Magic Estimation is a well-known fast-paced estimation technique for Agile teams, that was developed for in-room collaboration.

In our app, we have re-created the framework in the digital environment of Jira Cloud.

In this article, we will focus on the concepts of the original game and their implementation in the digital version.

Classic Magic Estimation game

The classic in-office Magic Estimation process contains 4 main steps:

1. Game preparation:

   • a moderator prepares a board with lanes for team’s standard estimation values, usually Fibonacci Sequence numbers or T-Shirt sizes;

   • the moderator prints PBIs

2. First estimation step:

   • the development team selects the first PBI as a baseline, usually, it’s the smallest 1 Story Point PBI;

   • the team spreads the rest of the prepared PBIs through the board lanes, this process must be done in silence, no discussion is allowed;

   • the Product Owner clarifies PBI-related questions;

3. Second estimation step:

   • the team silently checks the results, each participant can move any PBI to a different lane in case of any concerns;

   • the moderator tracks lanes changes for every PBI;

   • PBIs that have been changed a lot are discussed centrally until there is a consensus on the number of Story Points;

   • the exercise is finished when all the PBI’s are placed and everyone agrees with the final result.

4. Processing results:

   • the moderator has to transfer all results to an issue tracking system.

Read the following articles for more details:
https://agiletransparency.com/estimation-a-kind-of-magic/
https://medium.com/the-liberators/magic-estimation-5165df2be245

Magic Estimation game flow for Jira Cloud

In Magic Estimations, steps 1 and 4 are organically incorporated in the app’s flow, making the life of the moderator much easier.

Magic Estimation for Fixed Lanes board mode

The app’s Fixed Lanes board mode is considered the default for the Magic Estimation game because it supports lane values configured from the very beginning of the session.

Although our app is flexible and doesn’t limit users with one particular flow, we recommended the following flow of using Magic Estimation in the app:

1. The moderator creates, configures, and opens the session.

2. The moderator invites participants using the sharing option from the bottom Participants area:

   

3. The board is prepared by the team for the estimation:

   • a baseline PBI is selected and moved to the corresponding lane;

   • alternatively, Magic Estimations users add already estimated PBIs to the lanes for reference; this can be done manually or using the “Distribute estimated issues” option from the session controls of the moderator

     

4. When everyone is ready, the moderator starts Magic Estimation game:

   • the participants are automatically formed in sequence;

   • each participant gets 15 seconds (the default timer value that might be changed in session configurations) for their estimates;

   • during each turn participants can make 1-5 estimates depending on the session configuration;

   • PBIs can be moved from the backlog to the board or from one lane to another;

   • PBIs that are moved between the lanes several times must be spotted by the moderator;

   • the moderator can pause the game at any time, for example, to discuss the PBI that was replaced several times

     

5. The game ends when all the participants are fine with the board results. The moderator can pause the game and save the estimates.

Magic Estimation for Relative board mode

In the Relative board mode, where participants estimate without Story Point value in mind and thinking of the relative weight of a particular story, Magic Estimation process looks similar. There are 2 main differences with the Magic Estimation for Fixed Lanes board mode:

1. Estimation values are added in the end, so reference issue should also be added right before saving estimates.

2. The final extra step is adding estimation values to the lanes on the board.

^^top of topic