Andragogy is the art and science of helping adults learn. The art and science in this field are somewhat different from that for teaching children. This has technical and learning-style implications for teaching adults to improve at chess or advising them how to self-teach and self-train at chess.
First, a quick note about how this post was written, It was written by spit-balling with Google AI. I will call it spit-balling rather than researching. Excerpts of Google AI answers will be repeated verbatim but not in quotes and also they will be amended by my comments which will not be flagged as such either. The results may be taken at face value as possibly reasonable advice to be tested in adult self-training in chess. This is certainly the way I will proceed empirically to further amend and test my self-training efforts. I first note below why this should suit (very motivated) adult learners.
Some traditional chess training relies heavily on the "immersion method"—forcing a player to consume tens of thousands of random puzzles. For a child with an incredibly high rate of synaptic pruning and neuroplasticity, the brain automatically organizes this chaos into patterns. But for an adult improver, whose brain requires deliberate structure, high-relevance repetition, and error-focused feedback to overwrite existing neural pathways, this random approach is deeply inefficient.
The Cognitive Science Behind It: Adults learn best through error-amplification. When an adult brain makes a mistake, it releases epinephrine and acetylcholine, which chemically flag that exact moment for neural restructuring. By immediately flooding the user with similar, compounded variations of that specific mistake, your adult system would capitalize on that hyper-focused neurological window to force pattern recognition. (The epinephrine and acetylcholine claims could certainly do with research checking which I have not done.
The inability to execute chess tactics stems from three sequential bottlenecks. In cognitive psychology and chess coaching, this progressive failure is a well-documented breakdown moving from (1) limited pattern recognition (instinct) to (2) limited visualization (mental simulation) to (3) limited calculation ability, (managing branching complexity).
When any of these three layers fails, your tactical ability collapses. Here is a breakdown of how these failures happen and one possible targeted, systematic training regimen to address each level or phase:
Phase 1: Inability to See One-Move Tactics Fast Enough.
• The Root Cause: A weak pattern vocabulary. If you cannot spot a one-move tactic instantly, your brain’s "System 1" (intuition) isn't recognizing the configuration of the pieces. You cannot calculate a deep combination if you are spending all your mental energy trying to notice that a knight fork is possible.
• The Training Fix: Volume Drilling. One approach is to take pools of say 100 very easy tactical puzzles at a time using platforms like the Lichess Puzzle Themes.
Solve these under some time pressure at a rate that suits you, preferably where your success rate is in the zone of 85% to 95%. Re-do the ones you fail the next day. Lichess does not easily support more re-doing of two failures in a row etc. But the sheer volume (at least x hours a week, you choose) should ensure recurrence of failed motifs in variant forms.
• The Goal: Build automated chunks in your long-term memory so your brain flags close threats and opportunities (1 to 2 or maybe 3 moves) almost instantly without active calculation.
Phase 2: Inability to Calculate One Line Deeply (2+ Moves)
• The Root Cause: Poor visualization and a failure to identify forcing moves. Players often suffer from "wishful thinking," imagining their own active moves but failing to accurately visualize the opponent’s best responses, causing the mental image of the board to blur.
• The Training Fix: Quality over Quantity & "CCTV" Checks
• The Routine: Switch to medium-difficulty puzzles (Mate-in-2 or 3, or win-of-material lines).
• The Method: Strictly do not move a piece until you have visualized the entire line to the end. Sit on your hands. Use the CCTV scan (Checks, Captures, Threats) for both you and your opponent. For every move you make in your head, force yourself to look for your opponent's most forcing reply (their checks and captures). Write down your intended variation on a piece of paper before executing it to keep yourself honest.
• The Bonus Drill: Pawn Endgames. Solve simple king-and-pawn endgame puzzles. Because there are no long-range pieces, they are perfect for cleanly extending your visualization depth (e.g., counting out pawn races 5 moves ahead) without your mental board blurring.
Do at least x hours a week, you choose. Note the "x" means each of these Phases should be 1/3 of your tactics training time of this type.
Phase 3: Inability to Calculate Branching Lines (The Tree of Analysis)
• The Root Cause: A lack of calculation structure. When a position has multiple viable moves, an untrained brain jumps back and forth chaotically between lines, gets fatigued, forgets what it already calculated, and blunders.
• The Training Fix: Candidate Moves & Kotov’s Tree
• The Routine: Use complex, tournament-style puzzles where the solution isn't a dead-obvious sacrifice etc.
• The Method: Apply a strict, systematic mental protocol:
1. Identify Candidate Moves First: Before calculating a single line, list 2 to 3 attractive starting moves. Do not dive down the rabbit hole of move A without acknowledging move B.
2. Calculate One Branch at a Time: Pick candidate move A. Calculate it down its main line until the position stabilizes. Evaluate it ("White is winning a pawn"). Now drop it completely.
3. Switch to the Next Branch: Move to candidate move B. Calculate it cleanly, evaluate it, and compare the final evaluations of both branches.
• The Goal: Eradicate chaotic looping. Treat calculation like a checklist, not a random maze.
(I will add further possibilities to do with this if this post gets any interest or positive comments.)
(Note: some sources suggest setting Phase 1 at 25 % of this tactics training, Phase 2 at 50% and Phase 3 at 25%.)
I have considerably more I can post on these Andragogy issues. I have looked into it at some depth but it does remain speculative and experimental on my part. I do not say what I have written so far is news to chess coaches working with adult improvers but this information takes some digging out if you do not have a chess coach and you are attempting to adult improve at chess by self-coaching and self-training. I will also note I am old, 72 years. This puts me near the extreme end of the spectrum of adult learners (from a low base) and not the good end. But this extremity throws into high relief my need for self-applied andragogy if not gerontagogy if there is such a thing.
I hope @Benedictine and @NDPatzer among others will find this interesting.
Your post is hard to read. I guess that is the fault of Google AI.
"tens of thousands of random puzzles"
* That is poor training. Chess should be learned backwards: endgames first, then middle game, last opening.
"Adults learn best through error-amplification"
* That is why playing and analyzing lost games is so good.
"Solve these under some time pressure"
* You should solve puzzles as the same pace as you play.
It makes no sense to solve a puzzle in 5 minutes and play at 5 seconds / move.
"Kotov’s Tree"
* yes, Think Like a Grandmaster by Kotov is a great book.
"gerontagogy"
* the key to gerontagogy is physical exercise. The brain is an organ and to function it depends on good physical shape of other organs: heart, lungs, bowels, liver, kidneys... Most grandmasters have a physical coach. The elder you are, the more it is needed.
Kotov's "Tree of Analysis" sounds good in theory, but I believe that Adriaan de Groot (Thought and Choice in Chess) has shown that chess players don't strictly follow Kotov's idea. It's more a combination of branch analysis combined with positional "chunking". The stronger the player, the better he can analyze branches, but also the more "chunked" positions he can instantly recall.
I didn't read the whole OP but yeah, adult improves have to be more systematic about learning.
@tpr said [^](/forum/redirect/post/WJnbagXy)
> "Solve these under some time pressure"
> * You should solve puzzles as the same pace as you play.
> It makes no sense to solve a puzzle in 5 minutes and play at 5 seconds / move.
This is a good point. Taking time to solve puzzles & then playing bullet is basically wasting your study time.
In response to the replies above. My post above was about tactics training, not the whole of chess training. I should have been cleat about that. When it comes to tactics training for adult improvers, the three phases or levels suggestion for tactics training makes clear the idea you train at three different paces; a different pace at each level. It makes sense to me and I am going to try to fully implement that.
The fastest tactics training is for board sight and seeing mainly one-move and simple two-move tactics very quickly. Each phase after that slows down the training for medium and harder puzzles (relative to the player's ability). In each of these phases, the player tries to implement conscious, logical thinking by spotting loose pieces, pins already operating, vulnerable king positions and so on and then trying to find the tactic(s) by set protocols like CCT (Checks, Captures, Threats) and TIM (Take, Interpose, Move) for escaping checks and for checking that your checks on the opponent's king can lead to checkmate rather than its escape. There are more protocols too.
In summary,
Level 1 for board sight and fast, simple tactics.
Level 2 for longer but still mostly single branch tactics.
Level 3 for tactics of two cacalcation branches usually (and later on more.
Training in the protocols like CCT is for training only. They are "training wheels". In real games, especially Rapid and faster, you do not have time for this type of slow, formal logical thinking. The training is meant to slowly make these thinking methods become automatic and rapid without slow, conscious steps. This takes a longer time to achieve for adults and old adults.
In addition, it is recommended that adult learner do batches of like tactics/motifs in blocks (like forks etc.) for about 2/3 rds of this type of tactics training and then blocks of mixed puzzles (like Lichess Healthy Mix) for 1/3 of that training. The idea is to first get better at the tactic, like the fork, and then learn to find it in the jumble of a mix of tactics.
This is a short description of the tactics training idea. The method is meant for players like me (about 1525 in Rapid 15+10) who are still very poor at tactics and blunder a lot. This won't be for all and you might not agree that this is the correct way to train in tactics at any level. But I am having a go at doing it this way over time and in a more and more structured fashion as I go. If I can't get to some basic competence in tactics I will likely have to quit chess and I don't want to do that yet.
"My post above was about tactics training, not the whole of chess training."
* That was unclear, the title suggested otherwise.
"the three phases or levels"
* I do not believe in it.
"I am going to try to fully implement that"
* You can measure it by your rating going up or not.
"In real games, especially Rapid and faster, you do not have time"
* then play a slower time control: 15+10, 30+20, 45+45.
"tactics/motifs in blocks (like forks etc.)"
* I do not believe in thematic training. In a real game you have no clue about the theme.
"blunder a lot"
* Key is blunder prevention. It needs no training, only mental discipline. Think about your move, calculate, evaluate, decide, then do not play your intended move, but imagine it played on the board and check it is no blunder: does not run into checkmate, does not hang any piece or pawn, only then play it. A great help is to enable move confirmation in your playing profile. Play your move, check it is no blunder, confirm.
When you tell child, solve this puzzle. He will begin to solve.
When to tell an adult, solve this puzzle. He will answer I don’t want it, I don’t need it, and will not solve the puzzle.
@Blivot said [^](/forum/redirect/post/t7X6QJXo)
> still very poor at tactics and blunder a lot
Hey?! Do you solve 350 and 250 puzzles a day???
I don’t understand how you solved 108 is correct, 8 is incorrect, and for this you received a -250 rating?
@RemRus said [^](/forum/redirect/post/pzGG3jSP)
> > still very poor at tactics and blunder a lot
>
> Hey?! Do you solve 350 and 250 puzzles a day???
> I don’t understand how you solved 108 is correct, 8 is incorrect, and for this you received a -250 rating?
If you do puzzles on "Easiest" (-600) you will often get 0 or 1 point for a win and -45 (something like that) for a loss. I change the puzzle level for different levels of training. "Easiest" is to get quick puzzles for basic pattern reinforcement. I also do puzzles at "Easier" and "Normal". It sends my puzzle rating all over the place. I haven't worried about that but it makes next puzzle difficulty a bit too random.
What I could do (and might do) is get my puzzle rating to a set number. A rating of 1800 would be good. Then I would do all puzzles on unrated. I could do puzzle sets on different levels as required for different types of training (phases 1,2 or 3). The levels of easiest, easy and normal would then equate approximately to 1800-600 = 1200. 1800-300 (1500) and 1800. That would still not fully solve my issue with getting puzzles of the right difficulty. But it might be the best I can do with random, non-curated puzzles.
@Blivot said [^](/forum/redirect/post/CI8cefuH)
> If you do puzzles on "Easiest" (-600) you will often get 0 or 1 point for a win and -45 (something like that) for a loss. I change the puzzle level for different levels of training. "Easiest" is to get quick puzzles for basic pattern reinforcement. I also do puzzles at "Easier" and "Normal". It sends my puzzle rating all over the place. I haven't worried about that but it makes next puzzle difficulty a bit too random.
if you click/tap “analysis board” in one of your past games, and scroll down, you will see a hyperlink that says recommended puzzle training.
i think those puzzles relate to the game you played.