Partypooper.....I'm enjoying having a rant about longheld views...mates not interested in my theories and usually buy me a beer to shut me up...that works.
Your idea on faves would stand up if the price in every race was 2/1 the favourite.
What about races where the favourite is Evens..that would imply that the fave wins one in every two races...or 4/6 would mean the fave wins 3 of every five races?
If you go down the 'one in three faves wins a race' road then you are taking unders every time you back a horse at less than 2/1...the logical extension is that the higher the price of the favourite the better the value you are getting...this flies in the face of our herd instinct that says the more people that predict a certain outcome then the more likely it is to occur.
Winning fave Stats tell us that 2 out of 3 times the herd is wrong.
Might be a system in long priced favourites?

Thought of a system once where you(on the TAB) back the Friday paper poll topper for the last race Saturday.
Friday's picks are done on Thursday in a rational calculated way...unless the cadet/coffee boy is busy then he picks as quick as he can...in that case may pay to add up your favoured tipsters who have a reputation to consider.
The last race Saturday sees a lot of drunk money hitting the TAB...full of dutch courage playing up their winnings and going for the big collect...mugs chasing money...and all-ups distorting the pool.
You see blokes late every Saturday lurch their way to the terminal with a glance at the screen for the tote fave and on goes their money...the cumulative effect of the drunk and desparate money sees the fave firm and in come the fluctuation watchers adding to the momentum as the 'fave's' price tumbles.
A good time to get a bit of 'value' about the journo's Thursday pick.
Be interesting to see some work on a system using those rules.