I definitely agree. this may only be a good training tool if you like ar are good at blitz. it would help you see tactics more clearly. If you want, I could send all my 708 friends emails for a petition. Then I could talk to my three Admin friends and one national master friend to see if they could help me.
Tactics trainer needs a fixed time mode
Sort:
The Chess Tactics Server is always an option...
Agree completely. The entire methodology rewards you for guessing. This is one of my two biggest issues. The other issues is they never tell you what you are playing for then mark you wrong for simplifying to a won position.
One of many examples: Problem ID 0074507
I found the idea of Ba5 pinning the queen, losing my queen. I had seen the winning line where I play Bxc7 and get a protected passed pawn on the 7th rank in a King and Pawn endgame (after Kxc7,dXc6) with a bishop thrown in for good measure. Apparently I'm supposed to play dXc6 first, and black will move their king away surrendering the position (after which I take) instead of playing pawn moves.
I find it ludicrous that I get 0% despite finding the main idea of the problem and forcing 1-0.
I agree that TT awards guessing but fixed time limit is bad solution imo. Some problems are naturally more time consuming than others.
I believe that TT scoring system should be somehow tweaked in such a way that it better awards correct solutions and penalises the wrong ones. It somehow feels very wrong that if you take say 6 minutes to find a correct 5 move deep solution to a problem where average time is around 3 minutes but passing rate only around 20 percent then you score less than someone who just gets a partial solution inside 2 minutes.
Since I have been using it untimed, I have noticed something. After the problem is posted, they still give the rating.
Some of the problems are incredibly easy. In some cases, an enemy piece is hanging in front of a pawn, and the solution is to use the pawn to take the piece. That's it. I noticed that those sorts of problems often have a much higher rating than more typical problems that require finding a fork, or a mate in two, or something similar.
I can only assume that their high rating is due to the fact that they have extremely short time limits in rated mode. A good player will scan the board quickly and find it right away. A weaker player, like me, will tend to first look at whether the enemy king is under pressure, and if there is some pressure try to find whether or not there is a checkmate possibility. (After all, as noted earlier, if you take a queen and end up with a won position, but there was a mate in three available, you still lose.) The strong player might take five seconds to analyze the king's position, determine there is no checkmate possibility, and look elsewhere. It takes me 30 seconds to reach the same conclusion, with a lower confidence level. By the time I've decided that there is no checkmate available, it's too late to find the obvious move.
I never miss those in untimed mode, but in timed mode, I frequently overlook those moves. Apparently, so do a lot of other people, because their ratings are much higher than the true difficulty of the problem.
Shakaali wrote:
I agree that TT awards guessing but fixed time limit is bad solution imo. Some problems are naturally more time consuming than others.
No system is perfect, but, this suggestion is one that would be better than the current system, and easy to implement. Meanwhile, generally, problems that take more time are harder than others, and the ratings will reflect that. Hard, time consuming, problems will have high ratings and weak players won't get them in the allotted time.
This contrasts with the current TT system, where very easy problems often have higher ratings, because the shorter time limit is too much for the weaker players.
I suggested 3-5 minutes for the time limit because that is about the amount of time a typical player can spend on a problem before his patience runs out. At that point, you have stared at the problem a while, and when you do get the solution, you actually might learn something from the problem. In the current TT system, you spend very little time on a problem you miss, and throw out a guess on, and it just seems more natural to automatically hit the "next" button without giving it much thought.
You could fairly easily allow multiple time modes, based on player choice. A 30 second mode and a 5 minute mode, or even a 10 second blitz mode and a 20 minute thoughtful mode, with each problem getting a different rating in the different time limit modes. I'll bet the relative order of problems would be at least somewhat different in the different modes. Technically, it wouldn't be hard to do.
Of course, the possibilities are limitless. The main point is that the current system skews the problem rating by having variable time limits, and prevents the tool's optimal use as a training tool. One way to improve the situation would be to have a five minute, fixed time, but rated, mode.
I agree. The simplest would be to have two solving modes. One timed, as it is now, another untimed, with different rating (typically much higher than timed). This is already used in another site with tactics, and it works.
Ok, chess.com, the competition has it. Sounds like a market demand to me.
Meadmaker wrote:
The current system works ok to test tactical ability, but because it encourages people to guess as the clock runs down, it doesn't work as a training tool as well as it could.
In fact, the current setup does the opposite, the more your time runs down, the less you are rewarded for guessing.
Guessing as your time runs down only saves you a few percent of your grade compared to taking your time. Once your time is up, you can still get 20% for getting the problem right no matter how long you take. You will be more rewarded for taking your time, getting the problem right and getting 20% than you will for taking a chance with a guess as time runs down and trying to get 25%.
Loomis wrote:
Once your time is up, you can still get 20% for getting the problem right no matter how long you take.
If you take all of your time in tactics trainer, and then find the right answer, your rating drops. This encourages guessing when you see that time bar approach the yellow zone.
Some people have recommended a change in which the scoring system is altered so that your rating never drops for a correct answer. That, too, would be one way of addressing the problem with tactics trainer.
I understand that your rating drops if you take all your time. You only get 20% credit for the answer.
Consider that your rating drops by the same amount if you take twice as much time, because you still get the same 20%. But your rating drops even more for a wrong answer. As you get closer to being out of time, the risk/reward ratio for guessing gets smaller, not larger because you stand to get 0% for a wrong answer but you can always get 20% no matter how long you take.
I like Meadmaker's suggestion of different modes for tactics trainer eg. Blitz, Standard, Long.
If you search for tactics trainer according to problem time, there are some with massive available times. I wonder whether the time allowed is largely influenced by the time taken by the first few people to do the problem. I would be very interested to know what the rules are for altering the available time after each attempt (perhaps a simplified version if the maths is a bit complex).
Now I always think through all the possible lines before making my first move, even if the first move seems obvious. Its far more rewarding, and reveals many interesting side lines. However my rating stays around 1500, yet I can solve 2500+ problems with enough time, and I am being denied the chance to solve trickier puzzles in training mode.
I realize it has been a long time since this post has been updated; not sure if it's better to start a new one or append.
I agree with zyga65 ; I would appreciate it if Tactics Trainer had two modes (and different ratings to go along with them) -- the current, timed mode and an "untimed" (but rated) mode.
I use the tactics training on lichess.org, and find it very satisfying because it is untimed -- well, it times you, but it doesn't penalize you for taking a long time. It is extremely frustrating to spend the time to get a problem "right" on chess.com's Tactics Trainer only to get penalized for it. And, as EndgameStudent says, "my rating stays around 1500, yet I can solve 2500+ problems with enough time, and I am being denied the chance to solve trickier puzzles in training mode." -- my sentiments exactly.
So chess.com staff, if you are listening, please consider this. Thanks!
The 'mixed' set at ChessTempo gives you a rating and doesn't substract any point if you find the solution within 5 minutes.
As an added benefit, it offers both attacking and defensive problems 
This was posted on another forum by staff 11 months ago:
DeepGreene wrote:
However, with the advent of V3 later this year, we will be making another change to tactics logic (scoring this time) that we think most people will like (based on a recent survey). Here's what that will entail:
- Percentile score remains at 100% until the player uses up half of the problem's average time-to-solve. (This is an extension.)
- From that point on, the score drops steadily, reaching 50% at twice the average time-to-solve. At that point (a theoretical "draw" vs the tactic), the score stops dropping based on time. (It currently does this only after reaching 20%.)
- Any time after that, if the player solves the tactic correctly, the server will assess whether the player should get any points for a 50% score (i.e. if the tactic is rated higher than the player). If that's the case, then the player gets those rating points. In all other cases, the player's rating does not change. A player's rating will never drop as a result of correctly solving a tactic.
After much frustration, I switched to untimed mode in Tactics Trainer, which also means I switched to unrated mode. I think it is helping me a lot more than the old way, because now I actually think through every problem. In rated mode, I almost never got the "fast" problems correct. However, I miss the challenge of trying to improve my rating.
Tactics Trainer would be a better tool if there were a "fixed time" mode, of perhaps three minutes for every problem. Maybe even five minutes. It would be easy enough to do. Each problem could just have two ratings, the "classical" rating, where both the problem rating and the time limit changed based on player evaluation, and the "fixed time" rating, in which only the rating changed.
The current system works ok to test tactical ability, but because it encourages people to guess as the clock runs down, it doesn't work as a training tool as well as it could.