The new catchbob scenario is well received by Pierre. He always complained by the simplicity of the task; now he is more convinced. His point is that Bob should be something credible (fire? objects? birds?...) and we have to think about what this task will prove. We can complexify the task with:

object: static - mobile
number of bob: 1 -5
let them play on different levels of EPFL
jogsaw? each player have a different information?
they have to keep a certain distance between them?

Anyway, Pierre was pleased to see that in our experiments, players without the tool self-declare their position. It would be great of we find positive/better results in the condition without the location awareness tool. That led us to think about a new independent variable: the accuracy of the synchronous location awareness tool. An incredible result would be to find that group performance is not affected by the positioning accuracy. I think I'll tackle this issue in the near future. People here at EPFL would be really interested in this kinf of results!

Other points:

Keep on doing some experiments...so that we can have 5 groups with the tool and 5 without.
Use Multi-Level Analysis? because it is hard (as I stated yesterday) to define what is at the individual level or at the group level.