You find a combination you could not see a minute earlier. You smile and start another exercise. Or you win a game and immediately click “Play again”.
Wanting to continue is a real experience. Describing it in terms of dopamine is tempting; we still need to distinguish what was measured from what we imagine.
Reward is not a single sensation
Berridge and Robinson (1998) distinguish, among other things, the motivational attraction of a reward from its subjective pleasure. Their review discusses dopamine’s role in these processes. It does not justify saying that a pizza and checkmate produce exactly the same brain response.
You may want to start another game while no longer enjoying it much. This distinction helps describe your experience without claiming to measure neurotransmitters.
What prediction error shows
Schultz, Dayan and Montague (1997) link dopaminergic-neuron activity to the difference between expected and obtained reward. In the experiments discussed, an unexpected reward may trigger a response; after learning, the response may shift to the predictive signal. The absence of an expected reward produces a different pattern.
This is a framework for understanding some learning processes. It does not demonstrate that every opponent blunder triggers more dopamine than your own good move, or that one Elo point corresponds to a quantity of neurotransmitter.
The moment you see the solution
Kounios and Beeman (2014) review the cognitive neuroscience of insight, the experience of suddenly discovering a solution. Their methods and tasks do not constitute dopamine measurement during chess-puzzle solving.
You can therefore say “it clicked” without adding “my nucleus accumbens released a spike”. Lived experience, activity measured through imaging and the release of a molecule are different levels of description.
Why comparing games requires caution
The sources presented do not rank chess, Go, video games or gambling on a dopamine scale. Nor do they show that every classical game is healthy while every blitz game is compulsive.
An interface enabling successive games makes stopping easy to postpone: that is an observable property of the design. Saying it “saturates the circuit” or requires twenty-four hours of recovery would need specific measurements not provided here.
Observing what you actually do
For your practice, begin with concrete questions: had you chosen to play this game? How much time did you intend to spend? Did you continue to recover points after already wanting to stop?
You can decide an end time before starting, choose a study theme or alternate with another activity. These organise your time; they are not a dopamine-reset protocol.
Keep the option of appreciating a session without a final win. Understanding why a variation fails can be a useful achievement. You do not have to solve one last easy puzzle to “positively condition” your brain.
If playing regularly interferes with obligations or causes distress, seek support for the situation you are experiencing. A supposed neurochemical explanation does not replace that conversation.
Further reading: exhaustion in chess practice and tournament stress.
Rebuild the prediction before interpreting the surprise
Suppose a coach offers a puzzle with a visible reward: a point on the club score sheet if you find the winning move. On Monday you have never seen the pattern. On Friday you recognise it as soon as the board appears. You may earn exactly one point on both occasions. In ordinary language it is easy to call both events equally rewarding because they add the same number to the score. As learning events, however, they may differ enormously. Finding a continuation you did not expect can change your prediction of what you can solve; confirming a highly familiar answer may leave that prediction almost unchanged.
The work of Schultz, Dayan and Montague connects neural observations with models in which the discrepancy between an expected and an obtained outcome matters. The experiments discussed in that research do not involve a player winning a pawn or seeing an Elo rating update. Keep that distinction in view when using the idea at a chess club. We can use a prediction model to ask clear questions about our exercises. We have not measured a particular player's dopamine by observing the expression on their face, and cannot assign a chemical dose to a particular move.
Here is a deliberately simple calculation. Imagine that before receiving a result a learner assigns an expected value of four tenths to this puzzle. Give success a value of one and failure a value of zero. The prediction error on success would then be six tenths; on failure it would be minus four tenths. If our invented rule adds one fifth of the error to the old estimate, success moves the estimate from 0.40 to 0.52, while failure moves it to 0.32. The update has a direction as well as a size. The figures are values in a teaching model, not amounts of a neurotransmitter or ratings of how enjoyable Monday felt.
The model also throws away things a coach cares about. A learner might lose the point after finding the clever first move but overlooking a quiet defence. Another might guess the first move and collect the point without seeing the defence at all. If the update uses only zero or one, it cannot distinguish those cases. Ask the player what they expected the opponent to do. That answer can change the next exercise even when the score sheet records the same result.
Wanting another game and enjoying this one
Consider a player who has lost three blitz games and wants to regain the points before shutting the laptop. Ask two questions separately. Did the player enjoy the last game? What made the option to start another game attractive? They may dislike the experience while wanting to restore a number on the screen. Someone else might queue for exactly the same fourth game because a friend is online or because they promised to practise a particular opening. The decision to click again is observable; the explanation needs more information. The research below on the distinction between wanting and liking provides a reason to ask both questions, not a way to diagnose a player at a distance.
Try a session with a different ending condition. Before starting, write down a chess question such as “can I recognise when the defender has an intermediate check?” After a few games, stop at the first position that raises that question. Save the diagram and give yourself one attempt at the defence. You can complete this session after a loss if you have learned what to look for, or discover after a win that your original question remains open. This changes a feature of the activity that the player can actually control: what qualifies as worthwhile feedback.
Two cards that reward a better explanation
Make a pair of study cards. On the first, show a clean fork that wins material. On the second, show a similar-looking fork that fails because the defender has an intermediate check. Tell the solver that there may or may not be a win, and ask them to commit to a line before turning over the card. Give credit for correctly rejecting the fork as well as for finding it. A puzzle collection that promises a winning combination on every page can accidentally teach solvers to continue hunting for a sacrifice long after the strongest defence has answered it.
Give a second partner the cards in reverse order. Someone who just saw the defensive trick may hunt for one even in the straightforward position. The diagrams have stayed exactly the same while the solver's expectations may have changed. Write down what each partner anticipated and what they noticed after checking. Surprise in the losing line can be informative even when nobody received a point. The lesson to take away is a question about move order and checking replies, something you can verify by moving pieces on a board.
To see whether the session taught that question, prepare a third card with different pieces and a completely different board outline. Keep only the relevant choice: a tempting material gain and an opposing reply that needs to be examined first. If the player asks about that reply without being prompted and can show why it matters, you have evidence of learning on a new position. If they only recognise the drawing on the first two cards, you have evidence of familiarity. Either result is more useful to a teacher than an unmeasured claim that one diagram produced a larger dopamine spike.
A pleasant side effect of the exercise is that it gives you a good stopping point. Put the new diagram and your unanswered question in tomorrow's pile. The session has produced a piece of analysis that will still exist after the points on the screen have faded from memory. Whether to play another game can now be a separate decision, based on the time you set aside and what you actually want to do.
You can even change the scoring rule on the last card. Give a point for naming the opponent's best defence before naming an attacking move. Record whether the solver's questions change. The point, the written line and the decision to continue are all visible. A supposed neurotransmitter level would add a new claim requiring a new measurement.
Sources
- Schultz, W., Dayan, P., & Montague, P. R. (1997). A Neural Substrate of Prediction and Reward. Science, 275, 1593–1599. Article.
- Berridge, K. C., & Robinson, T. E. (1998). What is the role of dopamine in reward: hedonic impact, reward learning, or incentive salience? Brain Research Reviews, 28, 309–369. Article.
- Kounios, J., & Beeman, M. (2014). The Cognitive Neuroscience of Insight. Annual Review of Psychology, 65, 71–93. Review.
Comments