The SequenceStateBuilder needs to be improved

[probberechts]

I think the logic for building sequences could be improved. Below are a few issues that I noticed:

• Typically, an event stream starts with a few off-ball events such as an event that announces the lineups, the start of a period, etc. These events should not be part of the possession sequence of a team.
• Events in which the goal keeper picks up or claims the ball should be able to start a new sequence
• Take ons should be able to start a new sequence
• Successful interceptions should be able to start a new sequence
• A single failed pass should not start a new sequence
• When a sequence ends with a duel between two players, the lost duel should be part of the old sequence and the winning duel should be part of the new sequence
• A shot should not always end a sequence. For example, a rebound should be part of the same sequence.

[probberechts]

While looking up some definitions, I figured Opta uses both the concept of a "sequence" and a "possession".

events are organised into sequences and possessions.
Sequences are defined as passages of play which belong to one team and are ended by defensive actions, stoppages in play or a shot.
Possessions are defined as one or more sequences in a row belonging to the same team. A series of passes leading to a shot which is saved and results in a corner kick would comprise one possession since the same team retains control, but more than one sequence, since the ball has gone out of play. A possession is ended by the opposition gaining control of the ball.
-- https://www.statsperform.com/resource/introducing-a-possessions-framework

Hence, maybe we should keep the SequenceStateBuilder as is and add a PossessionStateBuilder?

[DriesDeprest]

Yes, I think it would be wise to make the distinction and use the definitions used by Opta.

[koenvo]

Would be great. Not sure how difficult this will make things, but maybe good to be able to pass some configuration