Agree with both of you: 1) we can predict transition point 2) strip helps us to initiate transition in order to improve stall characteristic at high angles of attack and low speed - without the strip it is more difficult to provide conditions for turbulent reattachment. This is what I have: using laminar algorithm I have Cdrag~0.004. In flotran I can turn on wall treatment properties only when I use turbulent model over the whole domain. With this options turned on, I have Cdrag~0.009. Reference value is 0.006. I guess, that these 2 results that I've got are top and bottom limits and the tuning for achievement reference value is confined in finding this transition over the airfoil (approx. at 40% of chord). Combining BL laminar flow over the leading edge and BL turbulent flow over the trailing edge will give me reference value. But I have a big problem: I can't set transition point in BL where soft should switch from laminar to turbulent wall treatment. I guess that wall treatment models are applied to all BL around the airfoil. Maybe you know how to do this in ansys/flotran or Star-cd? And one more question: as you suggested, we should know the locagtion of transition point on top and bottom surfaces of airfoil to adequately describe the flow. But how to know this? From real experimens? or we just point the location, see results, if they are wrong, change the location of transition...etc... doing it until results are valid?

For questions on the commercial codes you can try consulting the respective forums.

On guessing the transition locations: Sure you could play around with it and see which position gives you the best result, but that practice can't be called "prediction". Once again, you should look for more than just predicting the drag coefficient. A single parameter can always be "tweaked". Compare and analyse the pressure distribution over the airfoil.