Sunday, February 01, 2015

The Flaw in Extended PvP

As Gevlon astutely points out in the comments on the last post, there is a basic flaw in Extended PvP (PvP carried out over multiple game sessions). Players will defect to the winning team. PvPers often talk about looking for good fights. But when you look at their behavior, they'd rather join the best side and steamroll the opposition, rather than do their best and lose narrowly. Sometimes it's even a self-fulfilling prophecy. Everyone believes X will win, and thus everyone joins X, leading X to win.

Transient PvP doesn't really have this problem. Since teams are formed anew at the start of each game or play session, matchmaking and randomness can generate teams with a reasonably equal chance to win. Or if a player finds themselves on a bad team, well, next game or session might be different.

So then the question becomes: How do we incentivize players to stay on losing teams, or even join the losing side? Can we also do this without incentivizing people to lose on purpose?

Perhaps we can do something with the reward structure. Maybe the faction with a losing record gets a higher number of guaranteed points win-or-lose, and an even greater number of points when they win. The problem with schemes like this is that you have to be very careful, or there will be an optimum point where below you want to win, and above you want to lose.

The only other idea I had was the winners automatically absorb the losers. Like if Faction A defeats some of Faction B in battle, those Faction B players become part of Faction A. Of course, this might lead to a positive feedback loop, where Faction A becomes quite large and ends up winning everything. So you'd have to build in some incentive for A to "split" apart.

Multiple faction games can make the equilibrium more stable. But there's still a possibility that one faction will end up overwhelming the others if enough people defect. As well, there's also the issue that the factions may opt for stability, rather than trying to defeat the others. Arguably, Eve Online has fallen into this state. The major players all have their own empires, and are fairly stable apart from minor skirmishes along the borders.

What other possibilities are there to keep teams or factions balanced in Extended PvP, while still keeping things from stagnating?


  1. I can only imagine one stable scheme, the "zerging noobs vs the elite few". There are two teams with different reward structure.

    Let's say that there are 1000 players on the server.

    The "zerging n00bs" get 1 pt/player for every lost battle and 3 pt/player for every victory. Each member gets the same points unless AFK.

    The "elite few" team gets 500 pt, regardless win or lose, the points are distributed among the members unevenly, according to personal performance (so the individual member is still motivated to try to win).

    If everyone joins the "elite few", the average player will get 0.5 pt (500/1000). The points will be distributed unevenly, so the below-median players will likely get 0.1-0.3pt. They are better off joining the "zerging noobs", even if those are losing.

    If everyone joins the "zerging noobs", then the first player switching sides will get 500 point, so there is always reason to switch.

    In the equilibrium, the good players are all in the "elite few", the bad players are all in the zerging noobs and the worst "elite" gets the same reward in both teams.

    The winrate can be manipulated by manipulating the point pool of the "elite few". If they win too many times, decrease the pool to 450 and their weakest players will be better off leaving for the "zerging noobs"

  2. It's a difficult balance. If you try to lock a faction, then you're already acting too late b/c the faction is winning and the other factions, if they can't jump to the winning ship, will just quit playing. If you try to reward being on the underdog team you still are telling players that they have to lose, they'll just "get more". However we already know that players hate losing more than anything else; people happily game any "50/50 win-lose" system they can whether it's through smurf accounts or intentionally losing games (which in their mind isn't a loss b/c they didn't actually try).

    Trying to empower the losing faction discourages the winning one and they just give up or it just becomes a more prolonged loss for the losers. Plus if the issue gets too bad it just turns each loser into a mini-raid boss that dies a death of a thousand cuts (Wintergrasp Tenacity springs to mind). Some dev could try putting a less visible thumb on the scale (imagine it taking one or two more/fewer bullets to kill a target; in most games people don't have time to count) but that has its own risks.

    I think a game that successfully eliminated this would have to look at what is happening when someone is losing and adjust that so that you still have fun. To use planetside 2 as an example. If your side is losing then what is most likely happening is you are spawning and dying within a few seconds. You're not getting kills, little XP, and you don't feel like you're getting anywhere. Sitting in respawn screens and repeatedly getting slaughtered isn't fun. So you find a way to keep people active, even if they are losing. Or perhaps even design a game where a loss is somewhat inevitable (such as a "Score as many points before ____ happens." type game) so it is an accepted fact and it's more about what you do beforehand. Losing itself might not be fun, but it's even less when you're not even playing anymore. No one enjoys sitting out for 30-40 seconds after dying in a MOBA. In short, devs need to find a way to make the gameplay of losing identical (or substantially similar) to that of winning. If you're getting the same experience/rewards either way, then there's less reason to switch sides.

  3. tl;dr EvE isn't stagnant, it's just on a massively longer timescale than most games. Why would I spend 2 months grinding structures, another 2 months moving in and putting up structures if I will lose this space within 4-8 months?

    People feel that null-sec is stagnant in EvE because that is the currently accepted storyline among most powers (likely an attempt to keep it so). Realize that a null-sec war against a major power takes ~2 months to win and another ~2 months to truly complete.

    Last year we had several empires fall. It is true their space didn't go to new hands but went to old warlords, but the jump changes should (and are) making this un-tenable.

    I would expect 3 large wars a year in EvE. Basically ~1/2 of the player base is English first language. The English first language world really only notices wars involving English first language. Assuming people have an equal odds of fighting everyone else (likely not true) we'd only expect to really hear about 1.5 wars per year. Vast Russian empires change hands and all we hear about is what the Mittani ate for lunch.

    tl;dr EvE isn't stagnant, it's just on a massively longer timescale than most games. Why would I spend 2 months grinding structures, another 2 months moving in and putting up structures if I will lose this space within 4-8 months?

  4. This is a matter where you need to take a step back and look at the issue from the basic level.

    The basic question is "What is the prize for winning?"

    Looking at this from a wide range, the prize for winning in the kind of encounters in the real world that these forms of "empire" PvP are meant to represent is territory and resources; the ability to, in fact, eventually obliterate the opposition such as they cease to exist.

    So, you need to decide at the start how you want to develop your game as a result. The first question is whether you even want to reward victory in such a manner at all.

    If you want to reward victory, then timelines are very important. If a single battle can capture a resource or territory, then the whole metagame needs to reset very frequently or else you always get a single side with an overwhelming advantage quickly, thus attracting all the players. However, what if capturing required a much more long-term investment, much like modern conflict. What if each territory in Planetside 2 didn't take 10 minutes to capture, but instead took 10 days? 10 weeks? Each capture requires more investment from the attackers, and allows the defenders to make use of more and different tactics in order to retain it, and slows down the snowball effect of taking territory. While it may be thought that having quick turn-overs helps keep the balance closer, I think that it simply enables the snowballing.

    As another idea, what if territory or resource control wasn't directly fought over, but was instead based on factional results over a battle "season". Let's say that each month, there's a different location being fought over, and the faction with the best result over the month gets the benefits of it until it comes back up for grabs. Since the reward is over a longer period of time, it helps to diffuse the impression of "snowballing" of one faction.

    Regardless, though, enabling capturing territory or resources will eventually result in one side gaining dominance and requiring a game board reset. The only difference is with how long of a timeline are we talking about.

    However, I have been thinking more and more that I don't really think rewarding victory, at least not in the manner being discussed or traditionally used. If a developer wants a healthy, long-term PvP game, they need to not reward victory with power, because all that does is result in the people who win being able to get more power and thus keep on winning and thus getting more power and so on. I think that the only way such a system is sustainable is if a SINGLE loss resets the board, bringing everything back into balance.

    Victory can still give rewards, but I feel they should be cosmetic or reputational. Special cosmetic gear, achievements, and the like should be the only rewards for winning, because that's the only way you can slow down or prevent the exodus of players to the "winning side"; if there's no REAL benefit for winning all the time, then players will instead remain with the group they like.

    I understand that one of the points of such "empire PvP" games is to feel like you are helping to dominate the game world. But keep in mind that that exact mindset, that goal presented by the devs and embraced by the players, is exactly what is being discussed here as a problem. The "bill of sale" when you play an Empire PvP game is that eventually a faction will dominate the world.

    And I disagree that that's healthy for a game, but if that's the way players and devs want to build the game, they need to seriously think about the timelines, not about how to keep the sides "balanced". If they wanted factional balance, then why are they offering territory control?

  5. Excellent points. There's a lot of stuff to chew on in these comments.

    @Gevlon, that's a very interesting idea, having asymmetrical factions with different rules for point gain. The one problem I see is people not going for the optimum. For example, a number of poor players sticking with the "elite few" faction, even though they are getting less than 1 pt. While they're not getting many points, they will suck up some of the points in the pool, perhaps enough to drag down the distribution to the top players.

    The second problem, of course, is that your method requires mechanically evaluating performance, which is quite hard, especially when it comes to things like defending targets. Metrics can be gamed, and since your method doesn't require a win, the best course of action for one of the "elite few" is to do whatever nets her the most contribution, irregardless of if her team wins or loses.

  6. @Clockwork, that's a very good point. Losing not only gives you less rewards, it's also less fun gameplay.

    As for encouraging the losing team, I really liked the mechanic used in the Tarren Mill vs Southshore battleground. As you killed players, your Rank increased and you got a minor boost to damage/healing. Killing higher ranked people gave more points. Thus it lead to a very back-and-forth affair with good comebacks and an exciting finish.

  7. @Justin, that may be true. I'm just a casual observer of Eve Online. But from an outsider-perspective, it seems like the major powers have been the same for the last couple of years.

    As well, it doesn't feel like there's been a true war of conquest for a while. Perhaps something will evolve from the Phoebe changes.

  8. @RJ, excellent point about victory rewarding power being a positive feedback loop.

    As well, the idea of slowing things down, making battles much longer is also intriguing. For one thing, it reduces timezone advantages, where one side launches an attack at 3am and wins simply because there are no defenders.

    In some ways though, it comes down to feedback loops. If the feedback loop is positive, if victory makes future victories more likely, that leads to dominance. If the feedback loop is negative, if victory makes future victories less likely, that leads to parity and stalemate or stagnation.

  9. It's amusing that you refer to "stalemate" and "stagnation" as negatives. But in theory, these are the end goals of a balanced PvP system. After all, stalemate occurs by definition when the two sides are evenly matched.

  10. EVE is a good example of the system working. CFC are and have been for years the biggest because they are lead by the best, and that leadership core has done an amazing job at adapting to the changes in the game and keeping their 'foot soldiers' busy/entertained. EVE isn't, and never has been (on Tranq) one big blue circle, so the system is working.

    It also worked in DAoC, which was a different (pre-set factions) system for large-scale PvP.

    It worked in DF as well; no clan ever completely dominated, and even when one got close (EU server), everyone else got together to bring them down.

    Are there examples of bad systems leading to one side dominating? Sure. But the core issue isn't the idea of FFA PvP or faction PvP; the core issue is bad design. With good design either (or both) are viable.

    Hopefully Crowfall gets the design right.