If, like me, you have ever found yourself standing in front of a slot machine, losing game after game but still continuing to feed in your money in the hope that Lady Luck is smiling favourably upon you and that on your next go, the machine is sure to pay out - just like it did 10 minutes ago - you'll understand what I mean by the gambling effect. Because we won something once, we are prepared to continue our efforts and put up with getting nothing, time after time, for the chance of winning again - maybe even hitting the jackpot.
If after a while we haven't had a win, we might decide to cut our losses and give up. The chances are though that the next time we and the slot machine are together in the same room - even if this happens a few weeks later - we'll be flash with our cash and play it again because it's our wins that we remember, not all the countless games that we lost. This is because when we have a win, a big release of the neuro-hormone, dopamine, floods our brain. Dopamine is the stuff of desire, of want and addiction, and it's what keeps us feeding that slot machine in the hope that it'll pay out again soon ... and so long as we get an occasional payout, we'll keep putting our efforts into winning even though overall, we have probably lost more money than we have gained.
We can apply the principle of 'the gambling effect' with our dogs too. I don't mean that we can train them to play the slot machines for us while we sit and chill with a beer and then collect our winnings at the end of the evening ... although that would be pretty cool ... no, I mean that when our dogs receive something really rewarding, the same thing happens in their brain as it does in ours - dopamine is released, and what the dog was doing immediately before the reward is likely to be repeated in order for that reward to occur again. The trouble with using rewards such as food tit-bits in training is that the value of the reward decreases if it occurs every time - or rather the 'desire' for the reward lessens. The desire for the reward hinges on its unpredictability, and once a reward has become predictable its value goes down. In the presence of other rewards that don't occur very often, that once high-value reward is next to worthless, and no longer worth the effort to try and obtain unless there is nothing better on offer. At the other end of the scale from this, efforts made to obtain a previously once-gotten reward that consistently end in failure result in the 'why bother effect', and effort ceases completely.
Using the gambling effect to keep the value of a reward high, and so the chances of the effort required to obtain that reward high also, involves making the occurrence of the reward unpredictable. Not giving the reward every time our dogs do as we ask is difficult. We want to reward them because we want them to repeat that behaviour the next time we ask for it, however, what happens then is that we, not our dogs, fall under the influence of the gambling effect, and it is we who end up holding out for the desired behaviour, rather than our dogs holding out for the reward.
Once we are sure that our dogs have learned a new behaviour - which usually happens after only a few repetitions, sometimes even upon the first go if our timing is spot on - we need that behaviour to become a habit and so for a while, we need to reward it every time it occurs. Once the behaviour has become a habit then we need to start making the reward unpredictable, which means not paying out every time our dogs put the effort in. We can still give a cheery 'good dog!' if it makes us feel better, but the reward that we have used to train the behaviour in the first place needs to be saved for another time. So long as that reward does happen from time to time in the near rather than the distant future when the behaviour is repeated, and occasionally the reward is even better than expected, it will remain desirable and therefore worth the effort of our dogs doing what it takes in order to obtain it.
