Assessment of tasks
In assessing the complexity of tasks it is often difficult to give an absolute assessment. It is much easier to estimate the relative size of two tasks, that is, for example, to say that task A is twice as large as task B.
With the agile development process, you have to evaluate a lot of user stories.
Using pairwise comparison, the error in determining the estimates is reduced, and moreover, this error can be calculated.
Theory
Suppose we need to evaluate
n stories. Denote the size of each story by
s i .
To find the size of each story, we estimate the size of each story relative to all other stories and write the result
a ij to the matrix A
nxn :
I note that only the upper part of the matrix is to be filled, since
a ij = 1 / a ji .
Calculation of the size of stories
To calculate the average relative size of the history
mrs i, it is necessary to find the geometric average for the
i-th row of the matrix, i.e.
')

(one)
The average relative size of the story tells us little, so the size of the story, in the units we need, is calculated based on the story, the size of which is known to us by
reference :

, (2)
where
s right is the size of the reference history in the units we need (story points, hours, etc.),
mrs right is the relative size of the reference history calculated by formula (1).
Numeric example
Imagine that we have four user stories for evaluation, which we briefly describe as follows:
- Registration / Authorization
- Add post
- Show list of posts
- Post classification
Fill the matrix with relative estimates:
| one | 2 | 3 | four | Relative size mrs i | Grade s i |
---|
1. Registration / Authorization * | one | 0.33 | 2.00 | 3.00 | 1.19 | ten |
---|
2. Add post | 3.00 | one | 3.00 | 5.00 | 2.59 | 21.83 |
---|
3. Show list of posts | 0.50 | 0.33 | one | 0.50 | 0.54 | 4.52 |
---|
4. Post classification | 0.33 | 0.20 | 2.00 | one | 0.60 | 5.08 |
---|
We took history No. 1 as a reference value and estimated it at 10 story points. The size of the stories number 2, 3, 4 was calculated by the formula (2).
Calculation of the index of convergence and confidence interval
The variance of estimates in the matrix A
nxn is calculated by the formula [1]:
The square root of the variance is called the convergence index:
The lower the convergence index, the more consistent our estimates.
However, the convergence index is difficult to interpret. It is better if you translate it into a confidence interval for estimates.
To do this, we make the assumption that each story in our assessment makes the same contribution to the total variance, then the error and confidence variant is calculated using the following formulas:
Implementation
I tried to implement this functionality in Javascript. Put here
estimateit.ru .
Future plans
An obvious disadvantage of using this method is that with an increase in the number of evaluated stories, the number of necessary comparisons grows. This can be avoided using not all comparisons, but only the minimum necessary ones.
There is also a methodology for finding inconsistent estimates in matrix A.
Literature:
[1]
Sizing User Stories Using Paired Comparisons // Eduardo Miranda, Pierre Bourque, Alain Abran