Mind linking versions of DB / trinity / QT and GSR M3th0d ?
that is a shit load of keys lol
I think the low counts on A4 are an artifact of the way i'm checking against median. For example:
[Keys] Counts:
Act 1 => 43
Act 2 => 42
Act 3 => 42
Act 4 => 31
[Keys] Stats:
LF => 20.3
LQ => 33.8
M => 42
UQ => 42.8
UF => 56.3
IQR => 9
this would ignore A1 and do all 3 acts since they're not above median (42). So the lowest count doesn't get a chance to catch up.
[Keys] Counts:
Act 1 => 19
Act 2 => 15
Act 3 => 7
Act 4 => 26
[Keys] Stats:
LF => -13,9
LQ => 9
M => 17
UQ => 24,3
UF => 47,1
IQR => 15,3
[Keys] Counts:
Act 1 => 30
Act 2 => 23
Act 3 => 23
Act 4 => 31
[Keys] Stats:
LF => 11,4
LQ => 23
M => 26,5
UQ => 30,8
UF => 42,4
IQR => 7,8
Familiar problem here, Act 3 getting ignored for a total of 10 keys difference to other acts.
Currently testing HighestKeyCountId(keyActorID) from QuestTools\Helpers\CustomConditions.cs instead of KeyAboveMedian(keyActorID) as condition on BKF_data.xml lines 86,90,94 and 98. Not perfect workaround, but it's no longer skipping that act I was missing and hey it's 01:00am here ...
Still interested in learning what's "wrong" with that statistical approach on act skipping conditions or whatever is the problem.
[Edit]
Seems to work ok for me for the time being. Started out with
Code:[Keys] Counts: Act 1 => 19 Act 2 => 15 Act 3 => 7 Act 4 => 26 [Keys] Stats: LF => -13,9 LQ => 9 M => 17 UQ => 24,3 UF => 47,1 IQR => 15,3
and now it's at
Code:[Keys] Counts: Act 1 => 30 Act 2 => 23 Act 3 => 23 Act 4 => 31 [Keys] Stats: LF => 11,4 LQ => 23 M => 26,5 UQ => 30,8 UF => 42,4 IQR => 7,8