Bloated spreads

Just curious. I am working on a model which I am using Kenpom ratings. I am incorporating adj off & Def , tempo & the 4 factors. Some of the games i do relate pretty well with what the vegas odds would be for that game & some are just way off. For example Last NIght my numbers said TCU -5.96 & the spread at the time I did that was -5 but in contrast I did the Uconn / Marquette game & my numbers showed UCONN -17 way off from what the spread was. IM curious if this happens to other people & If there maybe something else I can add to this model to help. Any suggestions. 

What are the "4 factors" that you are capping?

@RebelTell2

@GTKiller

Efg %

To%

ORb %

Ft / fgM

I think by adjusting the % of those factors I have found a issue. 

Here are the 5 basics worth looking at:

True shooting % (FG/FT composite)

Rebound Rate (rebound composite)

Steals per possession % (You didn't list this)

Turnovers per possession %

Defensive Efficiency (Defensive composite)

These 5 categories will give you a snapshot of how the match-up shakes out statistically. Always keep in mind though: statistics can deceive big time. Part of capping is knowing when the numbers are conning you.

Hope this helps,

RT2

@RebelTell2

 Helps alot. I appreciate your input. Ill go back to the drawing board here.