« The Hidden Complexity of Wishes | Main | Lost Purposes »

November 24, 2007


My suggestion - no fees, but pay no interest on deposits. I think this would attract a lot more action and InTrade could earn their money by investing the deposits.

Alas I would prefer the opposite of Will - pay full interest on deposits, and set fees proportional to the actual transaction costs. This would allow more longer term claims.

I personally like the Betfair model... just a commission on profits made. It's simple, which is perhaps more important when it comes to actually getting a wider population of people to trade.

Admittedly, the lack of interest payment skews bets toward short-term turnover in this system, though, which won't help you in the '08 politics arena.

Haven't had time to look at these markets closely recently since there's no lack of more liquid and volatile ones out there, but selling longshots would be my preferred *Intrade* strategy. I mean, Gore was still at 5% for the nomination after announcing that he wouldn't be running.

Are you taking advantage of "advanced" margining, Patri? That is, among other things, being able to sell two 10% longshots for 80% margin (not 180%) because they're mutually exclusive outcomes and your worst-case loss is 80%. I'm surprised that these opportunities are still in place given the existence of that option. Also, with the fee issue if you assume that this bias is already in the price, changing your two ratios, the EV is closer to flat.

I agree with Robin on the interest question, although clearly that is potential revenue for exchanges and your typical punter is more interested in blasting 500 foot home runs than just making contact. It's a short-term revenue vs. long-term liquidity trade-off. Simplicity is preferable, but I don't think that not being paid interest really falls under "more simple".

Jed - I think that would avoid bias #2.

Jason - Selling longshots is what I did last election, but these transaction costs really eat into your profits. Selling multiple exclusive longshots is a good idea. Sure, the price may incorporate this bias - but that's a bad thing, because it means prices can no longer be directly interpreted as probabilities. The market provider should be trying to avoid adding bias.

Sure, the exchange likes earning the float. But that drag cuts down on volume and liquidity. Hopefully in time, competition will reduce these transaction costs.

Tradesports had the same deal then they went to 4% profit tax. Someone did analysis of how that altered the payouts depending on the trade price:


Agree with Jed as to the Betfair model of just charging winners. Simple and a real psychological advantage (even for people who 'get' that there is no free lunch, our behavioral biases are extremely hard to overcome and so it is easy to give into them in this case.)

Re margining, etc. - what is really needed (to develop long term to expiry) markets is a prime brokerage model where you separate credit/intial & variation margin from the execution/transaction element. They are different businesses with different capital requirements and expected returns. Alternatively you need to have a CCH (Central Clearing House) mechanism.

The problem here is - at least in the US - the underlying market needs to be seen as entirely legitimate/legal in order to justify the investments needed to develop either or both of these solutions.

The comments to this entry are closed.

Less Wrong (sister site)

May 2009

Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30