Old values:
With 3 Trinispheres, being able to cast it is = 19%
With 3 Trinispheres, not being able to cast it is = 12%
With 4 Trinispheres, being able to cast it is = 24%
With 4 Trinispheres, not being able to cast it is = 15%
New values:
With 3 Trinispheres, being able to cast it is = 21% (21998/100000)
With 3 Trinispheres, not being able to cast it is = 9% (9394/100000)
With 4 Trinispheres, being able to cast it is = 28% (28212/100000)
With 4 Trinispheres, not being able to cast it is = 11% (11763/100000)
So if you have 3 Trinispheres, you get +2% more, if you have 4, you get +4% more chance (adding 4 of either City of Traitors or Ancient Tomb).
It really becomes strange now, because adding 4 of the 2-mana lands can be difficult in all but MUD. You already have 4 Workshops and 5 Stripmines, so adding 4 more would result in 13 colorless lands. This can't leave too much space for the color mana though.
I don't think it really matters that much, because it doesn't have a strong bearing on the validity of your point. The major problem is not having the mana to cast Trinisphere turn 1, but rather having the card in your hand.
Well, the odds of having the Trinisphere in your hand can easily be calculated (add both). Something does puzzle me (the chance of getting a Trinisphere opening hand is 39%, as opposed to the 42% i would expect by calculating it using stats. How odd...
I wonder if anyone (other than njx from SCG.com) has tested the Trinisphere out?