Loading ...
Sorry, an error occurred while loading the content.

Partial points in sprint

Expand Messages
  • James Cooper
    Hi, If my team has an 8 point story which spanned two iterations, how do you credit the work? If you credit it to the iteration that it was finished in, you
    Message 1 of 17 , Feb 1, 2007
    • 0 Attachment
      Hi,

      If my team has an 8 point story which spanned two iterations, how do
      you credit the work?

      If you credit it to the iteration that it was finished in, you skew
      the velocity high for that iteration.

      If you try to spread it across then you get into quibbles about how
      many points to credit each iteration.

      I'm going to ask the team which way they would prefer, but I wanted to
      see whether anyone awards partial points or whether the consensus is
      it's a bad idea.

      thanks

      -- James
    • Matthew Knight
      i suppose its not answering your question, but one isn t supposed to split user stories across sprints, the user story should be broken down into a size
      Message 2 of 17 , Feb 1, 2007
      • 0 Attachment
         
        i suppose its not answering your question, but one isn't supposed to split user stories across sprints, the user story should be broken down into a size managable within single iteration.
         
        to answer your question, i'd say only allocate the points to when the user story was completed. this would be most akin to where, say, a user story had been started in sprint 1, but dropped due to not being able to complete it within that sprint, and then picked it up again in sprint 2. some of the work may have been done already, but in any case, its only done when its done, so sprint 2 is where you'd credit the work.
         


        From: scrumdevelopment@yahoogroups.com [mailto:scrumdevelopment@yahoogroups.com] On Behalf Of James Cooper
        Sent: 01 February 2007 19:25
        To: scrumdevelopment@yahoogroups.com
        Subject: [scrumdevelopment] Partial points in sprint

        Hi,

        If my team has an 8 point story which spanned two iterations, how do
        you credit the work?

        If you credit it to the iteration that it was finished in, you skew
        the velocity high for that iteration.

        If you try to spread it across then you get into quibbles about how
        many points to credit each iteration.

        I'm going to ask the team which way they would prefer, but I wanted to
        see whether anyone awards partial points or whether the consensus is
        it's a bad idea.

        thanks

        -- James

      • Laurent Bossavit
        James, ... I would - do the above - bitch about it - encourage people to break stories down into smaller ones A smooth burndown chart makes it easier to see
        Message 3 of 17 , Feb 1, 2007
        • 0 Attachment
          James,

          > If you credit it to the iteration that it was finished in, you skew
          > the velocity high for that iteration.

          I would
          - do the above
          - bitch about it
          - encourage people to break stories down into smaller ones

          A smooth burndown chart makes it easier to see when we'll cross the
          finish line, which makes it easier to manage the project. Smaller-
          grained stories make it easier to get a smooth burndown chart. Ergo,
          smaller stories make the project easier to manage.

          Partial points are lipstick on a pig - they'll make the burndown
          curve *look* smooth, but they're not actually helping you steer.

          Laurent Bossavit
          laurent@...
        • James Cooper
          ... Yes, this is the situation. We ve got at least one story that will not be completed in this sprint and will need to be completed in the next sprint.
          Message 4 of 17 , Feb 1, 2007
          • 0 Attachment
            On 2/1/07, Matthew Knight <matthew@...> wrote:

            > to answer your question, i'd say only allocate the points to when the user story was
            > completed. this would be most akin to where, say, a user story had been started in sprint
            > 1, but dropped due to not being able to complete it within that sprint, and then picked it
            > up again in sprint 2. some of the work may have been done already, but in any case, its
            > only done when its done, so sprint 2 is where you'd credit the work.

            Yes, this is the situation. We've got at least one story that will
            not be completed in this sprint and will need to be completed in the
            next sprint.

            thanks,

            -- James
          • Nathan Bobbin
            We don t do partial points at all. In the scenario you describe, we d get 8 points when the story is completed. I ve been calling this the squeegee effect.
            Message 5 of 17 , Feb 1, 2007
            • 0 Attachment

              We don’t do partial points at all.  In the scenario you describe, we’d get 8 points when the story is completed.  I’ve been calling this the “squeegee effect.”  The net result of the squeegee effect is that your later iterations will have a higher velocity than your earlier iterations.

               

              Smaller stories are the solution to the squeegee effect, but as we all know, they are not always possible – or at least not always apparently possibly.

               

              Cheers,

               


              Nathan Bobbin  |  Product Manager
              nathan.bobbin@...


              From: scrumdevelopment@yahoogroups.com [mailto:scrumdevelopment@yahoogroups.com] On Behalf Of James Cooper
              Sent: Thursday, February 01, 2007 12:25 PM
              To: scrumdevelopment@yahoogroups.com
              Subject: [scrumdevelopment] Partial points in sprint

               

              Hi,

              If my team has an 8 point story which spanned two iterations, how do
              you credit the work?

              If you credit it to the iteration that it was finished in, you skew
              the velocity high for that iteration.

              If you try to spread it across then you get into quibbles about how
              many points to credit each iteration.

              I'm going to ask the team which way they would prefer, but I wanted to
              see whether anyone awards partial points or whether the consensus is
              it's a bad idea.

              thanks

              -- James

            • Nicholas Cancelliere
              Can you demonstrate it to me? If not - then no credit. It s a bad practice to allow stories to span sprints. As a rule of thumb a story has to fit inside the
              Message 6 of 17 , Feb 1, 2007
              • 0 Attachment

                Can you demonstrate it to me?  If not – then no credit.

                 

                It’s a bad practice to allow stories to span sprints.  As a rule of thumb a story has to fit inside the duration of a sprint.  If you find that many times this isn’t working out then maybe you need to increase the size of your sprints (if you’re doing 1-week or 2-week sprints try 3 or 4 week ones, no more than 4 though).  Or you can (better choice) break down the story into smaller parts that can be demonstrated at the end of each sprint.  Don’t allow stories that span iterations.

                 

                The question you have to ask yourself is “If we released today – would this feature be there?”  If it’s not demo able then the answer is no, and you shouldn’t get credit on the story being “done.”  Now in the next iteration when you finish the work do you count 8 or 4?  If it were me I’d say “4.” 

                 

                The reason I say this is because I tell my team that velocity is what they’re able to estimate and deliver, so not having something done at the end of the sprint doesn’t cut it.  This gives more incentive to break down the stories into smaller parts.  I don’t like the idea of sliding the effort cost (squeegee as some say) because it defeats the purpose of velocity – which helps folks understand what the team can consistently delivery every sprint.  How would a product owner know what he can plan on getting in the next iteration or release when the numbers are bleeding over and spiking up and down?  It makes it very difficult to predict, and if it’s happening a lot I lose my confidence in the team and what they can reliably deliver.

                 

                -Nick

                 


                From: scrumdevelopment@yahoogroups.com [mailto:scrumdevelopment@yahoogroups.com] On Behalf Of James Cooper
                Sent: Thursday, February 01, 2007 1:25 PM
                To: scrumdevelopment@yahoogroups.com
                Subject: [scrumdevelopment] Partial points in sprint

                 

                Hi,

                If my team has an 8 point story which spanned two iterations, how do
                you credit the work?

                If you credit it to the iteration that it was finished in, you skew
                the velocity high for that iteration.

                If you try to spread it across then you get into quibbles about how
                many points to credit each iteration.

                I'm going to ask the team which way they would prefer, but I wanted to
                see whether anyone awards partial points or whether the consensus is
                it's a bad idea.

                thanks

                -- James

              • Ron Jeffries
                Hello, James. On Thursday, February 1, 2007, at 11:24:57 AM, you ... It s a bad idea. At the end of EVERY Scrum Sprint, all the stories are supposed to be
                Message 7 of 17 , Feb 1, 2007
                • 0 Attachment
                  Hello, James. On Thursday, February 1, 2007, at 11:24:57 AM, you
                  wrote:

                  > If my team has an 8 point story which spanned two iterations, how do
                  > you credit the work?

                  > If you credit it to the iteration that it was finished in, you skew
                  > the velocity high for that iteration.

                  > If you try to spread it across then you get into quibbles about how
                  > many points to credit each iteration.

                  > I'm going to ask the team which way they would prefer, but I wanted to
                  > see whether anyone awards partial points or whether the consensus is
                  > it's a bad idea.

                  It's a bad idea. At the end of EVERY Scrum Sprint, all the stories
                  are supposed to be Done==Done.

                  Ron Jeffries
                  www.XProgramming.com
                  Those who believe complexity is elegant or required don't understand
                  the problem. -- John Streiff
                • William Wake
                  ... My preference is to say - first iteration gets 0 for that story. For the next iteration, estimate the work remaining on the story at the start of the
                  Message 8 of 17 , Feb 1, 2007
                  • 0 Attachment
                    > If my team has an 8 point story which spanned two iterations, how do
                    > you credit the work?

                    My preference is to say - first iteration gets 0 for that story. For
                    the next iteration, estimate the work remaining on the story at the
                    start of the iteration, and get that much.

                    I think of the velocity as "amount our team can commit to and
                    deliver", not "amount of work we should get credit for." Some people
                    definitely want "credit". But I generally want to drive toward smaller
                    stories that are more likely to have the essential value and which
                    we're more likely to deliver.

                    --Bill Wake

                    --
                    Bill Wake William.Wake@... www.xp123.com
                  • James Cooper
                    Hi Ron, Understood. Does this create load fluctuations when using yesterday s weather to plan the next sprint? For example: Sprint 1: - Story A: 2 points
                    Message 9 of 17 , Feb 1, 2007
                    • 0 Attachment
                      Hi Ron,

                      Understood.  Does this create load fluctuations when using yesterday's weather to plan the next sprint?

                      For example:

                      Sprint 1:
                       - Story A: 2 points (done)
                       - Story B: 1 point (done)
                       - Story C: 8 points (not done)

                       Sprint 1 velocity = 3 points

                      Sprint 2 - carryover unfinished work and add 3 points of new stories
                       - Story C (continued): 8 points (done)
                       - Story D: 3 points (done)

                      Team gets finished with these with 5 days left (story C was close to completion at end of sprint 1), so we add:

                       - Story E: 2 points (done)

                       Sprint 2 velocity = 13 points

                      Sprint 3 - add 13 points of new stories?  Yesterday's weather is skewed by the carryover of the 8 point story.

                      thoughts?

                      -- James


                      On 2/1/07, Ron Jeffries <ronjeffries@...> wrote:

                      Hello, James. On Thursday, February 1, 2007, at 11:24:57 AM, you


                      wrote:

                      > If my team has an 8 point story which spanned two iterations, how do
                      > you credit the work?

                      > If you credit it to the iteration that it was finished in, you skew
                      > the velocity high for that iteration.

                      > If you try to spread it across then you get into quibbles about how
                      > many points to credit each iteration.

                      > I'm going to ask the team which way they would prefer, but I wanted to
                      > see whether anyone awards partial points or whether the consensus is
                      > it's a bad idea.

                      It's a bad idea. At the end of EVERY Scrum Sprint, all the stories
                      are supposed to be Done==Done.

                      Ron Jeffries
                      www.XProgramming.com
                      Those who believe complexity is elegant or required don't understand
                      the problem. -- John Streiff


                    • Steven Gordon
                      Calculate velocity as the average of the last N iterations where N is something like 3 or 4 instead of 1. Steve
                      Message 10 of 17 , Feb 1, 2007
                      • 0 Attachment
                        Calculate velocity as the average of the last N iterations where N is something like 3 or 4 instead of 1.

                        Steve

                        On 2/1/07, James Cooper < jamespcooper@...> wrote:

                        Hi Ron,

                        Understood.  Does this create load fluctuations when using yesterday's weather to plan the next sprint?

                        For example:

                        Sprint 1:
                         - Story A: 2 points (done)
                         - Story B: 1 point (done)
                         - Story C: 8 points (not done)

                         Sprint 1 velocity = 3 points

                        Sprint 2 - carryover unfinished work and add 3 points of new stories
                         - Story C (continued): 8 points (done)
                         - Story D: 3 points (done)

                        Team gets finished with these with 5 days left (story C was close to completion at end of sprint 1), so we add:

                         - Story E: 2 points (done)

                         Sprint 2 velocity = 13 points

                        Sprint 3 - add 13 points of new stories?  Yesterday's weather is skewed by the carryover of the 8 point story.

                        thoughts?

                        -- James


                        On 2/1/07, Ron Jeffries < ronjeffries@...> wrote:

                        Hello, James. On Thursday, February 1, 2007, at 11:24:57 AM, you


                        wrote:

                        > If my team has an 8 point story which spanned two iterations, how do
                        > you credit the work?

                        > If you credit it to the iteration that it was finished in, you skew
                        > the velocity high for that iteration.

                        > If you try to spread it across then you get into quibbles about how
                        > many points to credit each iteration.

                        > I'm going to ask the team which way they would prefer, but I wanted to
                        > see whether anyone awards partial points or whether the consensus is
                        > it's a bad idea.

                        It's a bad idea. At the end of EVERY Scrum Sprint, all the stories
                        are supposed to be Done==Done.

                        Ron Jeffries
                        www.XProgramming.com
                        Those who believe complexity is elegant or required don't understand
                        the problem. -- John Streiff



                      • Ron Jeffries
                        Hello, James. On Thursday, February 1, 2007, at 8:53:51 PM, you ... Look at the list. Estimate how much you can get done. Draw a line. Look each other in the
                        Message 11 of 17 , Feb 1, 2007
                        • 0 Attachment
                          Hello, James. On Thursday, February 1, 2007, at 8:53:51 PM, you
                          wrote:

                          > Understood. Does this create load fluctuations when using yesterday's
                          > weather to plan the next sprint?

                          > For example:

                          > Sprint 1:
                          > - Story A: 2 points (done)
                          > - Story B: 1 point (done)
                          > - Story C: 8 points (not done)

                          > Sprint 1 velocity = 3 points

                          > Sprint 2 - carryover unfinished work and add 3 points of new stories
                          > - Story C (continued): 8 points (done)
                          > - Story D: 3 points (done)

                          > Team gets finished with these with 5 days left (story C was close to
                          > completion at end of sprint 1), so we add:

                          > - Story E: 2 points (done)

                          > Sprint 2 velocity = 13 points

                          > Sprint 3 - add 13 points of new stories? Yesterday's weather is skewed by
                          > the carryover of the 8 point story.

                          > thoughts?

                          Look at the list. Estimate how much you can get done. Draw a line.
                          Look each other in the eyes and ask if you can do this. Ask if you
                          WILL do this. If yes, go ahead.

                          It's about the people, in my opinion, not the numbers.

                          Ron Jeffries
                          www.XProgramming.com
                          Speak the affirmative; emphasize your choice
                          by utterly ignoring all that you reject. -- Ralph Waldo Emerson
                        • Alexey Krivitsky
                          James, You should avoid partial metrics as much as you can, simply because: 1. it is hard to know if something is done for 55% or maybe 70%, much easier to use
                          Message 12 of 17 , Feb 2, 2007
                          • 0 Attachment
                            James,

                            You should avoid partial metrics as much as you can, simply because:
                            1. it is hard to know if something is done for 55% or maybe 70%, much easier to use a binary metric: not done vs.done, you don't want to fool somebody with your predictions based on the velocity, right?

                            2. you might be willing your team to be focused on finishing the stories 100%
                            in order to avoid a queue of unfinished tasks which is a major impediment in projects

                            if you don't calculate half-done things in the current sprint but move it to the next,
                            then your next sprint's velocity will be increased (if the unfinished things are done then finally),
                            in this case you can try to use an average velocity (over last 2-3 sprints) to predict your release date
                            with more accuracy.



                            On 2/1/07, James Cooper <jamespcooper@...> wrote:

                            Hi,

                            If my team has an 8 point story which spanned two iterations, how do
                            you credit the work?

                            If you credit it to the iteration that it was finished in, you skew
                            the velocity high for that iteration.

                            If you try to spread it across then you get into quibbles about how
                            many points to credit each iteration.

                            I'm going to ask the team which way they would prefer, but I wanted to
                            see whether anyone awards partial points or whether the consensus is
                            it's a bad idea.

                            thanks

                            -- James


                          • jay_conne
                            James - I like many of the responses you have received and would like to add one note to Bill s response. I have found that POs, sponsors and teams are willing
                            Message 13 of 17 , Feb 2, 2007
                            • 0 Attachment
                              James - I like many of the responses you have received and would like
                              to add one note to Bill's response.

                              I have found that POs, sponsors and teams are willing to be tough on
                              themselves about 'done' and taking credit. We have a conversation
                              about building integrity into the process so they can honestly get the
                              best of what Scrum offers.

                              Most people respond well to the honesty principle - not kidding
                              themselves or others.

                              Jay
                              www.jconne.com


                              --- In scrumdevelopment@yahoogroups.com, "William Wake"
                              <william.wake@...> wrote:
                              >
                              > > If my team has an 8 point story which spanned two iterations, how do
                              > > you credit the work?
                              >
                              > My preference is to say - first iteration gets 0 for that story. For
                              > the next iteration, estimate the work remaining on the story at the
                              > start of the iteration, and get that much.
                              >
                              > I think of the velocity as "amount our team can commit to and
                              > deliver", not "amount of work we should get credit for." Some people
                              > definitely want "credit". But I generally want to drive toward smaller
                              > stories that are more likely to have the essential value and which
                              > we're more likely to deliver.
                              >
                              > --Bill Wake
                              >
                              > --
                              > Bill Wake William.Wake@... www.xp123.com
                              >
                            • James Cooper
                              Hi everyone, Good discussion. I have a sprint review today followed by a sprint planning meeting, so these thoughts will help me focus our internal
                              Message 14 of 17 , Feb 2, 2007
                              • 0 Attachment
                                Hi everyone,

                                Good discussion.  I have a sprint review today followed by a sprint planning meeting, so these thoughts will help me focus our internal discussions here about how to credit this sprint and allocate work for the next one.

                                -- James

                                On 2/2/07, jay_conne <jay@...> wrote:

                                James - I like many of the responses you have received and would like
                                to add one note to Bill's response.

                                I have found that POs, sponsors and teams are willing to be tough on
                                themselves about 'done' and taking credit. We have a conversation
                                about building integrity into the process so they can honestly get the
                                best of what Scrum offers.

                                Most people respond well to the honesty principle - not kidding
                                themselves or others.

                                Jay
                                www.jconne.com

                                --- In scrumdevelopment@yahoogroups.com, "William Wake"
                                <william.wake@...> wrote:
                                >
                                > > If my team has an 8 point story which spanned two iterations, how do
                                > > you credit the work?
                                >
                                > My preference is to say - first iteration gets 0 for that story. For
                                > the next iteration, estimate the work remaining on the story at the
                                > start of the iteration, and get that much.
                                >
                                > I think of the velocity as "amount our team can commit to and
                                > deliver", not "amount of work we should get credit for." Some people
                                > definitely want "credit". But I generally want to drive toward smaller
                                > stories that are more likely to have the essential value and which
                                > we're more likely to deliver.
                                >
                                > --Bill Wake
                                >
                                > --
                                > Bill Wake William.Wake@... www.xp123.com
                                >


                              • Guy Davis
                                ... Hi James, Our team uses XPlanner to track our stories and tasks. When we first started adopting Scrum we ended up with many stories having some completed
                                Message 15 of 17 , Feb 2, 2007
                                • 0 Attachment
                                  James Cooper wrote:
                                  > If my team has an 8 point story which spanned two iterations, how do
                                  > you credit the work?

                                  Hi James,

                                  Our team uses XPlanner to track our stories and tasks. When we first
                                  started adopting Scrum we ended up with many stories having some
                                  completed tasks and some not completed. XPlanner offers a "Continue"
                                  feature which creates a duplicate of the story and its uncompleted tasks
                                  in the next iteration. The effort spent in the current iteration is
                                  recorded in the existing story while the remaining effort is set as the
                                  estimate on the new story in the next iteration. This avoids skewing
                                  the effort estimates in the current or next iteration.

                                  That said, we've been doing Scrum for a while now and it is very rare
                                  for us to need to Continue a story anymore. We're much better now at
                                  delivering on our commitments, but it took a while to get there.

                                  Guy
                                • Michael James
                                  ... But remaining effort doesn t equal original effort minus effort spent ... not unless you had perfect estimates in the first place (ahem!). This approach
                                  Message 16 of 17 , Feb 2, 2007
                                  • 0 Attachment
                                    --- In scrumdevelopment@yahoogroups.com, Guy Davis <davis@...> wrote:
                                    > The effort spent in the current iteration is
                                    > recorded in the existing story while the remaining effort is set as the
                                    > estimate on the new story in the next iteration.

                                    But remaining effort doesn't equal original effort minus effort spent ...
                                    not unless you had perfect estimates in the first place (ahem!).
                                    This approach seems to buy into the misperceptions of traditional
                                    project management that lead to 90% done half the time.

                                    It would make more sense to re-estimate the remaining work.

                                    --mj

                                    Michael James
                                    Danube Technologies, Inc.
                                    http://danube.com/blog/michaeljames/
                                    http://scrumworkspro.com
                                  • Michael James
                                    ... Sorry -- in the case of a backlog item it makes more sense to take zero credit for the item the first time, and then credit for the original
                                    Message 17 of 17 , Feb 2, 2007
                                    • 0 Attachment
                                      --- In scrumdevelopment@yahoogroups.com, "Michael James" <michael@...> wrote:

                                      > It would make more sense to re-estimate the remaining work.

                                      Sorry -- in the case of a backlog item it makes more sense to take zero
                                      credit for the item the first time, and then credit for the original
                                      (pre-committed) estimate the second time.

                                      I was thinking about task estimates. Tasks should be reestimated from
                                      one day to the next rather than trying to subtract "actual" hours worked.
                                      Some teams prefer not to use task estimates at all and just count them
                                      as ones and zeros.

                                      --mj

                                      Michael James
                                      Danube Technologies, Inc.
                                      http://danube.com/blog/michaeljames/
                                      http://scrumworkspro.com
                                    Your message has been successfully submitted and would be delivered to recipients shortly.