Rewards in Training
Hey all, hope everyone's doing well. 
---
So, I'm pretty sure we all know that, in training, when a dog performs a desired behavior [ sit, stay, leave it, touch, etc. ] we offer up some sort of reward.
My questions are, which reward do you use? Have you tried other forms of reward? Which has worked best?
Rewards, in this sense, are praise, toys, the pup's food, special treats, anything else you might use as a reward for a dog doing desired behavior.
Just curious ~
---
So, I'm pretty sure we all know that, in training, when a dog performs a desired behavior [ sit, stay, leave it, touch, etc. ] we offer up some sort of reward.
My questions are, which reward do you use? Have you tried other forms of reward? Which has worked best?
Rewards, in this sense, are praise, toys, the pup's food, special treats, anything else you might use as a reward for a dog doing desired behavior.
Just curious
~
Comments
PS - promise more Keiko pics coming soon ;o)
and it turns out Hanzo is partially clicker trained already, he just had to adjust to me but he's already better than Tsuki & Kitsune.
Occasionally, all it takes for Tsuki to do something is to know a good massage is waiting for her afterward. We have this routine when I sit on the couch, she'll sit, wait to be invited, lay on command and roll over on command and then the massage. She loves it!
I haven't yet met a shiba who will do something just for a verbal praise type of reward, most are completely focused on the food rewards!
Sake just learned Sit, she is doing so well. She will do ANYTHING for food I think, she is a little chow hound.
For now, I use sit & stay heavily with NILIF and he's responding very well to verbal cues and body language (he already sits when I snap my fingers). I haven't done any training of new stuff yet as we're still in the bonding/developing trust stage.
I know that it still isnt guaranteed that Honey will COME whenever I call, and I shouldnt bank on her doing so.....but I'm very happy at how well she does.
So it could go like;
Action, Click, Treat. Action, Click, Treat. Action, Click, Treat. Action, Click, Treat. Action, Click, Treat. Action, Click, Treat.
Action, Click, Treat. Action, Click, Treat. Action, Click, Praise. Action, Click, Treat. Action, Click, Treat. Action, Click, Treat.
Action, Click, Treat. Action, Click, Praise. Action, Click, Treat. Action, Click, Praise. Action, Click, Treat. Action, Click, Treat.
Action, Click, Treat. Action, Click, Praise. Action, Click, Praise. Action, Click, Treat. Action, Click, Praise. Action, Click, Treat.
Action, Click, Praise. Action, Click, Treat. Action, Click, Praise. Action, Click, Praise. Action, Click, Praise. Action, Click, Praise.
I think also as the dog continues to learn the behavior you start randomly offering praise instead of treats. If that all made sense...:o ~
The key to using food rewards effectively is to start off consistently giving them and then, after a week or so move to an intermittent reward schedule. If your dog knows that he/she will get the reward every time, they can loose motivation. By keeping them guessing and occasionally (and eventually frequently) skipping the food reward, you keep them engaged. This phenomenon is well documented in the human psychology literature. For example, with email, if somebody knows that they always have email waiting for them when they check, they will check it less frequently than someone who only gets occasional emails.
So, Mike, the moral of the story is, treats are your friend if you use the properly...especially with a Shiba!
If you are alone and just doing obedience training, any type of treats will do - if you are working with a dog on fear issues or training in a place with a lot of distractions, you will definately need something worth getting the dogs attention and worth focusing for.
I usually switch it up a lot too so they don't get bored of a certain treat - I will use really smelly foods like Tuna or Sardines so my dogs get the scent and can focus on me in a place with lots of distractions. For places with not as many distractions...like if we're going for a walk and I may need to re-direct...I bring cheese or boiled chicken pieces. I have also made mini bite size meatballs that my dogs go nuts for.
I really only use treats when we are out in public though - when we do training at home, its usually just vocal and physical praise.
But we got to the point of "random intermittent reward". That means I we only reward him every-so-often, to keep him expecting it. Going form rewarding every time to random intermittent takes a while, and is very progressive. Kuma responds pretty well to it, sometimes he even sits if I use the word in the middle of normal conversation.
Lindsay is correct in that you want your reward pieces to be very small. To a dog, size doesn't matter, but quantity does. They are just like children. If you ask a young child if they want a quarter or 25 pennies, they invariably respond that they want 25 pennies. I also condition my dogs to the word "yes" which acts like the "click" and means food is coming and that they did the correct behavior. As you train, it is easier to say "yes" at just the right moment than try to get your treat out just in time. During the initial stages, you may have to treat very regularly. But then it is best to move to the variable reinforcement pattern (ie, intermittent random pattern) of treating. Multiple researchers have shown (in a large variety of animals) that in training, animals "work harder" if they never know when their paycheck is going to come (ie, they get a treat after 2 attempts, than 5 attempts, than 1 attempt, than 7 attempts). Make sure to vary you pattern as it is known that if you always treat on the same pattern, the animal will learn it and "slack" off during the periods between treats.
Finally, as you shape the dog's behavior, remember that JACKPOTTING can be very helpful. For instance, if a dog suddenly has a "light bulb" moment and dramatically improves from 1 attempt to the next, you give the dog a jackpot (ie, multiple small treats in very quick succession, one right after the other). This helps cement the behavior in their brain so that they keep trying hard AND understand that they did what you want and got the behavior close to the finished product.
I just did a very interesting training session with my Kuma. My parents sent me this contraption that their neighbors made for their dog. It has a hinged lever/button that the dog presses with its paw which causes a bell to ring. It is to let you know that the dog wants to go outside. Well, as I suspected, my Kuma is shy about things like this and was nervous about the contraption (let alone the bell!).
I disabled the bell for the 1st 1/2 of the training. As I said, Kuma doesn't have a good food drive. So I had to have very yummy treats and at first coax him just to come near and sniff the contraption. Then I had to begin shaping him to raise his paw and come close to the contraption with it (LOTS of treats). Then to touch anywhere on the contraptions with his paw). Then to touch closer to the lever/button. Finally to touch the button and eventually to touch it hard enough to depress the lever (remember the bell is disengaged). We did this with 2-3 short training sessions over about a week. Once I got him to the point where he would reliably depress the lever on the command "touch", I re-engaged the bell. Well, he was terrified and jumped back about a yard. So I had to go back to treating very frequently and doing the same as initially, getting him to come close and conditioning him to hear the bell without fear. Another week went by.
Well, I am happy to say that Kuma has fully learned the new trick and when I say "touch" he will march right up to the contraption, press the lever, is unafraid of the bell ringing and will run to the kitchen door to go out! Success! With a lot of patience and treats!! :-)
I have the luxury of owning dogs with no known food allergies or intolerances.
string cheese / kraft singles
grilled chicken bits
hot dog bits
freeze dried liver
soft salmon chews
cheese tortellini
roast beef
deli meats
etc.
On our walks I usually pocket a few chicken jerkies from Dogswell. They are easy to tear into small bits and they don't make a mess in my pockets. They are perfect for intermittent rewards for "watch-me", not pulling, and sitting and ignoring other dogs. I also use them to reinforce recalls for whichever critter is on the flexi-lead.
Triepak is the most food motivated; Guska is the least.
For agility training classes I bring out the special toys too. There is the raccoon tail tug and the bunny-fur tug. These are highly prized toys and are only available during training sessions. (These are favorites of Guska's)
Both can be purchased by Clean Run:
http://www.cleanrun.com/index.cfm?fuseaction=category.display&category_ID=74&CFID=2656466&CFTOKEN=39745998
I use vocal praise too, but I do my best to not babble at them. I think some people talk to much and the dog tunes them out. Tuli tends to be excitable/anxious and there is a fine line between getting her revved and sending her right over the edge (zoomie lalala land)
Sometimes their reward is to run. Especially Guska who has a bit of difficulty with self control. If he demonstrates good attention or a good long down stay, I will often just let him cruise (in a safe contained area).
As far as the treat vs no treat debate: I use treats. The dogs are 7 years old and I still use treats/toys. For some reason people get hung up on this. They have an idea that the dog should do something simply because they like us! If they don't perform then they are lazy, stupid, stubborn. Well, I work for a paycheck, why shouldn't they want one? I would pose the idea that the smartest dogs are NOT the ones that do something over and over and over again, simply because they were asked. The Spitz brain is very different than a German Shepherd, Border Collie, or Golden Retriever brain. They were not designed to constantly look to a handler for instruction.
For rudimentary skills, treats are infrequent and intermittent. For new skills, treats are administered reliably and frequently.