Years ago, I majored in psychology and history at university. A course I remember very well was Learning theory. One of the experiments I recalled involved a Y Maze and laboratory rats. The purpose of the experience was to understand decision-making and learning patterns; I wanted to actualize how many attempts a rat had to make to learn the location of the cheese and make an informed decision. I brought a Y maze and a family of laboratory rats to my laboratory. Note: A y Maze, for those that are not familiar, is a Maze configured in the shape of a Y. I placed a rat at the opening of the Y maze (see diagram below). At first, I placed the cheese on the left and then to the right side of the Maze. The rats made several tries in error by going to the right side back to intersecting finally discovered the location of the cheese on the left side. Once the location was established the little laboratory rats instinctively went straight to the location of the cheese without hesitation. Great! I thought learning had taken place.
Reward
LeftRight.
Once and the rat learned the new location of the cheese. Again, new learning took place and superseded the old learning for establishing a decision—two thumbs up for success.
Then there was an epiphany! "A what-if moment, if you will"; what if I randomly changed the cheese's location? I proposed, i.e., right, left, and back again several times at random times and at different intervals.
Total confusion and chaos emerge; instead of making errors smelling the air and the established patterns as before, the rats suddenly stop at the fork of the Maze. After making frustrated attempts, they begin gyrating like spinning tops and collapse from exhaustion a few minutes later.
Take away:
- There is nothing like time and experience in any decisions.
- Trial and error are important components in the decision process.
- In any decision, know the goal and avoid being confused by the noise.
- Changes appear unexpectedly and switch from one side to the other.
- Establish historical patterns as a guide for a decision.
- Anticipate where the reward is moving not where it has been!
- I