Yes, there's a variety of things you pick to build your character. Main weapon, secondary weapon 3 perks, 3 killstrike abilities, 1 deathstreak ability—Then to unlock camos for weapons, or additional effect from perks you must complete challenges. The game incorperates an experience system and levels up to 70. To have access to any of the aforementioned you must have a pre-req level. In any event, it is quite persistant power building.

I like MAG quite a bit. 256 player battles can be intense. :) It also features similar systems with talent trees, experience, gear, etc.

Probably unavoidable for 3d shooters to become more mmorpg like. From my UT days I remember the playing skill difference being vast, making 3d shooters as imbalanced as your average open pk mud. Do modern shooters have some kind of ranking system to separate players based on skill? This could also work for muds with a large player base, placing newbies in a newbie group that can only attach each other, with the best newbies moving on to the veteran group.

I always had the most fun playing with friends who were roughly around the same playing skill.

Well, usually modern first person shooters seem to match players based on A) latency and B) experience with the game in online play. I've played a few in the last few years that attempted skill matching systems based on your win-loss. Either way it's certainly more interesting than completely random.

A player skill-matching system would be pretty interesting for a PK mud, too, but I think you'd need a fairly sizable playerbase for it to be worthwhile. Having said that, you could perhaps have it match you up with offline players as well, and use an in-game mail or board system for arranging match times.

I think there are some algorithms for chess skill ranking, with a newbie starting out at 1000 points or so, and the top players being around 3000. Then you could make it so that killing a low rank player would give very low experience, possibly combined with diminishing returns for repeat kills. This however allows for losing ranks, and not everyone could achieve the highest rank.

You're probably thinking of the Elo rating system and it's used in lots of games besides chess, including World of Warcraft. We use it for our PvP rankings system, though I decided against linking it to any other advancement.

Various online role-playing games use Elo ratings for player-versus-player rankings. In Guild Wars, Elo ratings are used to record guild rating gained and lost through Guild versus Guild battles, which are two-team fights. The initial K-value was 30, but was changed to 5 in early 2007. Vendetta Online uses Elo ratings to rank the flight combat skill of players when they have agreed to a one-on-one duel. World of Warcraft used to use the Elo Rating system when teaming up and comparing Arena players, but now uses a system similar to Microsoft's TrueSkill. The game Puzzle Pirates uses the Elo rating system as well to determine the standings in the various puzzles. Also Roblox introduced the Elo rating in 2010.

—

TrueSkill is a Bayesian ranking algorithm developed by Microsoft Research and used in the Xbox matchmaking system built to address some perceived flaws in the Elo rating system.

In layman's terms, TrueSkill maintains a belief on the skill of each player; every time a player plays a game, the system accordingly changes its perceived rank and acquires more confidence about this perception.

A player's rank is represented as a normal distribution \mathcal{N} characterized by a mean value of ? (representing perceived skill, on a 0-to-50 scale) and a variance of ? (representing how "confident" is the system on the player's ? value). A such characterized \mathcal{N}(x) can be interpreted as the probability one player's "correct" ranking is x.

Players start with ? = 25 and ? = 25 / 3; ? always increases after a win and always decreases after a loss; ? always decreases. The extent of actual updates depends however on how "surprising" the outcome is to the system and on each player's ?. Unbalanced games, for example, result in either negligible updates or huge updates according to whether the system's "expected outcome" is actually achieved.

Factor graphs are used to "pack up" each teams to single (?,?) pairs on which the update formulas are run; the rank updates are then correctly distributed to each player.

Good paper on the system - TrueSkill Ranking System Details

The problems I have with these ranking algorithms is that you can decrease in ranking. I know it defeats the purpose of ranking, because it's not your true ranking unless it's factoring in losses. However, loss encourages grind in the goal of making up what you lost. This also goes against the true nature of competitiveness in showing the true rank of any given player versus other players.

I'm leaning towards a fixed system where you don't lose rank or ratings on defeat. Players increase in rank based on wins, players gain no rank increase based on looses.

Modern Warfare 2 doesn't have any loss of rank that I've perceived, and even when you lose, you still earn a little bit of experience in the Team games based on your performance. They refer to it as a Match Bonus. Wouldn't know about how it works in Free-For-All matches as I tend to enjoy the Team game types more. Specifically the 18 man Ground Wars.

Interesting, I'll check into it.

Now that I think about it more and more. I wouldn't consider this a ranking system. You simply cannot rank yourself versus other players without having some type of loss in ranking. I'm very scared of that because when you have loss, even in rankings, people tend to fight less frequently or fight dirty (ganking). The quality of PvP degrades more in a small community setting where everyone knows you and the impact of what you do in game is greater.

So it may be better to correctly level it as PvP Progression System where the experience you gain is put forth towards a level. It only measures your experience value not your average value versus other players in PvP. Then debate if there should be a second system based on a formula where you don't lose experience on death; only rank standing versus the entire player base. That way if the ranking degrades PvP, it can be ripped out.

In larger games ranking works well. You can get away with having poor ranks because no one knows you as much as they would in a smaller community. When you have such intense ranking in a smaller MUD community then people know you by name or even real life names. The risk of losing is much greater and players try hard not to lose as much when they have people watching them or tracking them competitively. (so you're clear on my reasoning).

So I'll fully admit to not reading the rest of the thread and that I just felt like chiming in something.

When I used to seriously play PVP MUDs (GodWars, thanks for that Kav) a few years back (pre-Dystopia stuff :) I was one of the much more aggressive / hardcore people. The game I called home didn't grant anything extra based on kills, you lost nothing for deaths, so the entire thing of 'status' or even your K:D ratio was nothing but bragging rights. That was worth EVERYTHING though; and led myself and a few other top players to do nothing but kill anything that moved, for the most part (gotta have 1 or 2 temporary friends, till they were worth status ;), in an attempt to boost our K:D ratios as high as possible. A few people would actually delete their characters after 15 or so lifetime deaths, because after that it was close to impossible to even maintain above a 0.90 ratio.

I think what I'm trying to get at is if you have a game that is reasonably balanced and a staff who is truly hands off except in proven instances of cheating / bug abuse / rule abuse then that's all you really need. Attract the hardcore PVPers with adverts and word of mouth and the ones who stick around will be the ones that PVP for nothing but bragging rights on K:D ratios most of the time, not because you devised some fancy new ranking system.

I do actually have a ranking scoreboard for a PvP minigame, which is reset each week, and that has no loss for losing. It's actually pretty popular, so I may extend the idea to the main game.


You'll get people fighting dirty just to climb up the rankings, regardless of whether or not there's loss. But your other point - people fighting less frequently to avoid loss - is also something I observed with artifacts (unique lootable items) in GodWars1, which made the feature rather self-defeating (the idea had been to promote PK, but it did the opposite).

Nothing is worse than having your higher ranking players not PvPing in fear of defeat. Same things happen with Quest Armor. You could loot it in Devil's Silence, those who spent time to attain it were very powerful in combat. They never left safe without a ganksquad. Then you have more or less a ranking system where it's not factoring in everyone else. You just increase or decrease based on the rank of the opponent you killed or lost against. Those with higher ranking stood to lose more if they died to someone who had much less. Then you had loads of mid-to-low ranking players always forming ganksquads to catch the higher ranking players 10 on 1 in order to get easy PKILL Points. All of this combined left a lot of people in hiding or safe hugging. The only people who would fight would only advocate 1-on-1 duels leaving clan-versus-clan battles out the loop.

I'm trying to avoid all of that as much as possible with no looting, no loss in PvP progression and more systems designed to encourage fighting without diminishing PvP quality. It's hard to do, because all of those aspects give value to PvP. If you have nothing to gain but pride then you just jump in, charge your opponents without thinking about the consequences.

I still favour the idea of scheduled events. Give the players some nice prizes to compete for, and make losing no worse than not taking part at all.

It makes me very happy to see notation like this being used, and bonus points for the LaTeX. :smile:

Thanks for the link, it was an interesting read. By the way, they mention that sigma actually increases before factoring in an outcome into assessing a player's skill. They do later say that the sigmas only decrease with match outcomes. I'm guessing that this means that it's possible for the net result to be no decrease or perhaps a slight increase, depending on how they parametrize the initial increase and final decrease.

What I was thinking of is using an Elo-ish rating system to determine experience gain. So basically you have two rankings, one is your experience, which cannot be lost, giving certains levels. The other is an elo-ish ranking determining your experience gain. This will make dying in fact beneficial as it lowers your rank, and subsequently increases your exp gains (relatively to your previous rank).

This could be abused by players who don't care about dying (two players of equal rank taking turns killing each other - but diminishing returns could still be added), makes advancement without dying much more difficult, and makes killing newbies less worthwhile.

I think it's a bit too complex for the progression system. I rather follow the traditional leveling system where it just pits your level versus your opponents level with an additional cap so if you kill the highest ranking player, you don't jump insanely.

Ranking, determining who the best players are on the MUD is a bit more fitting. That way you can pit your rank against everyone else globally. That might require a whole new statistic all together where it also factors on PvP wins or loses.

I did some debating and I think a ranking system may not be needed at all. Considering some of my top PvP games I like to play, many of them don't actually have ranking systems. They only have PvP progression systems to determine how many kills you have (experience not rank versus other players). Players who care about ranking, finding out the best players on the game, design their own ranking systems off-site. Therefore, I would provide information to the player only that shows them their averages where other players cannot see. Then leave it up to them if they want to make it public knowledge or not.