Hi Stephen
Thanks for the figures. I'll include them in my mini-analysis. What do they mean by 'untitled'? I assume it means not an NM, GM etc. Which seems to assume that anybody who is good enough would automatically have gone that route. I don't know about you, but I didn't even start to play chess anything like seriously until my mid-fifties, and I've never set foot in a chess club in my life. So I guess I would be a prime suspect if I ever improved enough to win some games
(I hadn't played much at all for a while and suddenly I was playing 6 games at once in a no vacation tournament I got involved in. I lost 4 of the 6 games on time over Xmas while I was staying at my brother's, so my rating has plummeted. Then today I finally lost a game that I had been losing for ages, so it dropped some more 
)
In v3 there is no clock in TT. Or at least, if it does exist, it is switched off by default. And the algorithm has changed completely. Now you get a minimum of +1 if you get the problem right - you only get a minus mark if you get it wrong, which is fair enough. My TT rating has improved by over 200 points, just because of the new algorithm, which corroborates what I have always said about TT scores, namely: if you want your TT rating to improve then you should go with your gut instinct, take a chance and move quickly for a big plus score; but if you want to improve as a chess player, then try to see the problem right through to the end before you move anything. (Occasionally I might break this rule if, say, I can see how to drive the opponent's king into a blind alley but the end is several moves away and I can't quite see the finish.) I have played TT this way for years, even under the old system, so the new system has helped my TT rating quite a lot. I get an awful lot of +1 scores for getting problems right but very slowly, and I get a few big minuses and a few big pluses. Previously I would have had more minuses where I now get +1. Probably my pass rate has remained the same and (sadly) I doubt if my skill level has improved much at all. But my rating says it has.
What has all that got to do with the cheating debate? Well it proves, to my satisfaction at least, that looking at someone's TT rating and comparing it to their Daily Chess rating is a complete nonsense! The old scoring system was designed to reward people who were prepared to take risks. So is the new system, but to a lesser extent. So your TT rating could maybe be compared to your Bulllet or Blitz rating, but not to your Daily Chess rating (or VC performance) where you should never make a move until you are satisfied that you will not find anything better.
From memory my TT pass rate is around 65%. I have never reset my TT so the 65% includes a time when my pass rate was much lower (I was trying to beat the clock, and failing). Nowadays I pass around 75% to 80%. I once looked at the pass rates of some of the TT superstars, who are rated around 4000-5000 (but their ratings fluctuate wildly); one guy who was in the top 5 had a pass rate of about 55%! I'm not saying that I'm better than him, just that I'm playing a different game.
You should try TT in chess.com. I think it's pretty good now. My biggest complaint nowadays is the there are still some old problems with multiple good solutions, but the admins don't remove them when you report the bad problems. e.g. You see a mate in 4 and you go for it. The analysis says there is a mate in 2 so you are wrong. But, the analysis has not seen the mate in 4, because once it finds the mate in 2 it only searches to a search depth of 4 (in case there is an alternative mate in 2) and beyond that it doesn't care. Later the policy changed and they will cut the problem short, or scrap it, if there is an alternative "winning move", but they haven't weeded out all of the bad problems yet. The silly part is that their new Stockfish engine in TT shows up the bad problems for all to see!
After suspicions were raised in the Cheating Forum recently because my TT rating is so low compared to my Daily rating, I invited the cheat-detectors to carry out an analysis on my own games & this is what resulted (with comments):-
"Pretty good for an untitled player, but not meaningfully beyond what a strong untitled player should be capable of if he takes correspondence seriously."
50 games
UNDECIDED POSITIONS
Positions: 758
T1: 205/551; 37.21% (std error 2.06)
T2: 243/441; 55.10% (std error 2.37)
T3: 257/395; 65.06% (std error 2.40)
>0 CP loss: 369/758; 48.68% (std error 1.82)
>10 CP loss: 286/758; 37.73% (std error 1.76)
>25 CP loss: 196/758; 25.86% (std error 1.59)
>50 CP loss: 106/758; 13.98% (std error 1.26)
>100 CP loss: 40/758; 5.28% (std error 0.81)
>200 CP loss: 10/758; 1.32% (std error 0.41)
>500 CP loss: 0/758; 0.00% (std error 0.00)
CP loss mean 22.00, std deviation 42.98
LOSING POSITIONS
Positions: 56
T1: 5/31; 16.13% (std error 6.61)
T2: 9/22; 40.91% (std error 10.48)
T3: 13/19; 68.42% (std error 10.66)
>0 CP loss: 31/56; 55.36% (std error 6.64)
>10 CP loss: 25/56; 44.64% (std error 6.64)
>25 CP loss: 18/56; 32.14% (std error 6.24)
>50 CP loss: 10/56; 17.86% (std error 5.12)
>100 CP loss: 8/56; 14.29% (std error 4.68)
>200 CP loss: 5/56; 8.93% (std error 3.81)
>500 CP loss: 2/56; 3.57% (std error 2.48)
CP loss mean 57.73, std deviation 138.01
I'm hoping these might help as a kind of baseline for a player around 1900/2000 but who has no comparable OTB experience. That's to say for someone who has a very analytical style of play.
* BTW, my TT rating is awful because I disabled the rating function early on in order to disable the clock, which used to irritate me. I reason that being an exclusively long time control player, constantly being timed was of no benefit to me. I now use the TT facility on GameKnot where I have a rating of 1656 but haven't used it in more than a year.