Peterson,
RE: "it was introducing numerous errors in the cost loading when the schedule was reforecasted"
It might be due to the way P3 handles period performance, if you do not follow the updating of pre-defined financial periods cost distributions that span over one period will be redistributed.
RE: "If you work with subcontractors and they submit a contract schedule then force them to increase resources to meet the contract completion, then you may be more accurate, but then you are not scheduling - you are simply shifting responsibility."
The Contractor is always responsible to the Owner, the Subcontractor is always responsible to the General Contractor, there is no shift in responsibility. If you require them to meet plans you are managing, the initial schedule is just the beginning.
If you want to increase the probabilities of meeting contract completion I suggest always target for early completion in the hope you make it on time. A blow some 10 inches below waist line (ugh!) to many American schedulers that are required to use all available contract time into their schedule or otherwise it would be considered a reduction in contract time. Quite a stupid requirement I have been forced to follow before, that targeting for early completion is good is a concept well known in Europe, not necessarily in the USA. My solution always was to keep two versions of the schedule, one booby trapped for handling claims (sure why not nailing the Owner if he does not allows you to make good plans and force you to have two plans) and the real one that targets for early completion (do not nail yourself).
RE: "intuition tells me they have an accuracy of +/- 20% for forecasting cost and time"
I would say as a general rule schedules are optimistic, if not then Parkinson Law will make them optimistic, plan for early completion always. The great % variation is on time estimates, at times due to poor performance by the Contractor, at times due to poor performance by the Owner that does not provide answers in due time, but mostly due to unforeseen conditions. I would say on general cost estimates have an accuracy of +/- 5%, if a Contractor estimate for fixed contract under competitive bidding. Less than 5% might get them in red and go out of business, more than 5% might make them not to get any competitive bidding job and get out of business as well.
About the SureTrak Topic Activity and WBS Activity I would like to add that an added bonus of this activity type would be that compatibility with MSP and other software shall improve, summary activities will be converted to this activity type instead of a hammock and links to summary activities will be transferred. Also WBS activities shall be converted to this type just without predecessors or successors to external activities. Microsoft users will feel comfortable with Spider if their work from MSP is fully compatible, and who knows perhaps one way for creating the Topic/Group activity or whatever might be similar to MSP by indenting or out-denting selected activities.
Best regards,
Rafael