Only responds with treats

 Willy is my challenge dog.  He is too smart for his own good, a real problem solver.  You can actually see his head move as he checks my hand before he decides to follow a command.  "No treat?  No work, sorry Mum."

My trainer said that "I" trained him this way, ha haaa. Of course, it is always the handler's fault!  I'd give a command, he ignored it.  Gave it again, ignored again.  Went and got a treat, he'd see me go to the fridge, and then yes, he paid attention. OR, when he kept ignoring me, I'd take the treat out of my pocket and THEN he'd pay attention.

So the training has to back up a bit.  Put the dog on a leash, so s/he can't walk away.  Then WAIT HER OUT.  Don't move your hand for the treat in your pocket.    We've moved too fast, before the dog was ready, to giving commands when the dog is not contained, but has access to lots of distractions - the rest of the house, the rest of the yard.  Back it up, start at the beginning and rework that.

Willy?  He has me trained REALLY well -- I don't even try to give him a command without a treat in hand!

My other dogs are NOT like this.  

Jewlieee

You have to randomize the treat giving. Sometimes give her one sometimes not.

 You work with random treats once the dog reliably had the command.  That is, 90% of the time you get the action.  Then you start fading the treat and working to praise.

 I was taught that recall always, always ALWAYS gets a treat.  You never fade this out to praise.  BECAUSE . . . the recall could save the dog's life.  You need the dog to have 100% perfection on this each and every time.  

Dog chasing a squirrel who runs across a street, and you can see a car is coming.

Dog chasing deer across train tracks and a train is coming -- someone at the dog school I use actually lost her 6 month old puppy to the train when this happened.  She had not perfected the recall, but was allowed off leash out in the woods.  Horrible tragedy.

 When I started training Ribbon I would hold treats in my hand and have my pockets full for back-up. I soon realized that she sat, stayed etc., not because I told her to but because she knew she was about to get a treat. Right action, wrong motivation. What I found helped to take her attention away from the reward and back on the command was I emptied my hand and my pockets and put the treats in a cupboard - out of sight. When Ribbon follows my command, I give her head a rub, say 'good girl' and then walk over to the cupboard and give her a treat. For her, the time between doing the command and actually receiving the treat  seems to help a lot.

Another challenge with reward training is making sure the dog understands what actions are being rewarded. Ribbon likes to bark at other dogs when she is in the yard. I tried to train her to 'be quiet' on command. I started out using treats. When Ribbon barked I would say "quiet" and she would come running looking for her treat. It soon became apparent Ribbon thought she was actually getting a treat for barking, not for stopping. I had to be very careful with the timing of my command. Now I try to catch her before she starts to bark to give the 'quiet' command. If she comes running, she gets her treat. If she barks - no treat.

Random reward training means you randomly reward the behavior.  You vary the interval/frequency of rewards or the dog will soon figure out that every other time it performs the behavior, a reward is offered. You want to reward every third time, then every other time, then every second time, then twice in a row, etc. You want to  increase the number of repetitions the dog performs, before the dog gets the reward, until you can have the dog do ten reps with no reward,etc.  Don't set up a pattern.  The dog will soon realize that it never knows when a reward is coming so it will try to increase the probability it earns a reward by complying every time, even when there is no food reward. 

 The other thing that many people fail to do is to truly proof the dog. Use sit/stay for example.  They don't increase the time, distance and degree of difficulty they're requiring of the dog.  Asking for a faster response before rewarding, asking the dog to sit/stay as you gradually increase your distance and adding in distractions (tossing a stick, etc.) will all make for a solid reliable sit/stay.  You use the same type of intermittent reward schedule as you add these variations to sit or whatever command you're working on. 

I find people are so intent on rewarding their dog for any semblance of compliance that they never push for more.  Dogs are incredibly adept at figuring out that though they had been rewarded if they sat for three seconds but now the rewards aren't coming unless they sit for longer and longer periods. Don't be afraid to not reward.  Dogs want to figure out what we're trying to teach them but they will take the lazy way out if we simply toss them treats any time they half way respond to a command.  That is, if we've truly taught the dog the command.  You can't start demanding a dog lie down for five minutes if you're just starting to teach the command.

Understanding how dogs, and in fact most animals, learn is incredibly helpful when training.  My favorite introductory book is Don't Shoot the Dog by Karen Pryor.  I highly recommend it to anyone wanting to train their dog, spouse or goldfish. :)

Dogs always train us and I'm trained really well now. :)