RE: Re: Responsibility
"Charlie Poole" <[email protected]> Sun, 15 Dec 2002 13:13:07 -0800
| Newsgroups | gmane.comp.programming.extreme-customering |
|---|---|
| Message-ID | <[email protected]> |
Ron, > >> Here's a question: consider a project that assigns a dollar > value to each > >> story, and another project that only knows, for any pair of > stories, which > >> one is worth more. > > >> What differences could we detect in the decisions made in these > >> two projects? > > > Lovely question. > > Thanks. I was rather proud of it myself. ;-> > > > In the planning exercise that you do - although you weren't > > there when I did it with Chet and Anne - we discovered that > > knowing the actual value function would allow optimizing the > > order of doing stories so that the most value was delivered > > first. The actual measure to optimize was the ratio of the > > value points to the level of effort points. > > Ah, but consider the question further in the light of that game. Keep reading Ron!@! > I suggest that > always knowing which story is larger may be sufficient to > optimize almost all > the way. And certainly knowing any linear scale (this one is > twice as valuable > as that one) would suffice, would it not? Yes, any linear scale providing an unbiased estimate of "real" value would produce exactly the same optimization. And I agree that a ranking is almost as good in theory. I'll go further - I hint at this below but I want to spell it out. I think a ranking is subject to fewer errors. People tend to know more clearly the value rank of their wants than they do the relative values. We don't get much practice in deciding how much more we like steak than salmon, but we do get a lot of practice in picking one over the other for an individual meal. > > However, the results that we came up with using only rankings > > were within a few points of the optimum and about twice - if > > my memory serves - the value we created without any information > > given in advance. > > > So having perfect knowledge of the "real" value of each story > > should give us a slight improvement over having only rankings. > > > The problem with this is that perfect knowledge of value is > > really not reachable - nor is perfect estimation of effort which > > also plays in the optimization formula. I'd guess that normal > > variations in both estimates would pretty quickly wipe out > > any theoretical advantage. Combining that with the cost of > > coming up with estimates of value, I pick simple ranking as > > the winner. > > > This doesn't say anything against the usefulness of Cost/Benefit > > calculations done in advance of approving the whole project. > > In that case, we are looking at the expected value of the whole, > > which is probably something more than the sum of the parts, and > > the central limit theorem is standing by to help us out. > > Exactly! Well, on average I'm sometimes exact. > Ron Jeffries > www.XProgramming.com > It's easier to act your way into a new way of thinking > than to think your way into a new way of acting. --Millard Fuller > > > > To unsubscribe from this group, send an email to: > extremecustomering-unsubscribe-hHKSG33TihhbjbujkaE4pw@public.gmane.org > > > > Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ > > > > ------------------------ Yahoo! Groups Sponsor ---------------------~--> Get 128 Bit SSL Encryption! http://us.click.yahoo.com/CBxunD/vN2EAA/xGHJAA/NhFolB/TM ---------------------------------------------------------------------~-> To unsubscribe from this group, send an email to: extremecustomering-unsubscribe-hHKSG33TihhbjbujkaE4pw@public.gmane.org Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/