In Play NBA Final market feedback

Some thoughts:

1) Amazing job, Destry.

2) I hope that in future some of the "major events" include some international action. The Euro2008 soccer, the World Cup, even a cricket or rugby match. Most of all, the Olympics - there must be some events that would work well with in-play.

3) I agree with Donal on the hesitation about cutting off wagering too early. To my mind, it somewhat defeats the purpose of in-play. Half the fun of in-play is the excitement of turn-arounds, especially as the deadline approaches. That gets lost if we cut off too early -- it's just an opportunity to recalculate new likelihoods based on later information. Of course, in-play only works fairly if there is sufficient participation. Otherwise it's pretty well free money for the few.

I do see the value of cutting off with a minute left though. But whatever is chosen as cutoff, it must be clearly announced so as to be understood by the players. If I had one small criticism about the NBA questions, it was that in later games the cutoffs seemed a little too early, and so felt a bit arbitrary.

4) If there's a worry about people making too much profit on in-play questions (assuming the market has liquidity), I'd suggest maybe lowering the maximum single bet to $1000 for these markets. When a given outcome becomes extremely likely in the collective mind (say 90+%), the way to still make good money is to wager high (5-10 K). Although it's still a small risk, it's basically a safe bet. If we lower the max, 10 people could get in on the action, and the profits are spread around a little. (NB I'm not saying a limit on how much you can bet on a question overall, just a limit per prediction for in-play that is lower than the $10K for general questions.)

5) It would be nice if a group of in-play questions that all are to be suspended at the same time could be done so with one command and message by the admin or superuser.

6) If in-play is here to stay, and is seen as an important component of hubdub, then I think there will need to be some consideration of technical issues particular to it. Page reloads are troublesome now (probably because of the cross-site social bookmarks additions). Also, the important info is just the likelihoods and ones own predictions -- it would be great to be able to refresh these alone in-page.

But the biggest irritation for fast-paced predicting is the issue of what happens when the market moves between when you see the prediction page and when hubdub receives your prediction. (The cash-in feature works ok for this - it alerts you to the change but allows you with one click to continue or cancel.) But with predictions, I have found that you need to cancel the whole transaction, return to the question page, and start again, which kind of ruins the in-play fun.

I believe the whole "queueing of predictions in a market" protocol needs to be looked at again to make this work properly for in-play markets.