You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I was thinking about adding support for an OWN_TEAMOrientation which in the case for an EventDataset would behave the same as the ACTION_EXECUTING_TEAMOrientation, but in the case of a TrackingDataset it would orient the coordinates of each player with respect to its own side. Implying that e.g. both goalkeepers would have a starting position of +- (0, 50).
That sounds like a valuable feature! However, I would categorize it more as a symmetry transformation rather than an orientation change. In my conceptualization, orientation transforms merely switch perspectives without altering the underlying data, similar to changing the camera angle. On the other hand, symmetry transforms modify the relationships between data points.
So maybe we should create a different kind of transformation for this. There would be two such symmetry transformations: each team plays towards the same goal, and each action starts on the right side of the pitch. Combining both could be particularly useful, for instance, when analyzing corners it is common to aggregate them at a common point, like (0,0).
I was thinking about adding support for an
OWN_TEAM
Orientation
which in the case for anEventDataset
would behave the same as theACTION_EXECUTING_TEAM
Orientation
, but in the case of aTrackingDataset
it would orient the coordinates of each player with respect to its own side. Implying that e.g. both goalkeepers would have a starting position of +- (0, 50).What do you think @koenvo @JanVanHaaren ?
The text was updated successfully, but these errors were encountered: