I was thinking about this as I made my predictions for the new league we have:

 The average person spent a ton of VGC $ on the first prediction league. If we do wind up doing 3 games per week, plus bonus editions, it will definately strap alot of people with their VG$.

Not only this, there would be certain easy advantages each week by betting on specific games that are ultra-predictable with their weekly changes: Wii Sports, Wii Fit for Japan, and Wii Play in the US, and so on.

 So what should we do about games with easy exploits? I mean, if this is the way we do the leagues, it'd be way too easy for someone to bet $1,000 VGC or the max amount on those given games, and wind up with a huge advantage over everyone else.

 Or:

Do we ban specific titles such as Wii Fit/Sports/Play in Japan, and Play in the US, and other too-stable games (such as Guitar Hero 3/Rock Band) from the charts, as it'd be vastly too easy to predict them most weeks? Or am I crazy?

Or:

Do we limit the number of games per week, and focus on major weeks in given territories with major notable releases?

I just figured I'd ask before we continued on this new system. I really do like the new system, but it's going to be too easy for someone to spend 90% of their cash on games that they're almost always guarenteed to get 100% on, and double their money each week.

An interesting point. How about making two classes of predictions in a game: new releases and already released? Then you could limit the already released games to some small denomination (say VG$50).

New releases [games in 1st, 2nd or 3rd week] use this system now perhaps... but any games older than 3 [or 5?] weeks should have a different system where you need to be considerably more accurate.

That would split it nicely because any game important enough to be on the league after 5 weeks will likely be an easy title to predict. edit: iois system could work better though, except not $250 because thats still a pretty big amount to say Wii sports only goes up/down by a few thousand in normal weeks.

I am one of those who is banking their cash on an easy to predict title. I can't think of a reason on not to do it though. I am forced to bet my VG$ on the predictions so I need a way to cover my bets and stay at a level I like to keep playing.

The ony fix I can think of is to simply exclude the easy titles, but that becomes problematic for another reason. If I hadn't wagered big on Wii Fit, I would have done so on the DS. It is very stable and has a huge base number so even being off by a few thousand makes it a very profitable pick. It floats around 100k never being more than 20k off one way or another without a new release to push it. The Wii and PSP are looking to be easy money as well for similar reasons, but to a lesser extent.

The short version is I think it is impossible to fully get out of the system. The only way to limit it is to cap how much you can bet. Personally, I would like it better if I could not bet on anything, but that is not the way it is going to go.

You select the games to put in every week, right? I can't think of any single way that would be better than others, so how about you choose different types every week, and listen to the feedback? Just make sure no two rounds are the same, and see what people liked.

Well, it's upto ioi or Parokki's sugguestions. Either limit older titles (of which, the only real issues are weekly staples, as most holdovers aren't too bad), or exclude weekly staples most weeks.

Theres an easy way to bring people to bet their vg$ to more risky titles: If the overall accuracy for a title is very good, than it reduces the amount of VG$ you get in return. At the moment (according to the rules http://vgchartz.com/predictions/rules.php) the accurracy is divided by 50 to calculate your winnings. For a net profit you need at least 50% accuracy. If you instead use the average accuracy over all participating members it would result in a higher risk to bet on easy predictable games. Say a easy to predict game (WiiFit) will have a average accuracy of 90%. In this cas you need 90% to get a profit. On a harder to predict game (say DMC4) the average accuracy is 60%. You get a profit if you have a accuracy higher than that. If it works like that, people wouldn't bet the most vg$ on easy predictable stuff.

Perhaps only include weekly staples when there are not enough other games?

How do you choose them now? is it just new releases plus high charters?

 Hardware will be just as bad as the weekly staple titles though. ioi's suggestion is the best solution.

Or you could give different returns for different titles. That would be a programming nightmare though and I doubt ioi is up for it. By that I mean pay people base on an accepted 90% accuracy instead of 50% for the easy to predict or something similar. It kills the winnings and increases the risk.

I think there should be a multipler based on risk...

Something like sports game betting...

Less risky predictions give you a 1.1 multiplier while risk predictions give you 2.5 or 3.0 multipliers.