The Psychology of Quality and More

On Feedback, Delay and Damping

~ David Straker ~

Feedback and work

Consider the following story about a very patient and hard working man and his boss:

What is happening here? Sure, the boss is asking the man to dig the hole and then coming back perhaps later than he should have done, but what are the key factors we can extract from it?

The first point is that the ground level within a defined area is changing state, from the same as the surrounding land to several feet lower. Thus we have a system which is changing state and which we would like to transition to the desired state as quickly and easily as possible. In the above example, this is clearly not happening.

The second point is that there is a feedback control loop for determining the correct depth of the hole. In this case it is the boss.

The third point is that the feedback is delayed. In the above example, this results in the workman overshooting both when he digs the hold and when he partially refills it again. Although the hole eventually gets to the right depth, it takes more work than it should.

The final point is that the workman digs at a certain rate and that in this example, he is digging so fast that by the time the boss has come back, he has over-done it and dug too far down.

Smoothing out the ripples

Overall, then there are three variables: the transition required (depth of the hole), the speed of digging and the time delay on feedback. By playing about with the transition speed and the feedback, we can change the amount of ‘damping’ or smoothing out of the ripples.

Paradoxically, if the boss cannot come back sooner (thus speeding the rate of feedback), then a more efficient system would be for the worker to slow his rate of digging so that there is harmony between his dig rate and the speed of feedback from his boss.

Human transitions

So what does this mean within the context of mental states and meta-states?

Firstly, when someone transitions from one state to another, there is a rate of transition. They might gradually become more angry or snap suddenly into a state of anger.

Secondly, they get feedback that they are in a different state.

Thirdly, as a result of that feedback, they might decide to change or moderate the primary state to some degree.

If the feedback about change happens some time after the transition has occurred, then the person may have gone so deeply into the new state that it is difficult to come out again, have done something that they later regret.

Where does the feedback come from? Internally, it comes from a meta-state that looks down on the primary state and says “Hey, you’ve changed.” Externally, it may come from another person who points out to them, or it may come from noticing reactions from other people (External to Internal). Any of these can result in a delay.

So what can we do?

From an external viewpoint, we can ask other people to tell us that we have changed state (“You’ve gone into that angry state that you told me you are trying to avoid.”). From an internal viewpoint, we can set up a monitoring meta-state to spot the problem state before it gets to be too much of a problem. Then, as the meta-state controls the lower states, we can redirect or otherwise manage the situation.

How closely do we need to monitor? That again depends on the rate of change within the system. If we easily and quickly flip into a damaging problem state, then we need close monitoring. If it is more of a slow fuse, then we simply need to keep a weather eye out for the fizzing of the fuse.

The monitoring can be over-done, a situation equivalent to the boss standing over the worker and commenting after each shovelful, “You’ve not done yet.” If we monitor too closely then we will be spending more energy in the monitoring meta-state when the effort would be serve us better in the primary state.