The reward that rewards nothing: seven ordinary misses

You reward, and still nothing sticks. The problem is rarely generosity: it is almost always what got rewarded, precisely.

1. Too late, and it is no longer the same behaviour

Between the behaviour and the reward, a dog does plenty of things: it turns its head, steps forward, looks away. What gets paid is whatever has just happened.

Digging in your pocket, opening a tin, bending down: all that time counts. The reward lands on something other than what you were aiming at.

A short word said at the right instant, with the treat following a few seconds later, beats a fast treat that arrives at the wrong moment.

2. The treat sometimes pays for the wrong second

The commonest case: the dog sits, your hand comes closer, it stands up, it gets the treat. You have just taught sit then stand, and that is what it will offer next time.

The same mechanism explains dozens of odd behaviours people think they installed badly. They were installed very cleanly, on the wrong snapshot.

Watching what the dog is doing at the exact moment it receives, rather than what it was doing just before, finds the error in one session.

3. It decides the value, not you

A piece of its ordinary kibble may be plenty in the kitchen. The same piece out in the street, surrounded by smells, is worth nothing.

That is not greed, it is a scale. A reward only has value compared with whatever else is going on at that moment.

Which means a hard piece of learning is paid for dearly, and an easy exercise does not need to be. Using the same thing everywhere is the simplest way to wear yourself out for nothing.

4. Always the same thing, always at the same point, wears thin

A behaviour paid every single time for months becomes a predictable transaction. The day the transaction stops, there is no reason left to carry on.

A behaviour paid less predictably, once it is established, holds up far better. It is the same mechanism that makes certain human habits hard to drop.

The order matters: pay every time while it is being learned, spread it out afterwards. Spreading it out too early amounts to never having taught it.

5. A reward that is not one for the dog

A pat on the head, a good boy called from across the room, a slap on the flank: many dogs get nothing out of that, and some move away from it.

The answer is observable. After what you just did, does the behaviour happen more often, or not? If it does not increase, it was not a reward.

That is the useful definition, and it is an unpleasant one, because it takes no account whatsoever of your intentions.

6. You also pay for what you would like to see disappear

A dog that barks and receives a glance, a word, an outstretched hand has just obtained something. For a bored animal, an irritated response is still a response.

That is why some behaviours grow despite the telling off, and sometimes because of it.

The question to ask in front of an annoying behaviour is not how do I stop this, but what is it getting. The answer is usually in the room, and usually it is you.

7. A session that runs too long undoes what it built

A dog that has worked well five times in a row will not work better on the sixth. Tiredness arrives, success drops, and the session ends on a failure.

It is the last impression that stays closest to hand. Stopping just before things slide is not taking the easy road, it is method.

Several very short sessions across the day beat one session that runs until it goes wrong.