You say “The ideal aim of the game is for all of us to know each other and our codebase so well that we all agree first time around when we turn over the cards.”

I’d disagree with that. I would propose that if you are agreeing first time with too much regularity, then it is likely that you are missing out on one of the key values that collective estimation gives you – the ability to unearth assumptions, mis-assumptions, biased opinions etc. The point of planning poker is to stimulate more discussion that you otherwise may not have had.

So, for me, the “ideal aim of the game is to take as many rounds as necessary to ensure all reasonably forseeable considerations have been uncovered so that our estimate is as credible as possible”.

Just my 2 pence… :-)

The only problems we seem to have (at Huddle) are:

a) cards getting lost, or people forgetting them and just coming along using their fingers as a substitute. Using fingers is awful (I’ve done it) and defeats the point. All cards should be down on the table before everyone turns over so that no one is influenced in any way.

b) cards getting muddled and wasting time sorting them each planning session.

Two small things, but annoying. Your own design could be a good idea. Maybe name them for example so you can identify offenders who lose their set! Some sort of plastic case per set would be a good waste of money, but potentially quite cool for us geeks.

And then the golden rule if anyone forgets their plastic case, they have to go back upstairs to get it. The thought of having to go back up and down a set of stairs is probably a good enough deterrent to stop people forgetting.