2. First developed by James Grenning
“How to avoid analysis paralysis while release
planning”
The aim of Planning Poker is to create
estimates in a short time and involve the
whole team
David Parsons - Massey University
3. Like the Planning Game, Planning Poker is not
really a game
◦ Simply a way of using game-like activities to
perform some of the tasks of agile planning
One significant difference is that in Planning
Poker there are additional „pieces‟ – the
„cards‟ used to estimate stories
David Parsons - Massey University
4. The customer reads a story
◦ There is a discussion clarifying the story as
necessary
Each programmer selects their chosen
estimate card
◦ (Or writes their estimate on a note card, if no pre-
printed pack is available)
No discussion of estimates takes place at this
stage
Once all programmers have written their
estimate, all the cards are turned over
David Parsons - Massey University
5. If there is agreement, no discussion is
necessary
◦ The estimate is recorded and we move on to the
next story.
If there is disagreement in the estimates, the
team can try to get a consensus
If there is no consensus, it doesn‟t matter
◦ It is only one story out of many
It can be deferred, split, or the lowest
estimate can be taken
David Parsons - Massey University
6. Everyone in the team participates
◦ They have to make an estimate
◦ Everyone gains experience
Discussions are automatically triggered by
the more problematic estimates
Where estimates are straightforward, the
game enables consensus without unnecessary
discussion
David Parsons - Massey University
7. Save time of manually writing estimates
Cards also only have a subset of possible
estimated days
James Grenning‟s set:
◦ 1, 2, 3, 5, 7, 10 days and infinity
As the estimates get longer, the precision
goes down
David Parsons - Massey University
8. Maximum story size is under 2 weeks
if you estimate that a story is longer than 2
weeks, play the infinity card and make the
customer split the story
David Parsons - Massey University
9. Mountain Goat Software
◦ 0, 1, 2, 3, 5, 8, 13, 20, 40, and 100
◦ online version also includes a .5 card
◦ The „zero‟ value might look odd but it does not
mean it takes no time at all, rather that is closer to
0 than 1
Mike Cohn
◦ 1, 2, 3, 5, and 8 (Fibonacci sequence)
◦ or 1, 2, 4, and 8
StudioAlt
◦ ?, 0, ½, 1, 2, 3, 5, 8, 13, 20, 40, 100
David Parsons - Massey University
10. If the number represents days, why do some
card sets go up to 100?
Because not everyone sticks to „days‟ as their
unit of estimation
“Planning Poker can be used with story
points, ideal days, or any other estimating
unit”
– Mountain Goat Software
David Parsons - Massey University
11. As well as the estimation number cards, some
packs have additional cards
◦ „don‟t know‟
◦ „discuss‟
◦ „coffee time‟
◦ etc.
You can make up cards that you find useful in
your own processes
David Parsons - Massey University
12. 0, 1, 2, 3, 5, 8, 13, 20, 40 (in 5 „suits‟)
+ „fast forward/rewind‟ and „talk‟
David Parsons - Massey University
13. One suggestion for maintaining the speed of
the process is to use a 2-minute egg timer
for each discussion
This may be turned over once more for more
problematic estimates but then the next story
should be estimated
David Parsons - Massey University
14. With large teams, where there are many
stories to estimate, Planning Poker can be
played separately by smaller teams
However they will need to have done some
estimating as a whole team first, covering 10
to 20 stories
◦ This ensures that everyone is familiar with the
technique
◦ Also ensures that subsequent estimates are
consistent between groups
David Parsons - Massey University
15. A minor variation on Planning Poker is to use
poker chips instead of estimation cards, 1
chip for each story point
Possible to use different coloured chips to
indicate different estimation contexts
◦ “we had three team sizes we were considering for
the release and we used white, blue and red chips
to indicate the base story points and two levels of
increment”
Yip, J. (2007)
David Parsons - Massey University
16. Another variation is to use an on-line version
for distributed teams
You can also download versions for mobile
phones
planningpoker.com
David Parsons - Massey University
17. Moløkken-Østvold and Haugen (2007)
identified some measurable and potential
benefits
Haugen (2006) claimed that it improved
estimation in most cases, but that it
increased estimation error in the extreme
cases
David Parsons - Massey University
18. Cohn, M. (2005). Agile Estimating and Planning, Addison-Wesley
Grenning, J. (2002). Planning Poker or How to avoid analysis
paralysis while release planning https://sewiki.iai.uni-
bonn.de/_media/teaching/labs/xp/2005a/doc.planningpoker-
v1.pdf
Haugen, N. (2006). An Empirical Study of Using Planning Poker
for User Story Estimation, AGILE 2006, 23-34
Moløkken-Østvold, K. & Haugen, N. (2007). Combining Estimates
with Planning Poker – An Empirical Study, 18th Australian
Software Engineering Conference (ASWEC 2007), 349–358
Yip, J. (2007). Hands-on release planning with poker chips. 14th
Conference on Pattern Languages of Programs (PLOP 2007)
David Parsons - Massey University