OH NO! NOT INDIE GAMES!!!

You are just wrong, on many levels.

Hitman eredeti hozzászólása:

It looks like most of these games are amaturish,with a studio of 10 or less...I suspect that steam is going to get a reputation of pushing out crap....yes or i could be just wrong?

I wouldn't worry about it, I mean, it's not as if you're being forced at gun-point to but them. But it's not really a case of right or wrong - you simply have an incredibly shallow taste in games.

The purpose of Greenlight is to find the jewels in the pile of crap.
I think Valve is doing it wrong with the way greenlight is set up but it is probably better than nothing.

What would be the right way of doing it, Pnume?

- The feedback players give is binary. Binary is cheap on useful information. I mean information that helps you determine if the game will sale or not. There is probably no perfect system but they should be more creative. A rating system maybe.

- They should try to find what clues tell whether the game will be successful or not. At first it is trial and error only but with time you can try to find some decent correlations. At firstt I would use a rating system weighted by the profile of the voter ( Buy a lot or very rarely, buy day one or 3 years later with a 75% discount etc...)

- Screen the obvious trolls. They are a waste of time.

Binary feedback gives the information that is specifically useful here: "will somebody actually buy this?" and "will this help drive more sales to Steam in general." There doesn't need to be a rating system because most people vote either the highest 20% or lowest 20% anyway. Any high-traffic rating system driven by user response will show this. (see: Metacritic, the old youtube system)

Those clues are "consumer interest" which is gained from what I just pointed out.

They do.

@Erinduck:
It is true that a lot of rater gives ridiculously high or low rating to game but not everybody. It makes things harder though.

- A rating with a low spread (/5 or even /3) prevent this to some extent.
- This bias makes average rating irrelevant. Median or rating Repartition are not affected so much by that kind of behavior.

Another thing is that a lot of people vote out of sympathy. What greenlight needs is a way to screen them to some extent because their vote is meaningless. I expect they would end up a little bit above the middle rating.

When did things become so complicated?
If you like the game then vote for it.
If not then don't.
Or maybe I am too simple minded.
For people with very high standard for choosing game, I think they will find a hard time finding a game anywhere, not just greenlight though.

Pnume eredeti hozzászólása:

- The feedback players give is binary. Binary is cheap on useful information. I mean information that helps you determine if the game will sale or not. There is probably no perfect system but they should be more creative. A rating system maybe.

- They should try to find what clues tell if the game will be successful. At first it is trial and error only but with time you can try to find some decent correlations. At firstt I would use a rating system weighted by the profile of the voter ( Buy a lot or very rarely, buy day one or 3 years later with a 75% discount etc...)

- Screen the obvious trolls. They are a waste of time.

It is very likely that they have an internal behavioral system in place, but they will never give you access to it. And such a system are not telling them what is a good or bad game, only what percentage of their player base buys a given type of game etc.

Pnume eredeti hozzászólása:

@Erinduck:
It is true that a lot of rater gives ridiculously high or low rating to game but not everybody. It makes things harder though.

- A rating with a low spread (/5 or even /3) prevent this to some extent.
- This bias makes average rating irrelevant. Median or rating Repartition are not affected so much by that kind of behavior.

Another thing is that a lot of people vote out of sympathy. What greenlight needs is a way to screen them to some extent because their vote is meaningless. I expect they would end up a little bit above the middle rating.

A rating spread solves NOTHING because most votes fall into the top or bottom 20% and a third rating option won't do anything because there is NOTHING to be learned from a "maybe.
A binary system provides a more accurate view of whether or not a binary decision (to purchase) will be made.

And a lot of people do not vote out of sympathy. In fact, very few people do to the point where their votes are largely irrelevant. If that was the case at all the games that were Greenlit so far wouldn't have been.

@Erinduck

A binary system is certainly not MORE accurate. In the worst case it gives the same result.
Having a more complex system give them in the end more information to make regressions and find correlations. Maybe nothing good would come out of it. Who knows. But More information is always better than less information. The load of work on the voters side is not increased by this or at least not significantly.

People do vote out of sympathy. A lot of them. Just check the comments the voters makes on most pages. It leaves no doubt. This is speculation but I think this is the main reason why many greenlit games have been failure so far in term of sales. They were voted out of sympathy. Not great game people would buy right away but good enough for people to think that they deserved to be on steam anyway.

My opinion is that rating tends to be very high or very low mainly because people cares to rate only for game they either loved or hated. Average experience creates no passion. It is for games players have played.
Here on Greenlight it is only about maybe. If voter are passionate about a game it is a very good information. Let them vote high or low. Greenlight needs space for the vote of sympathy and it is somewhere in between. Right know they end up mostly on the Yes side.

Pnume eredeti hozzászólása:

A binary system is certainly not MORE accurate. In the worst case it gives the same result.

Actually when the statistic you are tracking is also binary then yes it is more accurate, you can only buy it or not buy it in this case, so adding more options will only muddy the results.

Pnume eredeti hozzászólása:

@Erinduck

A binary system is certainly not MORE accurate. In the worst case it gives the same result.
Having a more complex system give them in the end more information to make regressions and find correlations. Maybe nothing good would come out of it. Who knows. But More information is always better than less information. The load of work on the voters side is not increased by this or at least not significantly.

People do vote out of sympathy. A lot of them. Just check the comments the voters makes on most pages. It leaves no doubt. This is speculation but I think this is the main reason why many greenlit games have been failure so far in term of sales. They were voted out of sympathy. Not great game people would buy right away but good enough for people to think that they deserved to be on steam anyway.

You're dramatically overestimating how many sympathy votes there are by taking what a vocal minority says as a majority opinion.