Reduce exam pass cutoff
Opened this issue · 2 comments
lukasgabriel commented
Would it make sense to reduce the cutoff for passing to 0.5, because at this point, the chance of passing is theoretically higher than failing?
nk-coding commented
Well currently the chance of failing is ≈ 80%, which is reasonable for theoretical computer science
How about a second endpoint? Maybe a programming exam with a 90% chance of passing and a theoretical exam with a 20% chance of passing?
Seems to be more realistic
lukasgabriel commented
Ah, you're right. I assumed the chance of passing to be a perfect 0.5
We could gather some data about real-world exam grade distributions from open sources to model our predictions more closely to reality.