 |
PlayerepJuly 5th, 2006 |
You might start seeing new sigs cropping up. Playerep is a new service that is in beta. Tiny still, but a good idea.
You can follow any responses to this entry through the RSS 2.0 feed. Responses are currently closed, but you can trackback from your own site.
Trackbacks & Pingbacks
Reader Comments
Promising, but here’s what I’d like to see in a reputation system, and I keep talking to different scripters in SL and urging them to get at this, before LL puts in some creepy reputation system that might be like omidyaar.net’s or something, where group-think artificially keeps “nice, positive posts of happy people” pushed to the top and and “not-nice, unhappy negative posts” of even legitimate criticism pushed to the bottom or simply ignored/unrated/low rated.
o Colour coding, for very rapid recognition, blue, green, yellow, orange, red.
o Not cumulative scores, where 10 negs can cancel out 10 pos, but where clusters of scoring are viewed on a rainbow or spectrum — I want to see if there is a 20/80 relationship or 1/100 relationship and see if gaming is occurring if I see friendship clusters all green-rating (this is what is wrong with every rep or BBB type proposal that comes along, friends whitelist friends and blacklist enemies)
o balloon/tagging so that you can click and read who put in a score, and what they said, their explanation for their vote of a neg or pos
o possible hue changes of deeper green or something to show pos rates or neg rates of numbers of avatars, i.e. 100 avatars who pos rate all with verified accounts gives deeper hue on the spectrum
o a HUD or easy UI that can bring up an easy colour click when you have each transaction, so customer service, price, delight, all get some clicks, not just f2f
o rating not only by accessing avatar or object but by searching in name list
o a product code that has that spectrum with the rates’ colours showing perhaps?
o building/appearance/social or whatever are too limited categories. Dispense with categories. A green is a green. It could mean good price, or good service, or good sex, whatever.
myspace for MMOs? that seems worth a shot…..
I can’t speak to some of the internal designs you suggest for SL, but in terms of general gaming the service offers an overall reputation score alongwith a specific reputation and several attribute scores for an individual title. The color coding is within the scoring system, and we don’t do the one-to-one negation you mentioned (agreed, that doesn’t work). And we are working on adding voting capability to certain MMO’s. I’ll forward you an invite if you wish. Thanks.
(this is what is wrong with every rep or BBB type proposal that comes along, friends whitelist friends and blacklist enemies)
Why is this wrong? Surely this is exactly what the system is for, to divide the population into groups showing who trusts or likes each other?
If the system actively showed groups who liked and trusted each other than there wouldn’t be anything wrong with it, but since the idea is to give you an overall picture of the qualities of a person the results are skewed.
Think of it this way, who is most likely the better person to group with, someone who is good enough at being part of a group that another person votes for them because of their abilities or someone who forms a voting circle with three other players? It can certainly be argued that the person with the voting ring is more social but since that isn’t the intent of a ‘grouping’ score not taking steps to eliminate voting rings decreases the value of the score.
Invitation only!? :\
Oh yeah, I did sort of a thumbnail outline for a similar in-game system about three weeks ago on my own blog (link here). The first half of the post is rattles on about griefers and dealing with them but as I wrote it I started thinking more about the overall handling of socially noxious behavior.
It doesn’t break down scores into categories and is really meant more as a tool to help CS identify problem players while allowing other players to find helpful players, but I would think it could fairly easily be adapted.
The primary problem with these systems is that they need to be opt-in (unless built into the game directly of course); I could see this working if, say, it included a WoW plugin that allowed for personal notetaking as well.
Re: Why is this wrong? Surely this is exactly what the system is for, to divide the population into groups showing who trusts or likes each other?
Because there’s no impartial jury, no independent investigation, no recourse or remedy to slander or libel. The “who trusts of likes each other” is an entirely specious and subjective category open to heavy abuse. Most people beginning these BBB’s in fact are just clustering together a tribe of friends from a niche group and declaring “we’re the smart ones, surrounded by idiots,” and approaching others, who often are their competitors in the same sector, with suspicion. It’s nothing at all like the real consumer’s reporting and advocacy that occurs at RL http://www.bbb.org
People can have all sorts of subjective reasons why they want someone on a quest or not, with this or that skill or attitude. They make snap, visceral judgements. In this, gaming is like sex — you either like the person and want sex with them, or you don’t. It’s hard to make it up. If you do, it’s not fun. Business is like this too, but it does have to have some more formalities and impartialities built over it to prevent just mobs of people knocking a person who they hate for entirely arbitrary reasons and harming their livlihood.
The BBB example isn’t comparable at all, I don’t think. I mean, we’re back in the territory of the posts I made on “trust,” but what this system is explicitly measuring is a history of interactions. There’s no such thing as impartiality there precisely because it’s measuring subjective experiences.
The BBB is equivalent to the server that Playerep is running; neither one checks up on reports and verifies them actively because it’s too much work. Both are vulnerable to reputation attacks of the sort you describe. Neither is truly an impartial jury…
But the BBB case, you can appeal and have something corrected in your record. And the BBB can also do things like give awards and so on…
Anyway, you’re mostly echoing one of the points I made here, the one about single-stage reputation systems.
I agree that these systems idealy should be built on convience if at all possible. Click on the person, give them a Yes vote, and go back to the game. A concern that is voiced in ATITD everytime someone wants to introduce such a system is that some people would seek out poor reputations, and thus it might encourage griefing instead of discouraging it.
I think it’s worth noting for reputation attacks that it’s one reason why we have both decay and non-linear scoring. Along with reputation weight and other things with reciprocal voting. If the concern is that people will max out their reputations, or just benefit (or be victimized) by spammers — that’s not possible with our system. We don’t reward prolonged homogeneity. And you will always be able to know who is voting for you and reciprocate (although we don’t reveal vote type). There is no inevitable reputation really.
As mentioned, gaming experiences like RL experiences are transitory. People’s behaviors and judgements will vary over time. You need to reflect that. We can’t enforce behavior in a game, or ever dictate what’s a “good/bad” transaction in a title. We can only reflect what a group of people think of someone over a prolongued period. Over many games even. But we also know the kinds of bias this kind of scoring can bring, and we’ve put in constraints.
>The BBB example isn’t comparable at all, I don’t think. I mean, we’re back in the territory of the posts I made on “trust,” but what this system is explicitly measuring is a history of interactions. There’s no such thing as impartiality there precisely because it’s measuring subjective experiences.
The reason I raise the BBB, is because periodically everybody at SL lurches through this debate. There’s yet another attempt by yet another gold-star bunch who say they are just simple people, trying to serve the community blah blah. Because it’s business and not all a game, the BBB example becomes more important than the questing stuff.
There’s actually something quite impartial about *lots* of subjective experiences of *lots* of people unrelated to each other going through the rating motions. Indeed, the whole concept of any open system is based not on absolute objectivity (truth is one, only one path to the truth) but on the multiplicity of subjectivity (many paths to the truth) that always keeps the door open — the truth might be somewhere in the median, and the better reputations are those which show balance, not fake nobility. (”A scratch embellishes an empty face,” Gorky).
Now, why do you all get so fastidious about things “spiralling out of chaos” (!) These are *games*. Can’t there be chaos in games?! The reputation system is merely one more game-within-a-game. We all had a lot of fun in TSO collecting all those balloons which fed into greening and acquiring new interactions. Or you got redlined or “gloved” and it sank your energy. Yes, people deliberately collected redlines or “gloves” (so-called because an enemy would go through an interaction to lay down a gauntlet at your feet) as bads of dishonour in mafias and such, and that was part of what made it interesting. TSO only got interesting for many people when they went through elaborate strategems to avoid gloving or conversely collect lots of gloves to end up on the “most bad” list. Will Wright had it set up to accommodate that wish to be “most bad” and have “most people hate me” as much as “most popular”. You could pick and chose where you wanted to be on the spectrum.
>The BBB is equivalent to the server that Playerep is running; neither one checks up on reports and verifies them actively because it’s too much work. Both are vulnerable to reputation attacks of the sort you describe. Neither is truly an impartial jury…
Impartial enough, as I said, throught the aggregate of multiple subjectivities which is still better than one truth imposed upon all. Looking at Playerep, I see it might be useful but pulling up an outside site each time and figuring out how to map values of SL on to it are challenging. For one, phrases like “plays well” when you want “provides good customer service” or some such.
>But the BBB case, you can appeal and have something corrected in your record. And the BBB can also do things like give awards and so on…
What’s most important about the RL BBB is that it really isn’t about celebrating businesses set in stone on golden tablets (that’s what many people imagine it to be, when they are thinking of some ideal Chamber of Commerce), but a consumers’ reporting system to gather information about businesses and post them. If some claim was wildly slanderous or uninformed, it might be filtered out, I’m not sure of their methodology. But they also have a requirement that you first make the complaint to the business and get their response as a check against that sort of abuse.
>Anyway, you’re mostly echoing one of the points I made here, the one about single-stage reputation systems.
Raph, I can’t possibly know if I’m echoing some point made in some back pages of a voluminous blog, you’ll forgive me. I wouldn’t think of burdening you with back pages of SL and long-winded posts about BBBs and whitelists, I just try to summarize the issues here.
Can’t there be chaos in games?!
Personally, I think it’s more appropriate in reality.
Sorry, Prokofy, that came out the wrong way. I didn’t mean it as a slam or anything, but just as a helpful link. That’s what I get for writing half-distracted by dinner. 
Personally I only believe in positive-only reputation systems, but I’m all for Playrep trying to innovate in this area.
Positive-only reputation systems are too limited. It’s true that they can help you find popular people but they do nothing at all for warning you who you should stay away from.
Additionally they are, in my opinion, easier to abuse. A person forms a voting circle with a group of like-minded individuals and they will have a good score regardless of how badly they act since there is no way for their score to be lowered.
Of course systems that allow negative votes can lead to social griefing which is why I think a workable system needs to be able to be able to give votes greater or lesser weight dependent upon the originator.
As I’ve said before I think you need to tie reputations to specific transactions such as “sold an item to” or “grouped with for >1 hour” that can be verified by the game itself. Like eBay really. And even then eBay gets gamed. If this system became important it would soon be so broken as to be noise only.
Imagine people selling + votes for in-game gold or on eBay. From pools of alts created just to trade reputation back and forth. Supported by other pools of alts created just to generate fake reputation. Worse than that, imagine the same sorts of stuff perpetrated to blacken a players reputation.
If it is some sort of transaction that both players have a vested interest in and invest either in-game assets or in-game time to then it will be much harder to “fake” anything, not to mention too risky given the value of one’s own reputation as (presumably) such transactions will only occur between legitimate characters. Even then I think there would be problems.
Prokofy, I asked a fairly terse question so I figure you deserve a decent reply for having answered it.
Most people beginning these BBB’s in fact are just clustering together a tribe of friends from a niche group and declaring “we’re the smart ones, surrounded by idiots,” and approaching others, who often are their competitors in the same sector, with suspicion.
I agree, but if the members of this (and every other) niche tribe can’t use the in game trust system to see that they trust each other what use is it? If one of this niche group invites a RL friend to the game he’ll mark him as trusted. Now would the rest of this niche group want to trust this new player? Probably, depending on how much they trust the opinion of the person who invited him, but if you only have one trust value the in game highlighting won’t work and they’ll have to do things the old way anyway. Yes guilds do perform this role, but very coarsely.
I’m in my own niche, and I’ve seen one or two of this group of players stealing a kill/ripping off my friend. I don’t want to trust the new guy, and I might like him to be marked in some way. People who trust my opinion might like to have my dislike for this group passed on.
Those people might also be able to say, “hey I kinda like those guys, why are they marked down?”. When they see it’s me they might decide to stop trusting me as much, or the system might handle that for them if they mark the other guys up.
The maths for all this may or may not be possible without ways to game it, I’m a fan of Advogato like systems (yes I play Eve..) which Raph noted in his previous post don’t work too well for newbies. I think there’s ways around that problem, but a system with a single answer can’t cope with even fairly simple situations.
In WoW PvP Alliance and Horde would likely mark their sides good fighters up and the other sides good fighters down, so they know who to kill first. Which side would the single answer favour? You could have different trust systems for each side, but in general groups can’t be split with offline information and people organsing themselves into groups is inevitable.
I actually do not know what a BBB is (nor am I currently interested in finding out), but it seems to me that Lacero and Prokofy’s systems are not that far apart. If I understand what they’ve said (and given my batting average this week, that is by no means a certainty), they’re essentially tracking a list of author-explicit annotations using some kind of scale.
If this system became important it would soon be so broken as to be noise only.
At the risk of making a statement that is easier said than done; don’t make the system important and filter the noise.
The first part of that statement is fairly easy to do. If it’s just a feedback system that lets you see a person’s score then it isn’t going to be gamed as heavily. If you make a system that gives big in-game rewards for a high positive score then you’ll see it abused more than if the benefit of a high positive score is that more people will be asking you for help (because you’re high on the list of helpful people).
Filtering the noise is harder, but not impossible. By deprecating the value of a vote based on proximity between people on the social web the impact of reciprocal links and voting rings can be reduced. The voting links people possess can also be used to determine the ‘trustworthiness’ of their votes to allow for additional corrections.
I think Tide is fighting the current and has the right tools to do it. It is easy to sit here and say this is something that can’t be done. Hell I would say it can’t be done, but then again this is something I recommended long ago when talking about Dungeons and Dragons Online. The idea of having a REAL DM to adminster the quests in game. Players could play as DM or adventurer… and a rating system such as Playerep would be in place to keep the bad apples out.
If this system became important it would soon be so broken as to be noise only.
Evan has a pretty good ideal there, but really I think you only have a problem when you are taking the data too seriously. If you look at a persons profile page and judge that person you are the one making the mistake. The idea is to formulate a community that feels responsible for their actions/inaction. It’s a stepping stone to growing a social circle more efficiently.
Final note. I think the word of mouth on this could be very good. Uber raider guild #1231 implements it and that gets their enemies into the site. Then it gets the guilds recruits into the site. All of a sudden you have a nice little circle of feedback to go through the grinder. And Tide seems very competant with the grinder.
Art (10)
Game talk (1372)
Gamemaking (43)
Mailbag (13)
Misc (181)
Music (81)
Open thread (22)
Reading (86)
The Sunday Poem (125)
Watching (37)
Writing (106)
All contents of this site are ©Copyright 1998- 2008 by Raphael Koster. All rights reserved.
The views expressed here are my own, and not necessarily endorsed by any former or current employer.
See popular posts »
About the blog »
A Theory of Fun
for Game Design
Available on CD
$14.99
Cat Fishing
Ash Grey T-Shirt
$16.99
"The world the way they thought it was..."
[...] Raph Koster posted a small snippet about a new beta service called Playerep. [...]
[...] Comments [...]
. We needed to do some throttling up front, but all is well. Please sign up ! Secondly, we’ve had some very positive feedback. It’s a very big deal when a homebrew project gets people like Lum and Raph Koster using and commenting on the service.Raph mentioning us on his blog was very generous and that has spawned some good discussion there. A quick thank-you also goes out to SirBruce as well of MMOG chart fame for helping with the invites. The reception has been positive with some thoughtful skepticism from a few folks.
Oh, you know her, “Miss Groupie Supreme” Yeah, you know her, “Vera Vogue” on parade Found at Raph’s place: Playerep. Read How it Works and the FAQ , then come back. You know, if you want. Why game producers don’t want to build global player rep systems into their game: Players are more apt to negatively review a player than