Planning game confusion

[pcwijntjes2000]

I have a question regarding the planning game. From Kents book I
understand that working with points is "not done" anymore, but I
keep getting confused on velocity vs estimates. I have a small
example:

Assume the following situation:
Initial release plan:
Estimates from planning game
Story 1: 3 days
Story 2: 5 days
Story 3: 2 days
Story 4: 2 days
Story 5: 3 days

Story 1,2 and 3 are put into the first iteration. We estimate our
velocity at 10 days per iteration
After iteration 1 story 1 and 2 are finished, story 3 has consumed 3
days.
Based on new insights we now think that story 3 and 4 each will cost
a total of 4 days of work.

What does this mean for the second iteration?
- What will be our velocity based on yesterdays wether?, my
suggestion:
o 8 days since we finished story 1 and 2
- What will be our new estimates for the remaining stories, my
suggestion:
o Story 3 : 4 Days (can we say 1 remaining day?)
o Story 4: 4 Days
o Story 5: 3 Days

Is this correct? It seems that when we use this approach the error
in estimating story 3 is used twice in the calculation proces (for
lowering the velocity and for an increase of the estimate). And
shouldn't we re-estimate story 5 as well?

anyone?

[Laurent Bossavit]

Peter,

Post by pcwijntjes2000
After iteration 1 story 1 and 2 are finished, story 3 has consumed 3
days.

You're not saying that story 3 is finished - I assume it's not ? Then
your velocity is 8 (3 from Story 1 + 5 from Story 2).

Post by pcwijntjes2000
Based on new insights we now think that story 3 and 4 each will cost a
total of 4 days of work.

Initially, your total scope was 3+5+2+2+3 = 15. Now, your remaining
scope is 4+4+3 = 11. You have at least one more iteration to go (11 >
8), with the project ending sometime during the third iteration. That
is your best estimate based on actual evidence of how long the
project will take.

Post by pcwijntjes2000
What does this mean for the second iteration?

Assuming stories are ordered by business value, you'll plan Story 3
and Story 4 (totaling 8) for the second iteration.

Post by pcwijntjes2000
And shouldn't we re-estimate story 5 as well?

It would be a good idea to do so.

Post by pcwijntjes2000
o Story 3 : 4 Days (can we say 1 remaining day?)

Suppose you do not, and keep the 4 day estimate. What might happen in
that second iteration ?

I think I would prefer to pretend - for planning purposes - that
Story 3 is not only not done, it's not even started. I like to
pretend that there are only two possible states for a story - 0% done
or 100% done.

Cheers,

-[Laurent]-
If a program is useless, it will have to be documented.

[pcwijntjes2000]

Laurent,

if this approach is used and the (improved) estimates are correct
story 3 will end within 1 day of the second iteration and +4+3 (for
stories 4 and 5). so stories 3,4 and 5 will be completed at the end
of the second iteration.

Assuming that we have more stories for iteration 3 , our new
velocity will then be 12! (Stories 3,4 and 5 complete within 1
iteration).

Still confused...

Post by Laurent Bossavit
Peter,

Post by pcwijntjes2000
After iteration 1 story 1 and 2 are finished, story 3 has

consumed 3

Post by Laurent Bossavit

Post by pcwijntjes2000
days.

You're not saying that story 3 is finished - I assume it's not ? Then
your velocity is 8 (3 from Story 1 + 5 from Story 2).

Post by pcwijntjes2000
Based on new insights we now think that story 3 and 4 each will cost a
total of 4 days of work.

Initially, your total scope was 3+5+2+2+3 = 15. Now, your

remaining

Post by Laurent Bossavit
scope is 4+4+3 = 11. You have at least one more iteration to go (11 >
8), with the project ending sometime during the third iteration. That
is your best estimate based on actual evidence of how long the
project will take.

Post by pcwijntjes2000
What does this mean for the second iteration?

Assuming stories are ordered by business value, you'll plan Story 3
and Story 4 (totaling 8) for the second iteration.

Post by pcwijntjes2000
And shouldn't we re-estimate story 5 as well?

It would be a good idea to do so.

Post by pcwijntjes2000
o Story 3 : 4 Days (can we say 1 remaining day?)

Suppose you do not, and keep the 4 day estimate. What might happen in
that second iteration ?
I think I would prefer to pretend - for planning purposes - that
Story 3 is not only not done, it's not even started. I like to
pretend that there are only two possible states for a story - 0% done
or 100% done.
Cheers,
-[Laurent]-
If a program is useless, it will have to be documented.

[Jeff Nielsen]

For the planning game/yesterday's weather to give you a smoother velocity
you need to have at least half a dozen stories per iteration. I tell my
teams that they should create stories small enough that they can fit 6-10
per iteration. Your example of only 2-3 stories per iteration illustrates
one of the problems with stories that are too large (a bumpy velocity
graph).

The granularity of the stories relative to iteration size seems to be an
important concept. You want to give the customer the most control possible
when you are on day n-2 of the iteration and it is obvious to all that you
won't finish all of the stories in the original iteration plan. If there
are several smaller stories that you do have time to finish instead (or if
you can break up the story that won't get done into smaller ones), this
gives you more options.

Jeff

--
Jeff Nielsen | Chief Scientist | Digital Focus
703.561.5854 | fax: [phone]

Post by pcwijntjes2000
Laurent,
if this approach is used and the (improved) estimates are correct
story 3 will end within 1 day of the second iteration and +4+3 (for
stories 4 and 5). so stories 3,4 and 5 will be completed at the end
of the second iteration.
Assuming that we have more stories for iteration 3 , our new
velocity will then be 12! (Stories 3,4 and 5 complete within 1
iteration).
Still confused...

Post by Laurent Bossavit
Peter,

Post by pcwijntjes2000
After iteration 1 story 1 and 2 are finished, story 3 has

consumed 3

Post by Laurent Bossavit

Post by pcwijntjes2000
days.

You're not saying that story 3 is finished - I assume it's not ?

Then

Post by Laurent Bossavit
your velocity is 8 (3 from Story 1 + 5 from Story 2).

Post by pcwijntjes2000
Based on new insights we now think that story 3 and 4 each will

cost a

Post by Laurent Bossavit

Post by pcwijntjes2000
total of 4 days of work.

Initially, your total scope was 3+5+2+2+3 = 15. Now, your

remaining

Post by Laurent Bossavit
scope is 4+4+3 = 11. You have at least one more iteration to go

(11 >

Post by Laurent Bossavit
8), with the project ending sometime during the third iteration.

That

Post by Laurent Bossavit
is your best estimate based on actual evidence of how long the
project will take.

Post by pcwijntjes2000
What does this mean for the second iteration?

Assuming stories are ordered by business value, you'll plan Story

3

Post by Laurent Bossavit
and Story 4 (totaling 8) for the second iteration.

Post by pcwijntjes2000
And shouldn't we re-estimate story 5 as well?

It would be a good idea to do so.

Post by pcwijntjes2000
o Story 3 : 4 Days (can we say 1 remaining day?)

Suppose you do not, and keep the 4 day estimate. What might happen

in

Post by Laurent Bossavit
that second iteration ?
I think I would prefer to pretend - for planning purposes - that
Story 3 is not only not done, it's not even started. I like to
pretend that there are only two possible states for a story - 0%

done

Post by Laurent Bossavit
or 100% done.
Cheers,
-[Laurent]-
If a program is useless, it will have to be documented.

Yahoo! Groups Links

[Laurent Bossavit]

Peter,

Assuming that we have more stories for iteration 3 , our new velocity
will then be 12! (Stories 3,4 and 5 complete within 1 iteration).

Yes.

Still confused...

What seems confusing ?

Cheers,

-[Laurent]-
Press Ctrl-Alt-Del to continue

[Kent Beck]

Many people still use points. I discuss the pluses and minuses in XPE2. I
prefer real time, but that's not a decree or commandment. The goal of
planning is to be able to say with confidence what you will be able to do.
You should adopt whatever mechanism allows you to do that.

In the situation you describe, story 3 was underestimated. Why did you not
meet your commitments in the first iteration? Was it because story 3 blew up
or because you went slower in general? It sounds like it was because story 3
blew up. In that case, it sounds like you can confidently estimate that you
can do 10 days worth of work in the next iteration.

Another factor I would include in this planning scenario is slack. If I
expected to be able to get 10 days worth of work done in iteration 1, I
might plan stories 1, 2, and 6 (something less important). Then iteration 1
would have been a success in delivering all the most important stories.
Packing iterations with only the most important stories guarantees that if
something goes wrong, something important will have to be dropped.

Does this help?

Kent Beck
Three Rivers Institute

-----Original Message-----
Sent: Tuesday, February 15, 2005 1:51 AM
Subject: [xpe2e] Planning game confusion
I have a question regarding the planning game. From Kents book I
understand that working with points is "not done" anymore, but I
keep getting confused on velocity vs estimates. I have a small
Estimates from planning game
Story 1: 3 days
Story 2: 5 days
Story 3: 2 days
Story 4: 2 days
Story 5: 3 days
Story 1,2 and 3 are put into the first iteration. We estimate our
velocity at 10 days per iteration
After iteration 1 story 1 and 2 are finished, story 3 has consumed 3
days.
Based on new insights we now think that story 3 and 4 each will cost
a total of 4 days of work.
What does this mean for the second iteration?
- What will be our velocity based on yesterdays wether?, my
o 8 days since we finished story 1 and 2
- What will be our new estimates for the remaining stories, my
o Story 3 : 4 Days (can we say 1 remaining day?)
o Story 4: 4 Days
o Story 5: 3 Days
Is this correct? It seems that when we use this approach the error
in estimating story 3 is used twice in the calculation proces (for
lowering the velocity and for an increase of the estimate). And
shouldn't we re-estimate story 5 as well?
anyone?

[Laurent Bossavit]

Kent,

It sounds like it was because story 3 blew up. In that case, it sounds
like you can confidently estimate that you can do 10 days worth of work
in the next iteration.

I'm the one who's confused now. In the first iteration, our velocity
was 8. You seem to be saying that in the case where a story "blew up"
in the previous iteration, "yesterday's weather" doesn't apply, and
we can plan more work for the next iteration than "yesterday's
weather" would suggest. Do I have that right ?

Cheers,

-[Laurent]-
All models are false; some models are useful.
G.E.P. Box

[Kent Beck]

Laurent,

The OP asked how much functionality he should commit to in iteration 2. I
described my answer (10 but with a slack story). I didn't follow Yesterday's
Weather, and I explained why. YW gives a different answer, but it doesn't
make sense to me in this situation.

Does that make sense?

Kent Beck
Three Rivers Institute

-----Original Message-----
Sent: Tuesday, February 15, 2005 2:59 PM
Subject: RE: [xpe2e] Planning game confusion
Kent,

It sounds like it was because story 3 blew up. In that

case, it sounds

like you can confidently estimate that you can do 10 days

worth of work

in the next iteration.

I'm the one who's confused now. In the first iteration, our velocity
was 8. You seem to be saying that in the case where a story "blew up"
in the previous iteration, "yesterday's weather" doesn't apply, and
we can plan more work for the next iteration than "yesterday's
weather" would suggest. Do I have that right ?
Cheers,
-[Laurent]-
All models are false; some models are useful.
G.E.P. Box

[Laurent Bossavit]

Kent,

Post by Kent Beck
Does that make sense?

I think it does. What I hear you say is that "the story blew up" is
going to be the exception, not the norm; so we have no reason to rely
on YW in this case. Is that correct ?

Cheers,

-[Laurent]-
All models are false; some models are useful.
G.E.P. Box

[Kent Beck]

Correct. Intelligent decision making trumps simplistic rules.

Kent Beck
Three Rivers Institute

-----Original Message-----
Sent: Friday, February 18, 2005 9:36 AM
Subject: RE: [xpe2e] Planning game confusion
Kent,

Post by Kent Beck
Does that make sense?

I think it does. What I hear you say is that "the story blew up" is
going to be the exception, not the norm; so we have no reason to rely
on YW in this case. Is that correct ?
Cheers,
-[Laurent]-
All models are false; some models are useful.
G.E.P. Box
Yahoo! Groups Links