ELO rankings work because it is always a single player on single player or single team on single team. You will note that in the MSDN blogs on the subject that the Trueskill system breaks down in many game types.

" Our assumption is indeed that the skill of a team can be calculated as the sum of the skills of the team members. This is in many ways the simplest assumption possible and leads to an efficient update algorithm for the player skills. However, we are fully aware that this assumption may not be valid for all types of team games. Take, for example, a Team Sniper game in Halo 2. It would be very well possible that the game outcome is determined by the skill of the best sniper in each team, not by the sum of the skills in each team. One can also think of games where the weakest player dominates the outcome. Think, for example, of a team racing game in which the last team member to cross the finishing line determines the outcome. Clearly, your weakest player would determine the fate of the team. However, these are all game type–specific decisions. TrueSkill, in contrast, aims at providing a unified ranking and matchmaking for all games. In the future, we may think of adapting TrueSkill to specific game types. "

Given that the three BG's are very differant in the taticts used (WSG is offensive, AB Defensive, AV is split.) you are assuming that a great WSG team is good in AV (where a zerg rush is very beatable by skilled defenders) is a bad one.

Again, it all comes down to you assuming that team skill/number of players = player skill.

Maintain said...

I think this would be an improvement over the existing system, but I would worry about PVP becoming too PVE-like. With your suggestions, I can only imagine that larger guilds would simply dominate, just as they typically do in PVE situations. This proposed system actually encourages ridiculous amounts of twinking, which further unbalances how the Battlegrounds run.

While this would be great for established guilds, it greatly lowers the barrier to entry for new PVPers. I do agree that skill should be a major factor in determining honor rank, but don't think it's fair to exclude everyone who isn't in an advanced guild and has their toon twinked to the nines.

Of course, the current system is far from perfect and definitely needs improvement. With the exception of dishonorable kills, World PVP seems about as good as it can get under Blizzard's system, which is kind of sad. Unfortunately, other than fun, there's currently no incentive to do any World PVP because there are no rewards for doing so. It certainly seems like Blizzard wants all PVP to take place in battlegrounds, but the current system is obviously broken, so we're right back to where we started...

One thought I had was to remove the concept of "honor" from BG's and focus on faction instead. That way, rewards are given for actually accomplishing goals in the BG rather than standing around fishing. If you don't participate, you don't get faction, simple as that. Say you could get faction for (from most to least faction using WSG as an example):

1) Winning the match
2) Turning in the flag
3) Capturing the flag
4) Killing a flag carrier
5) Killing an opponent inside your base

BG's could still be "farmed" by organized teams for wins, but close matches between two teams would guarantee fair amounts of faction for both sides. If you camp the graveyard and cripple the opposing team, you lose out on potential faction, same thing if everyone ends up fighting in the middle of the field or hiding in their own bases. To avoid people fishing, let's say that faction for winning is only given to those who gain at least one point of faction otherwise.

I know the above isn't a perfect solution, but it does encourage you to actually be a good player (if you return a flag or successfully defend your own, you get more faction), offers potential rewards for both sides while encouraging actual wins, and hopefully eliminates "bad" BG practices like camping graveyards and non-participation. Unfortunately, it doesn't take World PVP into consideration, but I don't think that can be improved in the same manner as BGs.

I dunno, it's just a thought, I'd like to know what you guys think, whether it would be an improvement or not.