I am trying to understand the calculations behind the gate rank so that to get a feel of how our actual marks,ranks and gate score is mapped.
Scorecard mentions the formulae used to be
Gate Score = Sq + (St - Sq) *{ ( M - Mq )/ ( Mt - Mq ) }
where
M = marks of candidate
Mq = qualifying marks (25 or (μ+σ), where μ,σ are mean and standard deviation respv.,whichever larger)
Sq = Gate Score(GS) corres. to qualifying marks (Mq) in our exam mentioned on scorecard as 350
Mt = mean of marks of 0.1% i.e mean of top 107 people (as 1,07,893 students appeared) or mean of top 10 (larger of these)
St = GS corres. to Mt and for 2018 exam it is 900
i am using M = 58.33, St = 900, Sq = 350 and GS = 747
now to estimate Mt by using the rank prediction app (accuracy is 90-95%) we see that top 10 ranks score b/w 76-83 and mean around 79 which is evidently more than mean of 0.1%(top 107) which would be somewhere around 70 so Mt = 79
if we let Mq = 25 and try to estimate Mt we get around 71 so why this discrepency ?
and if we let Mq be (μ+σ), and solve for Mq using Mt=79 we get Mq=10 which is contradiction as per definition of Mq
Can someone help me resolve this ?