Love the updated Tactics Trainer

So what has changed to one step forward, eleven steps back?

one step forward, twenty steps back?

The ratings on TT are absurd.  They change TT so much that its ratings are worthless in gauging improvement.  When I continue TT, rather than my other software, apps, and platforms, I'm sure I'll be 2500, despite having a USCF rating in the 1800's.  Such a joke.

I kind of agree with Milliern.   With the updated TT I've quickly pushed my tactics rating up over 150 pts to from ~2080 to 2250+ !   This seems rather absurd to me, as I've never had any other kind of chess rating above 2000 before in my life.

Yeah! I just busted 1900 a few minutes ago. I have been playing 1 minute for so many years, I had pretty much forgotten how to think deep. TT has given me back the joy of what chess is really about. That crucial moment, where you need to make the right decision to outsmart your opponent and win the game. I hope the TT  rating system is close to what actual FIDE rated players score, because my 1 minute and even the daily aren't even close to 1900. It does seem easier now.

Yeah I looked at some of the higher rated players TT ratings and saw that. Still I very much enjoy them. One nice thing about the TT rating is it is cheat proof.

I don't care about my tactics rating. I just want to learn tactics. 

 Yeah, it happen to me. In my last session I get the 5 free tactics right and I think I increase like 70 points or more. Before they give +7, now I'm getting like +20 if I get them in time. 

Correct.

It looks like chess.com is not aware that a change in tt-ratings behavior is devaluating the existing

rating from people they gained over a longer timeperiod. Not comparable anymore.

I still love what they have done. Previously, the trainer bound you to problems near your current rating. If you didnt get the problem right in a short time, plus 1 point. Get the problem wrong in any amount of time and you got a minus 11. Before, I noticed that rest allowed me to temporarily increase my my tactics by a hundred or so points in a session. The next session, during lunch or immediately after work or late before bed, I would solve a few, miss one, tap a wrong square, or get the sequence out of order but clearly figure out the motif. This led to a downward spiral to back where I was. Assuming I do learn from my mistakes, it was discouraging to make little progress. Consider my journey almost a year ago, getting back into chess after a 35 year layoff, I have made steady tactics improvement over the year with a spike at the end of 2016, when I got plenty of rest each night between Christmas and New Years. After that, I pretty much held my tactics rating for months. Yet, I know I recognize motifs I didnt even know back last year. Only now does the rating reflect that, because I can solve more difficult problems that never presented itself and Im not slaughtered for missing a problem that is near my current rating. My uscf rating began in Oct 2016 at 1433P, and it should be close to 1500 after my last tournament where I tactically destroyed a 1714 in a transposed Smith Morra Gambit opening and went 2 of 4, losing one game to the winner, a former 2050 player and another to a really solid 1700. I think my current rating of 1800 in TT is near what I believe my tactical ability is at this time. I base this on the assumption my tactics score should be higher than my uscf rating because it takes time and experience to apply what I have learned in a rated tournament game.

For anyone who doesn't follow the site development news, the tactics trainer changes were announced at the end of April. They devalued a number of stratospheric ratings (people approaching 10,000 tactics ratings) and announced a "target time", which was supposed to take your rating into account on how long you should take to get full credit on finding the solution. Unfortunately, the people gaming the previous system appear to have skewed the target and average times, leading to ridiculous target times of only several seconds for some intermediate problems. As a result, people solving problems slowly would get only 1 point but anyone missing a problem would lose 11.  Many ratings started heading downward.

There was an immediate outcry as some people started losing hundreds of points in a day. After a few days of people threatening to quit if the system wasn't changed back, they did take a look at all the comments and obviously made some updates a few days ago.

It appears they may have overcorrected.

I appreciate what they are trying to do. It is not easy to change such a system without somebody getting upset. Could we give them at least a few weeks/ months to work out the kinks here?

 Well despite TT being way off actual ratings, I finally realized why I am so high and my real rating are so low. TT represents the "meat" of the game. How well and how fast you can assess a situation and come up with the best move in middle and endgame. It doesn't factor in the "opening library". My father taught me the game when I was 12 or 13 not sure which. When I beat him for the first time at age 15, he never played the game again. That was in 1970. In 1972, when Bobby brought chess in the US to an all time high, I bought Fred Reinfield's book on chess, which did have all the openings and their names, but it didn't explain how these opening lines needed to be memorized, or you would never make it to a winning position by midgame.  Now I am too old to memorize them all, but my high score on TT, evidence of how well I can think, once I get to the fun part of the game. Anyone rated over 2000 knows all the opening lines, which is a monumental task and takes years. Deep Blue didn't beat Kasparov until the techs loaded the computer with all known opening lines. Bobby read 4000 books on chess, before winning the world title. I love chess, but admittedly I was too lazy to "do the work" it takes to have that opening library. I feel not taking into account the opening strength of a player, skews the TT rating, but that doesn't explain why GMs have much higher TT scores as well.

I think that Chess.com has always missed the point that TT is a training tool.  As I get better at tactics, the problems should become more difficult.  That means that I will miss more problems until I learn the techniques for the more difficult problems.  Then the process should repeat itself.

With the emphasis on ratings, as the problems get more difficult, and I miss more, my rating goes down, and I then get easier problems because "I've forgotten what I learned before."  The problem isn't that I've forgotten, it's that I have more to learn.

The analogy that comes to mind is when I was in school.  I wasn't sent back to 7th grade math class because I was having trouble with 8th grade math.  I was given 8th grade math material until I learned it.

I like the school analogy woton, thanks. I also noticed my TT rating went too high, even though I dont think I have improved that much. I guess its the stabilization period of the new algorithm. It just feels too easy sometimes.