innolution agile estimating rubin

8/8/2019 Innolution Agile Estimating Rubin

1/25

Copyright (c) 2006-2007, Innolution, LLC and Mountain Goat Software, Inc. All Rights R

Agile Estimatingby Kenny Rubin

Includes slides from:


2/25


3/25

Copyright (c) 2006-2007, Innolution, LLC and Mountain Goat Software, Inc. All Rights Reserved.

2

[email protected]

(303) 827-3333 (office)

(303) 589-5938 (mobile)

InstructorKenny Rubin

Managing Director


3

Which Are We Talking About?

4Automate tests

5Write tests

12Code middle tier

6Write test fixture

8Code the UI

2As a frequent flyer,

I want to


I want to


I want to


I want to


I want to

Sprint2

Sprint1

Product Backlog Sprint Backlog

Weretalkingabouttheserightnow


4/25


4

Estimates What You Need toKnow

Estimates should be created by the team

Not by the most senior developer

You need estimates before you canprioritize

Estimates are not commitments

Every estimates comes with a probability

Estimates should be of relative size,rather than absolute size


5

Estimate Size; Derive Duration

SizeSize CalculationCalculation DurationDuration

300 Points300 PointsVelocityVelocity

==

2020

300/20 =300/20 =

1515

SprintsSprints


5/25


6

Traditional and agile measure sizedifferently

Traditionalmeasuresof size

Measures of Size

Lines of CodeFunction Points

Agilemeasuresof size

Story pointsIdeal days


7

StoryPointsStoryPoints


6/25


8

Story Points

The bigness of a task

Influenced by

How hard it is How much of it there is

Relative values are whatis important:

A login screen is a 2 A search feature is an 8

Points are unit-less

As a user, I want to beable to have some but notall items in my cart giftwrapped.

As a user, I want to beable to have some but notall items in my cart giftwrapped.

5


9

Where do we get the points?

Mostly we make them up!

Suppose were sailing

we see this island through our telescope

How far away is it?


7/25


10

How far away is the island?

Theres nothing else in sight so theresnothing to compare it to

Hard to guess then how many milesaway it might be

So, lets just say the distance is 12


11

Lets add more islands

0 12 24

Distances to otherislands are

estimated relativeto the first island

Distances to otherislands are

estimated relativeto the first island


8/25


12

What do we know and not know?

We know...

That three islands are about the same distance (12)away

That another island is about twice as far away

We dont know...

how long it will take us to sail a distance of 12

But we know...

That we can sail to all the islands 12 away in aboutthe same time

That it will take twice as long to sail to the islandthat is 24 away


13

Estimating the first voyage

0 12

The captain asks us howlong it will take to get to thefirst island

Its 12 away We guess we sail at 2 per

hour so we say 6 hours

The captain asks us howlong it will take to get to thefirst island

Its 12 away We guess we sail at 2 per

hour so we say 6 hours


9/25


14

We sail to the first island

0 12

Sailing to the first islandtakes 6 hours.

We were right!

We now know that we sail2 units per hour, or 12units in 6 hours.

Sailing to the first islandtakes 6 hours.

We were right!

We now know that we sail2 units per hour, or 12units in 6 hours.


15

How long will it take?

0 12

6hou

rs

We return to port. How long will it take to sail

to the second island? The third? The far island?

We return to port. How long will it take to sail

to the second island? The third? The far island?


10/25


11/25


12/25


20

Animal P ictures

Bear Giraffe Gorilla Hippopotamus

Kangaroo Lion Tiger Rhinoceros


21

IdealTimeIdealTime


13/25


22

Ideal Time

How long something would take if:

Its all you worked on

You had no interruptions And everything you need is available

The ideal time of an American football gameis 60 minutes

Four 15-minute quarters

The elapsed time is much longer (3+ hours?)


23

Monday has 8 hours Each week has 40 hours

Ideally

Elapsed Time vs. Ideal Time

So,thisdeveloperwill

onlymakefourhoursof

progressonMonday.

Itwilltaketwocalendar

daystocompleteone

idealdayofwork.

Howlongwillthistake?Areyou

answeringwhatisbeingasked?

Monday has: 1 hour of meetings 3 hours of email 4 hours left for the project

But instead...


14/25


24

Ideal Time vs. Elapsed Time

Its easier to estimate in ideal time

Its too hard to estimate directly in

elapsed time Need to consider all the factors that affect

elapsed time at the same time youreestimating


25

Specialization

First, dont worry about it too much

Were usually better off with fairly rapid,imprecise estimates than spending moretime

Second

Just add up the components and reportone total estimate of ideal days

= 8

= 8

= 8


15/25


26

The Great Debate

Story points or ideal days?Story points or ideal days?

Which do you prefer:


27

Comparing the ApproachesInitial Ease of Explanation

Easy Hard

Initial Ease of Estimating

Easy Hard

Confront Time Wasting Activities

Easy Hard

Risk of Misinterpretation

Low High

True Measure of Size

True False

Estimates Decay

More Likely Less Likely

Could Imply False Level of Precision

More Likely Less Likely

Ideal Days

Story Points


16/25


17/25


30

Triangulation

Confirm estimates by comparing the storyto multiple other stories.

Group like-sized stories on table orwhiteboard

Story AStory A

3pts

5pts

8pts

Story CStory C Story DStory D Story FStory F

Story BStory B Story EStory E


31

Disaggregation

Breaking a big story into smaller stories

You know how long the smaller tasks take So, disaggregating to something you know lets

you estimate something bigger you dont know

Sometimes very useful

But disaggregating too far causesproblems

Forgotten pieces Summing lots of small errors can be big

number


18/25


32

How Much Effort?

A little efforts helps a lot

A lot of effort only helps a little more

Effort

Accuracy


33

Use the Right Units

Can you distinguish a 1-point storyfrom a 2?

Can you distinguish a 17 from an 18?

Use units that make sense, such as

1, 2, 3, 5, 8, 13 1, 2, 4, 8

Stay mostly in a 1-10 rangeInclude0

and ifyou

want


19/25


34

time

1

2

time

Remember Estimates are Binned


35

Planning Poker Overview

An iterative approach to estimating

Steps

Each estimator is given a deck of cards,each card has a valid estimate written on it

Customer/Product owner reads a story andits discussed briefly

Each estimator selects a card thats his orher estimate

Cards are turned over so all can see them Discuss differences (especially outliers) Re-estimate until estimates converge


20/25


36

Planning Poker - An Example

Estimator Round 1 Round 2

Susan 3 5

Vadim 8 5

Ann 2 5Chris 5 8


37

Why P lanning Poker Works

1 Jrgensen, Magne. 2004. A Review of Studies on Expert Estimation of Software DevelopmentEffort.

2 Hagafors, R., and B. Brehmer. 1983. Does Having to Justify Ones Decisions Change theNature of the Decision Process?

3 Brenner, et al. 1996. On the Evaluation of One-sided Evidence.4 Miranda, Eduardo. 2001. Improving Subjective Estimates Using Paired Comparisons.5 Saaty, Thomas. 1996. Multicriteria Decision Making: The Analytic Hierarchy Process.

Those who do the work, estimate thework1

Estimators are required to justifyestimates2, 3

Focuses most estimates within anapproximate one order of magnitude4,5

Those who do the work, estimate thework1

Estimators are required to justifyestimates2, 3

Focuses most estimates within anapproximate one order of magnitude4,5


21/25


38

Why P lanning Poker Works

Combining of individual estimates6 Through group discussion7

Leads to better estimates Emphasizes relative rather than absoluteestimating

Estimates are constrained to a set of values so wedont waste time in meaningless arguments

Everyones opinion is heard Its quick and fun

Combining of individual estimates6 Through group discussion7

Leads to better estimates Emphasizes relative rather than absoluteestimating

Estimates are constrained to a set of values so wedont waste time in meaningless arguments

Everyones opinion is heard Its quick and fun

6 Hoest, Martin, and Claes Wohlin. 1998. An Experimental Study of IndividualSubjective Effort Estimations and Combinations of the Estimates.

7 Jrgensen, Magne, and Kjetil Molkken. 2002. Combination of SoftwareDevelopment Effort Prediction Intervals: Why, When and How?


39

www.planningpoker.com

Free

to use

Free

to use


22/25


40

Re-estimatingRe-estimating


41

Re-estimating

We expectvelocity = 12

We expectvelocity = 12

But we finishhalf that much

But we finishhalf that much

33

3333

33

3333

3333

3333

3333

3333

3333


23/25


42

But a velocity of 6doesnt seem right;so we double those

stories

But a velocity of 6doesnt seem right;so we double those

stories

But if all stories arethe same size, we

need to double all ofthem

But if all stories arethe same size, we

need to double all ofthem

6666

3333

3333

3333

6666

66666

666

6666


43

Six Sw imming Stories

5As a coach, I can have the system recommend whoshould swim in each event subject to restrictions abouthow many events a swimmer can participate in.

F

1As a coach, I can upload meet results from a file exportedfrom the timing system used at the meet.

E

3As a coach, I can see a text report showing eachswimmers best time in each event.D

3As a swimmer, I can see a pie chart showing how manyfirst, second, third, and lower places Ive finished in.

C

5As a coach, I can see a line chart showing the progressover the season of all swimmers in a particular event.

B

3As a swimmer, I can see a line chart of my times for aparticular event.

A


24/25


25/25


46

A Third Scenario

Re-estimating when relative size changes 33

Whatcanyou

concludeabout

re-estimating

fromthesethreescenarios?

Stories A, B, and C are all graphing stories; graphing istwice as hard as thought so double all these

What was the velocity of the first Sprint? Can the team do that much in the second Sprint?

Summary

Estimates should be created by the team

Not by the most senior developer

You need estimates before you canprioritize

Estimates are not commitments Every estimate comes with a probability

Estimates should be of relative size,rather than absolute size

innolution agile estimating rubin

Documents