Statistical Forecasting: Estimation made easy (ABE)

Instead of relying on vulnerable human memory we use the actual historical data and run statistical simulations to produce a forecast. Not only is the outcome significantly better than what expert guess provides but also it requires less work from development teams.

Transcript

1. ESTIMATION MADE EASY STATISTICAL FORECASTING
2. Paweł Brodziński @pawelbrodzinski Tomek Rusiłko @rusilko
4. EXPERT GUESS
5. IF A PROJECT HAS NO RISKS, DON'T DO IT. Tom DeMarco & Timothy Lister DON’T DO IT T. DEMARCO, T. LISTER: WALTZING WITH BEARS
6. ACCURACY IN ESTIMATING DID NOT IMPROVE AS INFORMATION ACCUMULATED, WHILE CONFIDENCE INCREASED CONSISTENTLY. Claire Tsai, Joshua Klayman, Reid Hastie ACCURACY OF ESTIMATION SOURCE: TSAI, KLAYMAN, HASTIE: EFFECTS OF AMOUNT OF INFORMATION ON JUDGMENT ACCURACY AND CONFIDENCE
7. SCIENTISTS AND WRITERS ARE NOTORIOUSLY PRONE TO UNDERESTIMATE THE TIME REQUIRED TO COMPLETE A PROJECT, EVEN WHEN THEY HAVE CONSIDERABLE EXPERIENCE OF PAST FAILURES TO LIVE UP TO PLANNED SCHEDULES. A SIMILAR BIAS HAS BEEN DOCUMENTED IN ENGINEERS' ESTIMATES. Daniel Kahneman, Amos Tversky ESTIMATION BIAS
8. STORY POINTS
12. LEAD TIME WORK TIME LEA WAIT TIME
13. WORK TIME LEAD TIME FLOW EFFICIENCY= LEAD TIME WORK TIME LEA WAIT TIME
14. LEA WAIT TIME LEA WAIT TIME
15. COUNTING THE NUMBER OF STORIES METRIC DOESN'T TAKE THE SIZE INTO ACCOUNT. IT TURNS OUT IT DOESN'T MATTER. THE SIZE OF STORIES IS GELLED TO A VERY COMMON SIZE. WE COULD USE THROUGHPUT VERY SUCCESSFULLY WITH THE RESEARCH. Larry Maccherone THROUGHPUT VS STORY POINTS
18. estimation.lunarlogic.io
19. EXPERT GUESS DOESN’T WORK
20. YOUR PLANS UNIVERSE’S PLANS
21. YOUR PLANS UNIVERSE’S PLANS source: DOGHOUSEDIARIES
22. ALTERNATIVE?
23. STATISTICAL FORECASTING BASED ON THE WORK OF TROY MAGENNIS, DOUGLAS HUBBARD AND LARRY MACCHERONE
24. PREDICTING THE FUTURE FROM HISTORICAL DATA
25. YOUR PLANS HISTORICAL DATA “REMEMBERS”
26. JAREK A FORECAST FOR
27. JAREK’S TASK HISTORY story start date end date #1 2016 / 02 / 29 2016 / 03 / 02 #2 2016 / 02 / 29 2016 / 03 / 03 #3 2016 / 02 / 29 2016 / 03 / 01 2016 / 03 / 03 2016 / 03 / 03 2016 / 03 / 04 2016 / 03 / 04 2016 / 03 / 05 2016 / 03 / 05 2016 / 03 / 06 2016 / 03 / 06 2016 / 03 / 07 2016 / 03 / 07 #10 2016 / 03 / 08 2016 / 03 / 08
28. Mon Tue Wed Thur Fri Week 1 Week 2 Week 3 Week 4 JAREK’S TASK HISTORY
29. LEAD TIME INPUT #1
30. WHY LEAD TIMES? Mon Tue Wed Thur Fri LEAD TIME 5 DAYS
31. WHY LEAD TIMES? Mon Tue Wed Thur Fri BUG FIX SICK LEAD TIME 5 DAYS WT
32. Mon Tue Wed Thur Fri 1 1 2 1 1 1 2 1 3 2 LEAD TIME
33. 1 1 2 2 3 2 LEAD TIME DISTRIBUTION Lead Time Count 1 Day 6 60% 2 Days 3 30% 3 Days 1 10% Total 10 1 1 1 1
34. LEAD TIME DISTRIBUTION 2 Days 3 Days 1 Day 1 1 2 2 1 1 2 2 3 1 16 Total Lead Time Days
35. LEAD TIME DISTRIBUTION FOR 10 TASKS IN THE BACKLOG 2 Days 3 Days 1 Day 1 1 2 2 1 1 2 2 3 1 16 Total Lead Time Days
36. LEAD TIME DISTRIBUTION FOR 10 TASKS IN THE BACKLOG 2 Days 3 Days 1 Day 1 1 2 2 1 1 2 2 3 1 16 Total Lead Time Days
37. Mon Tue Wed Thur Fri Week 1 Week 2 Week 3 Week 4 JAREK’S TASK HISTORY
38. Mon Tue Wed Thur Fri Week 1 Week 2 Week 3 Week 4 JAREK’S TASK HISTORY
39. Mon Tue Wed Thur Fri Week 1 Week 2 Week 3 Week 4 JAREK’S TASK HISTORY
40. Mon Tue Wed Thur Fri Week 1 Week 2 Week 3 Week 4 JAREK’S TASK HISTORY
41. WORK IN PROGRESS INPUT #2
42. Mon Tue Wed Thur Fri Week 1 Week 2 Week 3 Week 4 JAREK’S TASK HISTORY 0 0 0 0 0 2 0 1 1 1 1 1 1 1 1 1 1 1 1 1
43. WORK IN PROGRESS 1 Task 0 Tasks 2 Tasks SIMULATE THE NUMBER OF TASKS JAREK WORKS ON IN A DAY
44. 1 Task 0 Tasks 2 Tasks WORK IN PROGRESS SIMULATE THE NUMBER OF TASKS JAREK WORKS ON IN A DAY 1 0 0 DAY 1: DAY 2: DAY 3: …
45. 1 Task 0 Tasks 2 Tasks WORK IN PROGRESS SIMULATE THE NUMBER OF TASKS JAREK WORKS ON IN A DAY 1 0 0 DAY 1: DAY 2: DAY 3: …
46. WORK IN PROGRESS Each square represents a work day 0 Lead Time Days 0 Work Days Days
47. WORK IN PROGRESS Each square represents a work day 1 1 Lead Time Days 1 Work Days Days
48. WORK IN PROGRESS Each square represents a work day 1 0 1 Lead Time Days 2 Work Days Days
49. WORK IN PROGRESS Each square represents a work day 1 0 0 1 Lead Time Days 3 Work Days Days
50. WORK IN PROGRESS Each square represents a work day 1 0 0 3 Lead Time Days 4 Work Days Days 2
51. WORK IN PROGRESS Each square represents a work day 1 2 1 0 1 1 0 1 0 1 1 1 1 2 0 1 2 0 16 Lead Time Days 18 Work Days Days
52. DAYS 18
53. 18 19 12 20 19 19
54. 18 19 12 20 19 19
55. 18 19 12 20 19 19 12 21 22 21
57. 0 250 500 750 1000 Workdays 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
58. 0 250 500 750 1000 Workdays 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 50%
59. 0 250 500 750 1000 Workdays 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 50% 90%
60. GETPROJECTR.COM
61. A NEW PROJECT VERSUS AN ONGOING PROJECT
62. 90% CONFIDENCE LEVEL MEANS THAT OUR ESTIMATE IS CORRECT 9 TIMES OUT OF 10. KINDA. Pawel Brodzinski 90% CONFIDENCE INTERVAL?
63. UNCERTAINTY SOURCES OF UNCERTAINTY
64. S-curve
65. Project Alfa
66. Project Bravo
67. Project Bravo: Part 1
68. Project Bravo: Part 2
69. Project Charlie
70. Project Charlie: Part 1
71. Project Charlie: Part 2
72. ANY PROPOSED FORECASTING METHOD JUST HAS TO BE BETTER THAN WHAT YOU DO NOW, OR AT LEAST LESS EXPENSIVE WITH A SIMILAR RESULT. Troy Magennis BE BETTER
73. SOURCES & RESOURCES ▸ Troy Magennis original work: http://www.lkce13.com/videos/magennis/ ▸ http://focusedobjective.com/wp-content/uploads/2013/05/Modeling-and-Simulating-Software- Projects-Troy-Magennis.pdf ▸ http://blog.lunarlogic.io/2016/how-we-estimate-at-lunar-logic/ ▸ https://www.chicagobooth.edu/research/workshops/marketing/archive/workshoppapers/s06/tsai.pdf ▸ Planning Fallacy: https://books.google.pl/books?id=R- syxO7M67AC&pg=PA9&q=&redir_esc=y#v=onepage&q&f=false ▸ https://www.amazon.com/Thinking-Fast-Slow-Daniel-Kahneman/dp/0374533555 ▸ Flow efficiency: https://hakanforss.wordpress.com/2014/06/17/flow-thinking-aceconf/ ▸ http://zsoltfabok.com/blog/2013/12/flow-efficiency/ ▸ https://www.infoq.com/presentations/agile-quantify ▸ http://brodzinski.com/2015/02/story-points-velocity-the-good-bits.html ▸ https://estimation.lunarlogic.io/ ▸ https://www.agilealliance.org/estimation-and-forecasting/ ▸ Projectr: http://getprojectr.com/