mathcam wrote:I mgiht do this later, but I'm leaving soon, so I'll defer to someone else who's bored and is looking for something to do:
I wonder what the day 1 trends are. I would have suspected that the top bid on day 1 had been slowly decreasing, but then kerplunk went and took that theory and shred it to pieces last game.
That's becouse 100,000 is a nice round number. Everyone knows that the highest bid will be around that. And people don't want to go much lower becouse the last few games were very short and if someone was more than $5000 behind after the first round they had no chance to win. It did make sense to take the risk of dropping out in round 1 becouse if you were too conservative you lost all chance for a win anyway.
I'm not sure about games with more rounds but I think the highest bid in the first game has a very strong effect on the game in one round games and the tendency is towards 1,000,000/number of players. There is a strong reason of going just under the previous highest bid, of course. You don't want to be Too Greedy. But if you go just over and someone else does this too you already have a 50% chance of winning. Now the best part is, that there's no obvious way to cheat. Let's say you expect that at least one participant will go over the previous highest bid. But maybe not only one participant but two or three. You can't be sure and if you expect that at least someone will go over it really makes sense to take a risk.
One extreme is the one round game where winning is practically the only in game goal. The other extreme are very long games, where the first round goal is maximalizing the expected return. I think these games might have a downward trend in the first round highest bid, especially if the number of players is sufficiently low.
matchcam wrote:It's amazing how much a little collaboration could do...I wonder if it should be allowed. For example, if rOver and I agreed secretly to pick a random value in the range of 190,000 to 210,000 (accurately randomly), then as long as we actually do so, the one of us who picked the lowest is almost automatically going to win.
I was thinking along the lines of $1,000,000 and $999,999 but dropped the idea becouse I found no good way to enforce cooperation. I think 2 players will actually try this but one of them will cheat. More likely both of them
matchcam wrote:On the other hand, as long-term strategy, it doesn't work so well...because then people can safely pick 189,000 every time. So the new strategy for us would to be to randomly choose an relatively large value and then randomly choose wihin a 10,000 dollar interval around that value.
The highest bid would immediately jump to the pot/players range so there would be no way for the winner to return the favour. I won't even start playing Prisoner's Dilemma type games unless the other player can prove that the game will go on indefinitely. Or at least he, himself doesn't know when it will end.