Tactics trainer: Kindaaa dumb, do you agree?

...a lot of the problems, anyway. I'll start out by acknowledging that for lower rated players, the easier tactics trainer problems (up to ~1500) of just finding the hanged piece or simple checkmate pattern are pretty important.

But the ones I've been seeing recently are getting a little ridiculous. The main complaint I see again and again in the comments for the problems is this: you have the option of choosing a line where you get a huge material advantage and then can safely win shortly thereafter, but this is the wrong answer because there was a convoluted, more risky way to get a checkmate in 2 fewer moves.

Ok, obviously, it's better to see the checkmate. I can't argue that. If you could be sure that it would work, it could save you in a temporally desperate situation whereas the 'safer' line could take just too long and screw you.

But despite being totally 'deterministic' in the sense that there aren't literal luck factors like a die being rolled, chess is still 'probabilistic' if you don't have infinite computing power (I'm guessing most people don't). This means that you play moves that don't have a direct tactical purpose, but will "probably" yield good results in the long run. That's really all positional play is. I'm just saying this because, given a situation in a game where I'm not incredibly pressed for time, I will take the safe route of grabbing their queen and being way up in material and ending the game in 10 moves as opposed to trying a checkmate in 3 moves that puts a lot of my material on the line, even if I'm fairly sure there's nothing wrong.

Sometimes in actual live games I try a risky maneuver like the ones TT teaches. Sometimes they pay off and I have a clever little victory, but usually they don't, I lose the material I put on the line, and lose the game. Why? Because, unlike the problems in TT, where you know there is a clever little solution, in real games there just often aren't ones. That's my main complaint with TT: It's basically teaching you to constantly look for slick little combinations that are basically the glamor shots of chess, but are actually rare enough that spending any significant amount of time on them is silly and maybe even harmful.

Related to this is the queen sacrifice in TT. It's really becoming kind of a joke, and you can see it in the comments too. Whenever there is a problem involving a queen, my first thought is "Ok, how can I sacrifice the queen in the most non intuitive way?", and, honestly, that mentality is effective in TT. But I don't think it's very effective in the majority of games. Effective queen sacrifices are pretty rare. I suspect the disproportionate amount of them in TT is because the queen sacrifice is the most popular and romantic move in chess, and when a person gets to do it they get all hot and bothered.

I don't think it's bitterness at not getting a problem, too. Sometimes I'll get a problem wrong and say to myself, wow, that was a good one, I really should (and could) have figured it out. But with a lot of them I just find them obnoxious, because I look at the solution and think, this is dumb, I didn't get this here, and definitely wouldn't get this in a real game.

After all my complaining, what are my suggestions? I dunno, I just like complaining. But I do think the most educational problems I've had are the ones with common mating patterns I actually get to use and also slightly more advanced ones that just combine a few basic tactical things (like fork + discovered check, or a pin + overloading).

I'll leave you with this:

Chesstempo.com has an excellent tactics trainer and has improved my score by 200 points so far. I have a few complaints about chess.com Tactics trainer that's why I just use chesstempo.com and it's free. Chess.com is still my most favorite site, TT needs alternate moves that can be made,comment filter, and a way to show a diagram of a chessboard where you post a comment.

@OP: Nice post with good points.  TT has its flaws of course.  I don't know what you're expecting to get out of it. I use it as a test each morning to get my brain thinking tactically before I jump into a game.  If you're seeking a rating improvement, I agree that there is a limit to what TT will bring to your game above 1500 or whatever.  Ultimately, I think what you're after is a pat on the back for coming up with a workable solution rather than "the" solution and that's the source of frustration.  It's an aid, not a teacher, and getting it wrong doesn't mean your solution wouldn't have won you the game, you're right about that.  You seem to be approaching it in terms of whether it would win you the game, but that's not the goal of TT. TT just aims to improve your ability to find tactics.  So yeah, it's kinda stupid, but it's also kinda helpful.

Don't play the first move of the tactic until you see the combination to the end. 

The timed thing is dumb and forces you to gamble on those intuitive moves, which as you said doesn't work in real games.

If you absolutely can't see it and give up, then try to cheat your way though the tactic one move at a time... I say "cheat" because again, that would never work in a real game would it

I actually have the nagging suspicion that many people do that anyway. Otherwise I can't explain all those "easy" posts at daily puzzles in cases where the first move was maybe somehow "obvious", but there were still quite some variations to calculate to make sure it really works. As a correct move played solely by intuition usually prunes the remaining search tree quite alot, the whole problem may then seem easy from hindsight.

Why not have multiple correct responses instead? Mate is mate and in real games multiple mates are often possible. Having several equally correct moves might be more of a programming challenge obviously, but it strikes me that such a situation is representative of real life.

I absolutely agree. 

Fair enough answer. 

I agree with wafflemaster about the timed element.  I saw a post from an IM somewhere that said if you want to improve with TT, turn the timer off, and I fully agree.  What's important for the improving player is to see and get the tactic right, not to get it in X amount of time.  The timer is unnerving and as wafflemaster said, forces you to take the intuitive move.  You end up losing points if you take too long getting the problem right.  It trains you for blitz playing, but not for a "full clock" game.   I think the timed rated system should be an option.  But I wish there was a way to do the untimed option and get, if not a rating, a least a percentage of the number you've got correct.   Other than that, I think it's a wonderful tool. 

I made a similar posting in the "My Ideal Tactical Trainer" thread but thought I would repeat it just in case chess.com staff might see it.

Many times when I miss a move on the TT, I don't understand the solution and my only recourse is to "view analysis and source" which many times doesn't make sense (to me). Why can't the puzzles be carried out to an obvious conclusion. I'm sure higher ranking players can understand but for us patzers, playing out the puzzle to the obvious would help in understanding the objective and therefore aid learning.

Thanks.

Doesn't the puzzle have an "ELO" rating that is pitched against the puzzler and adjusted according to who "wins" (i.e. the puzzler or the puzzle)? At least that's how it is done over at ChessTempo's. A puzzle may be over- or underrated initially, but after a while its rating reflects its difficulty well.

I agree.  And I do it too, of course, thinking "this looks like a thematic first move, I'll try it and see if it's right"

But especially if you're trying to improve, this really isn't a good way to do your puzzles.

Chesstempo.com's tactics puzzles are rated by how many people pass and fail the problems. Wehn you don't understand a problem the comments will always walk you through notation or a mini board. Any dumb comment is usually voted down so you don't have to read it. The best part is you can play timed and untimed rated. All free.

Chess.com is still the best!!!

I think it is a bit more complicated. If a higher rated tactician fails a problem, the problem should gain more rating points than if a lower rated tactician fails it. Just as you lose a lot of points when you fail a problem that is rated much lower than you and vice versa - at least that's what I always thought how it works.

I use chesstempo.com mainly for tactic definitions

Your right bubatz.

I've given up paying attention to my Tactical Trainer rating. I take my time in order to avoid jumping to a move to beat the clock. I found myself carrying over the "quick move" syndrome to my games. Now I give my TT puzzle moves the same discipline I use in a game. I figure that over time, I will still learn the tactical patterns, won't get frustrated with rating loses due to quick moves and will continue to  reinforce my actual game discipline.

If you look at what happens after you answer a puzzle in the TT you will see that what you are saying is quite obviously not true (because you will see the rating change based on whether you got it right or wrong).  The puzzles are rated completely objectively.  They are given ratings which are constantly changing as players get them right and wrong.

Moreover the puzzles don't even have "authors".  They are taken from actual games (which you can see by viewing the analysis and source).