[kplato] definitions of effort and risk

[bilbo]

On 21 Jun 2001, at 19:26, Jim Sabatke wrote:

> Normally, the effort column would be modified to include risk.
> 
> By entering values in the Pessimistic, Expected and Optimistic columns; 
> and by including a column for risk level (confidence level) the effort 
> value would be calculated.  In the project management discipline, the 
> effort value is normally calculated as follows (with a 99% confidence 
> level):
> 
>                 Effort                           + 4 Effort            
>       + Effort
>                         pessimistic                       Expected    
>             Optimistic
>           -------------------------------------------------
>                                                                  6
> 
> 
> There are other calculations if the risk is greater, which it normally 
> is for software
> 
>                 E(p) + 4 E(e) + 2 E(o)                                
> E(p) + 4 E(e) + E(o)
>                 ----------------            or even       ---------------
>                                 7                                        
>                           3.2
> 
> The last formula generally gives a 95% confidence level.  It is possible 
> easily calculate std. dev. from the formulas and do a Z score lookup to 
> get an exact confidence level; but the formulas listed are considered 
> adequate for project work.
> 
> Accordingly, the expected effort field should have the first entry.  The 
> effort field will be the expected effort value unless the other (p) and 
> (o) efforts and confidence level have values entered; then the effort 
> value will be computed.
> 
> Does this make more sense?

Yes - even though the calcs above aren't very clear - that's the fault 
of ascii email. I'd just read 
http://www.koffice.org/kplato/docs/x94.html, so I can see what you 
mean.

The thing I'm wondering about is how/when to assign a task to high 
or normal risk, presumably this is a judgement call. Is it possible to 
use the values of E(e), E(p) and E(o) to determine a suggested risk 
which can then be overwritten?



regards,
        Bill