First, be sure that you are paying attention to the basics, that is, correct timing and being sure that you are not varying your criteria unintentionally. Again, videotaping is very useful. Assuming that you are paying attention to such basics (and, please don't just assume that is a given), your remedial training program should include LOTS of trials. A rule of thumb (that is only partially corroborated in the lab) is that for every one bad reinforcement you must have 10 good ones, While that may not be exactly true (and I don't really know what the figure would be, though it probably would depend on the animal, the behavior, the trainer, and a lot of other factors), my own experience tells me that I want to have a LOT more correct than incorrect reinforcements.
When you are in remedial training mode, it is usually best to shorten the distances in order to reduce the time per trial in order to have MORE TRIALS, thus more quickly overcoming the power of the incorrect reinforcement. There are some other reasons, but the upping the number of trials is the main one. Certainly, one other reason is that the dog can do fewer behaviors that you don't want if the distances are shorter - essentially your control can be better at shorter distances (I did not say will be better, just CAN be better).
I still suggest that you consider short latency in ALL of your training. The more the dog learns that <COMMAND> always means <COMMAND> NOW!, the faster all of the dogs behaviors will begin right away. It is interesting that oftentimes there is a general improvement in overall behavior when latencies are shortened. I would guess that has to do with the dog's focus, but that is simple speculation.
It is NOT a matter of being forgiving. it is a matter of being consistent. If your criterion is that the dog is BEGINNING to do the behavior by the count of onethousandandone, and that if it is not beginning, you make a clear turn and give a short time out, and then try again. Wait until the dog is doing this correctly 8 out of 10 times and then shorten the requirement a bit more. Don't be in too much of a hurry, but don't get the animal in a rut either. That is why we are just after 8 out of 10 times rather than 10 out of 10. Gradually require faster and faster response. Remember too, you must overcome all of the training where latency was NOT a requirement.
I hope you understand my explanation of building short latency. If you plan to CLICK for fast and not click for slow, you will be faced with deciding trial by trial and instant by instant, when to click and when not to. If you do not have an objective measure of time, you will almost surely be varying your timing of the click.
The point here being if the dog responds slowly (both in terms of latency and speed of performance) at short distance, there is little point to extending distance. Right now, forget distance. Get it close in. Then start out extending the distance. When you extend distance, don't just keep extending further and further. There should be some short trials, some medium trials, and some long trials, and some even longer trials, and occasionally, a really long trial. This is sometimes called "bouncing" or bounce trials. The dog should never really know how far out a trial will be when it starts. At first, emphasize shorter runs. Later, longer trials can become more frequent.
Copyright of all posts is the property of the original author. Please obtain permission from the original author before copying, quoting, or forwarding.
List and Site Owner: Melissa Alexander, mca @ clickersolutions.com