Develop better algorithm for choosing Sprint projects
Posted: Sat Aug 19, 2023 4:44 pm
The fiasco of this latest Sprint contest with Asteroids points out that projects that do not have enough WU's to last through the contest length shouldn't be put into the selection pool.
All it takes is to look at the peridicity of work availability using the Grafana graphing tool to see that the Asteroids project was at the bottom of its work availability cycle and it was going to run out of work after the first day of the contest.
https://grafana.kiska.pw/d/boinc/boinc? ... -6M&to=now
I think that work availability for the duration of the contest should be a selection criteria. Or just run the Sprint for a shorter length of time to stay within the work availability limit of the selected project.
The arbitrary use of existing real international competitions to sync with the Sprint is a detriment since you can't start a Sprint at the logical time where work will be available for the duration of the Sprint.
All it takes is to look at the peridicity of work availability using the Grafana graphing tool to see that the Asteroids project was at the bottom of its work availability cycle and it was going to run out of work after the first day of the contest.
https://grafana.kiska.pw/d/boinc/boinc? ... -6M&to=now
I think that work availability for the duration of the contest should be a selection criteria. Or just run the Sprint for a shorter length of time to stay within the work availability limit of the selected project.
The arbitrary use of existing real international competitions to sync with the Sprint is a detriment since you can't start a Sprint at the logical time where work will be available for the duration of the Sprint.