I like the general idea — gamblers can do useful work for society — but one area that needs improvement is the kind of things that the participants will be betting on.
"IBM wants to offer the minimum amount that the candidate will accept," however the gamblers are not asked to figure that out, instead the "entire wager pool is awarded to the Turks who had the closest answers [to what IBM actually offered]".
Thus the smart gambler tries to figure out not what the candidate would accept, but what IBM is likely to offer. At this point the prediction market does not predict what the candidate will accept, but what IBM will offer.
One might be able to construct a prediction market where the gamblers incentive is to identify the lowest number the candidate would accept, but this is not it.
I know where the name comes from but still, the word Turk being used as the name of the unit workforce while making automation by using mass manual labor is a bit inconvenient. We already have an ambiguity with the name of our country in the English language, give us a break :)
I thought you were just being a stickler referring to Mechanical Turk, but then I saw the PDF, and it didn't actually refer to a casino in Turkey. Wow, this is some next level confusion right there. What were the authors thinking?
What's complicated is that a "mechanical Turk" isn't a person from Turkey, and a "Turk" isn't an anthropomorphic automaton. "Turkish" means "from Turkey", not "a place where mechanical Turks operate".
It's like calling Wikipedia "wiki". A wiki is already a thing, using the word for the set to refer to an element is confusing, although I guess this is the opposite, and would be a bit like calling MoinMoin a "Wikipedia".
If one can make this work this would be huge. One could offer casino style games at a house edge of 0% (house edge = the amount that the house wins on average on every bet). This would not only be a super compelling value proposition to gamblers, it would be a huge step towards eliminating problem gambling: A gambling addict that plays at a 0% house edge would not loose money in the long run.
I run a small bitcoin casino called fastbets.io. We should talk to see if there is a potential for a collaboration (feel free to email [email protected]).
> A gambling addict that plays at a 0% house edge would not loose money in the long run.
This isnt true because the house edge is the average return. the individual players can have a positive or negative edge depending on their skill. this would be like playing poker without a rake.
Alexjikim, you are 100% right. I was thinking about classic casino games that are pure games of chance (where my statement would be correct). However, the OP crablar wants to offer games of skill where alexjikim's analysis is spot on. Would still be a huge improvement over the status quo though.
I find that Wikipedia article on "Gambler's ruin" pretty misleading. What is not mentioned there is that in a 0% edge game, the party with the higher bankroll has a lower chance of loosing it all, but she'd be losing a lot in that case. If she wins (very probable) she'd only be winning a small amount. That's still a fair game.
The hard part is making work into a game. I think Michael and Jane Banks would agree this is the foremost reason they prefer Mary to other possible nannies.
Something like this could be extremely valuable inside an organization. A lot of very valuable knowledge is locked up by employees with no incentive to reveal it.
They provide entertainment value sometimes, and for winners they provide monetary value. This is like saying watching a movie or playing football has no value. The particularities of the game may create the value. E.g., poker is a well balanced game.
One potential problem is that instead of assessing problem objectively, since the payout is based on "most popular" answer, people will choose which answer they expect to be most popular, rather than the one they necessarily believe in. This drives down the value of the crowdsourced data - because it's not real "insight", is already tainted by the very gamification you're looking to monetise!
1) Make a casino with slot machines/etc. Connect slot machines to a stock market day trading system that acts at random on each lever pull (I.E. buying and then selling 10 seconds later, when the price is a bit different)
2) Gamblers will gamble as they usually do, but since the stock market grows on average by 7% a year, they should theoretically come out ahead.
3) Casino can take 3.5% for setting everything up.
A slot machine based on this principle exists and was presented at ICE last year (forgot the name). However keep in mind that the stock market grows at single digit percentage points per year. Slots machines only run for a couple of seconds, so there's no real profit made in that time frame. (As far as I can recall the rational for this system was of legal nature).
The whole appeal of a slot machine is that you might put in a dollar and get back a thousand. The prospect of a huge payoff is what keep s people pulling the lever.
Since that would essentially never happen on a 10-second buy/sell action on a real stock, I doubt any actual gambler would care to participate.
You can make payout tables similar to what they are now but instead of having a 1% house edge there's actually a small player edge. The house then covers the player edge and makes money by the stock investment.
The issues would be
1) If you assume a 7% annual return, the return on the 10 second investment would be extremely small. 7 percent / 250 working days per year / 8 market hours per day / 60 minutes per hour / 6 ten second periods per minute = 9.7e-6 percent return every 10 seconds. On a billion dollars play through in a year you'd be looking at a $97 return.
2) I don't know what commissions are like on the volume of small, short trades you'd be doing but I imagine they would be more than $0.0000001 cents per trade which means you're losing money.
You'd have to run it as a sort of recurring scheme where the current players are earning an expected value based on the investments of the previous years' players - start with a negative expected value until the backing fund achieves a decent annuity, then trend it toward positive.
"IBM wants to offer the minimum amount that the candidate will accept," however the gamblers are not asked to figure that out, instead the "entire wager pool is awarded to the Turks who had the closest answers [to what IBM actually offered]".
Thus the smart gambler tries to figure out not what the candidate would accept, but what IBM is likely to offer. At this point the prediction market does not predict what the candidate will accept, but what IBM will offer.
One might be able to construct a prediction market where the gamblers incentive is to identify the lowest number the candidate would accept, but this is not it.