Economists from the Erasmus School of Economics from the Netherlands have written a research paper on Deal or No Deal that reveals strong support for the behavioral “break-even-effect” of Thaler and Johnson. See

ssrn.com/abstract=636508

The interesting thing about Deal or No Deal is that it uses a series of gambles and thus shows the effect of previous outcomes. Contestants are less risk averse when they have seen their expected winnings tumble. "Losers" tend to continue playing the game even if this means rejecting bank offers in excess of the average of the remaining prizes.

Their study was covered on the front page of The Wall Street Journal last January, see http://people.few.eur.nl/vandenassem/fi ... -01-06.pdf

That's a really interesting article (If a bit sideways in places )

Cheers for posting it.

Very interesting. I read that a while back actually.

One player, Irene, did indeed turn down an offer above the mean of the remaining values after a disastrous run of bad luck, and most of the players dealing at sums that were mean for the board have been those with very strong boards. (Jessica, the first to have the quarter-million in her box, was one example; Lyndsey, who dealt at £15,500 with both six-figure values left in the final eight, was another.)

We have just received the first series of episodes from the UK and we are planning to include those in our analysis. Unfortunately it will take some time to watch all those (>60) episodes, to get all the relevant data into Excel and to do the analysis. I would be very surprised if the pattern does not show up for the UK, because the evidence so far is overwelming.

The initial research focused on the 26-case games, where offers are made after every case once only six are left. Does the very different structure to the UK game, with much greater risk in the later stages compared to these 26-case versions (especially in situations where one value is far higher than the others), cause a significant change in the analysis?

Also, there appears to be much greater variation in the offers in the UK version, and the offers will often be dramatically less generous when large sums of money (particularly the £250,000) are removed. The offers often appear to be influenced by psychology as much as mathematics, although I have been reliably informed that the Banker is given an upper and lower bound on the offers based on the remaining values. (Indeed, in one episode, the host relayed a message from the Banker implying as much and explicitly stating the offer was near the lower bound. This came after she was extremely tentative in rejecting the previous offer.)

The Banker also claimed to Andy (game 27, broadcast on 30 November 2005) that his £20,000 offer was the absolute lowest he was allowed to offer. Frankly I don't believe this given that there have since been worse offers on similar boards.

Van, if you'd like a copy of my Excel spreadsheet containing details of the order of the amounts revealed in every UK game so far, together with the offer amounts, just let me know.

The problem with the number of briefcases to be opened in the UK could be that we find less narrow bounds for the contestants risk attitude due to the "big steps" in the game that are taken with each round. However, this range not only depends on the number of briefcases, but also on the pattern in the bank offers, so I'm not sure about the exact consequence.

MisterAl, I'd like to receive your Excel file, that would be great. Of course we will watch all shows ourselves, but your information can serve as a preview and as a check. My e-mail adres is vandenassem@few.eur.nl.

It will take some time before I can post some outcomes, but I will leave a notification here.

Martijn