Gaming the System
by lacy - June 20th, 2008
“You have to learn the rules of the game.
And then you have to play better than anyone else.”
~ Albert Einstein
.
“Any violation by you of the terms of the foregoing may result in immediate and permanent suspension or cancellation of your account and the
forfeiture of any amounts
otherwise due to you from the Company.”
~ namethis terms of use
.
This message has been brought to you by the Kluster Gaming Commission.
Gorblimey.
I’m confused. Has this affected knewsroom too?
I’m just wondering how exactly do you define gaming the system ?
I ask my friends sometimes who I invited to join to check out or invest in my articles does that count as gaming ?
Gaming the system would include such activities as creating multiple accounts in order to flood names with investment. We encourage inviting your friends so long as they are not imaginary
I’ve done the same thing, but i just checked out and there’s a guy on namethis that almost always wins, on the last 7 names he has been top 3 in 5 i think winning in one name he had the 1 spot and the 2 spot, fishy…..
i just tell my friends to chech it out, no false accounts
The trouble is that kluster isn’t big enough yet. If someone wins on namethis or gets articles up in knewsroom because their friends have invested in them, then that’s pretty harmless, isn’t it? And, in any case, they might be voting because they really like the story or suggestion. When kluster has more active members, and with the clever algorithm thingy going full pelt, the crowd should theoretically be large enough to overcome small statistical anomalies such as friends’ and well-wishers’ votes – at least that’s my take on the whole situation.
“You have to learn the rules of the game. And then you have to play better than anyone else.”
- Is the rules of the game available? I recall there was something for knewsroom, but namethis?
I think that if people vote for an article or name just because a friend has invested in it or was the one who posted it, it goes against the concept of identifying the name that people like most. But I agree that it would be near impossible to figure out if someone is investing because he likes the name or because he wants to help his friend. Even if there is some suspicious activity, the benefit of the doubt should be given. “Innocent till proven guilty”.
It might be a good idea to introduce some sense of human objectivity into the judging algorithm, and have the people paying namethis to have a say in which names they really like. Any algorithm, as long as you know how it is written, can be gamed.
@ Jon - agreed. As more users get into the mix, statistical anomalies start to wash out.
@ Wilson Lin - the rules of namethis are simple. Submit names, invest in names. One account per person.
I’m impressed with the rigorous way in which kluster have gone back and examined previous winners. It bodes well for the future. Keep it up!
@ lacy , thanks for the clear up ^_^
I still think we shouldn’t be able to invest in our own names.
I think the investment process is the most vital aspect of namethis, and has granted it a definitive “stock market” feel. The only real way to eliminate people wanting to help friends, family, and even names they feel will win because of the popularity of the submitter, is to completely remove money out of the equation. Many online forums that have point based systems are highly successful and have fantastic attendance and loyalty and yet do not incorporate a payout systems.
When money is involved things change a bit. A completely different aspect of a persons life becomes involved in their internet activity. A person that previously would have only used this site for fun and enjoyment is now using it as a venture, an activity that while fun and engaging is now also tied to their income. Now, the best change that was made was the lessening of the original Knewsroom reward- at the previous rate is was impossible for someone to not include earnings into their livelihood. A restaurant server could not have busted enough tables to come up with the possibility to win 150, or 300, or even 450 on a single day if multiple articles were chosen.
My son introduced me to this site after he read about it in a magazine, I think Readers Digest, and I come on once a day either from home or our office building and check the days submission. I am always more keen to invest my watts if I like one of my sons names than not- but that hasn’t stopped me from investing elsewhere. I am not a “kluster acolyte” in the sense that my goals are completely aligned with that of the Entrepreneurs that invented Kluster, nor are they polarized. I come to this site for a small amount of enjoyment and to support my son and occasionally my husband (though he is not much for naming–he wanted to name Drew “Butch’)
I think that ESPECIALLY as this site grows and finds more users who were not here at the beginning the community will be hard pressed to find users that are polarized to the exact goals and mission of the founders, maybe because their names wont go down in the internet history books or because they have their own goals and ambitions.
Instead the financial gains should be curbed and a stronger award system should be set in place. Cream always rises though, and not because it tricked the milk.
I respect kluster perhaps more than the majority of community supported websites I have participated in. I respect them for three primary reasons (and a boatload of secondary reasons).
1. kluster has shown an unparalleled commitment to share earnings with the community - something that very few web 2.0 companies are willing to do (cough…youtube).
2. I have rarely witnessed such willingness to consider, respond to, and implement feedback from the community.
3. kluster has been continually innovative and when something doesn’t work they are willing to change it, or come up with new ideas rapidly.
It is the combination of these elements that makes kluster (and its websites) great. But as Angela has pointed out, with a lucrative reward system, there will inevitably be people that are more interested in capitalizing on the pay-out, whether or not they contribute valuable content. At the same time, taking away the rewards lessens the quality of content, and robs the community.
Fortunately, there are ways to provide rewards and prevent fraudulent activity. Right now there are loopholes in the system. If someone has multiple accounts they could easily push their content to the top, and reap the rewards within their primary account. The easiest way to prevent this is through careful monitoring of IP addresses as well as suspicious activity (for instance accounts who continually invest in the content of a single person).
But that does not take care of what Angela is referring to, that is preferential investment. It is almost impossible not to notice that her son Drew has placed in the top 3 over ten times, the majority of those in first place. A closer look shows that Drews parents Don and Angela have invested almost exclusively in Drew’s name entries, often with investments that far exceed what I would suspect is an average community wide investment for name suggestions (no including an individuals own suggestions). While I don’t understand the complexities of the algorithm, these investments likely pushed Drews names to the top.
As I understand it, this type of preferential investment is not banned by the ToS. At the same time, if this is the case, then suggestions that won based on preferential investment are not providing an accurate reflection of the communities preferences for naming suggestions. This is problematic for obvious reasons.
I think there is a potentially easy solution to this problem: Place a limit on how much influence an individual could exert on a name with their investments. An example of this would be as follows: investments between 0-300 watts would exert relative influence over names, but all investments over that threshold would only exert the same influence as a 300 watt investment - the only difference being someone who invested 1000 watts in a name would receive a larger share of the rewards than someone investing 300 watts. With such a system, the influence of preferential investments would be minimized and the winners would better reflect actual name preference (rather than namer preference) by acting more like a vote, while still maintaining some relative influence.
There are possibly more complex systems that could analyze connections, or how frequently specific users invested in other users suggestions, but those are likely to be programming nightmares (if not impossibilities).
Regardless, I am confident that the kluster team will eventually get the system right (namethis is only a few weeks old!)
P.S. Thanks for adding a favicon to kluster!
@Gabriel — I love the idea of the 300-watt “vote cap”! (Here I will refrain from comparing this system to the many-donors-at-a-few-dollars fundraising that propelled my heartthrob Barak to the top.) We’ve just got to be sure that there’s no incentive to vote 300 once, twice, three times…
I also agree that it’s amazing that kluster is as flexible and fast-moving as it is. Further, the general attitude of transparency is completely amazing and could revolutionize business.
However, in fairness to all those who have been ignoring the client’s directives, or voting for their friends and relatives, let’s be clear that the current (now previous?) algorithm encouraged same. Those who did it were, as Einstein says, playing the game better than anyone else. It just wasn’t exactly the game kluster wanted them to be playing, so they’re fixing it.
@stacy: to clarify the “vote-cap” system would no be an actualy cap on how many watts you can invest in any name. It would simply cap the amount of influence your investment would have on the algorithm. In the “eyes” of the algorithm a 2000 watt investment would look the same as a 300 watt investment. The only advantage of putting more than 300 watts into a name would be the returns. If that name won, you would receive a reward proportional to your investment.
As far as fixing the problem of names that ignore the clients directive (or names that are poached from other companies within the same field), perhaps allowing comments or a flagging system could help solve the problems. Of course there are potential problems with both methods, but I think it’s worth a try. Another tool that might be useful would be a popup before name submission that reads something along the lines of: “Please make sure you have carefully read the clients directions,” and then displays the directions and Q&A. Not sure if that would do anything at all, but might be worth a shot.
@Gabriel: I got the vote-cap thing. I just wanted to be sure that repeated 300 votes wouldn’t count repeatedly as another “yes” vote in the algorithm. In other words, the program would know it was from someone who’d already topped out when it came to name-choosing.
And yes! yes! yes! to comments and flagging and the pop-up. All would be worth a try, imho. While we’re compiling a wish list, I’d love to be able to privately flag names I like so I can wait and vote nearer the end after most of the entries are in. It’s happened quite a few times that I invest in a clever name then see one I like better later.
Hmm. Two namethis projects finished a while ago (luxury spa & online teen resources) but we don’t see the winning entries yet?
And the winners are here! That was quite the delay–probably combing for lice.
Hey guys at kluster, in the project “Executive Search company name in the Media field” MediaHunter got second place, but I wrote Media Hunter as well, so how come for only being one single word the other person won, I thought the system didn’t let you write down equal names…