I prefer FR on PokerStars and tourneys on FT. Not sure why that is, it's just always worked out for me that way. So I keep a FR
bankroll on Stars and a tourney roll on FT. I prefer the Stars software client, although with the last update FT has almost caught up. I also have accounts on SportsBook (Cake) and UB, but haven't really played them enough to know what works better for me there -- I don't really care for the software on either though.
As a software engineer myself who has done quite a bit of RNG work, I can attest that the RNGs used by these major sites are not going to result in one flopping more paint, or whatever, than another. That's just another beginner's perception and fuels the "rigged" conspiracies that so many players have -- typically new, mediocre and/or downright bad players. You never hear the solid players whining about RNG's and sites being rigged. They understand that the variance is magnified when playing many many times more hands online than they would ever see live.
Theoretically, science tells us there is no such thing as "truly random" machine-generated numbers, and some beginning players who have heard this try and use it as the basis of their rigged arguments. Practically, however, even the simplest RNG built into the core runtime library of most programming languages is sufficiently random to pass any test of human perception of "randomness," particularly when compared to a human manually shuffling a deck of 52 cards. And to my knowledge, the sites don't depend strictly on a canned RNG built into a library, they develop their own variations and extensions to make it even more "random." For example, one key factor in the distribution of random numbers is the value used to seed the RNG. This starting value needs to be as random as possible -- if you seed an RNG with an identical value, you will generally get identical results. So most sites use a mix of "noise" data gathered from clients (date/time down to the millisecond, mouse movements captured by the client, etc.) to constantly shuffle and reseed their RNG. This type of data is virtually never repeatable and therefore never going to result in the same seed values, which makes their RNGs non-deterministic and non-repeatable.
Sorry for channeling my inner geek, I just get frustrated by the frequent suggestions and false information about how sites are rigged or that their RNGs are being compromised somehow.
One thing that is different between sites, however, (besides the varying quality of players) is how/when they shuffle the deck. Some, like Stars, use a set deck that is only shuffled between hands. Others, like FT, reshuffle the remaining deck between every deal of the cards. To me, this makes no difference as far as random distribution if you believe the deck was truly randomized to begin with, you're just taking an already randomized sequence and re-randomizing it. But some will still try and argue one way or the other that it makes a difference. The only difference it makes is when you're trying to second-guess your fold because you see that the board would have hit you if you stayed in the pot. With a reshuffled deck, you would have seen different cards had you stayed in so you can never really know what would have happened.