Dog Training

How do we increase (boost) reinforcement rate? (4 points)

1) Drop Criteria (or split) 2) Train faster (decrease ITL) 3) Pay Setup behavior 4) Shop (train at the same level and push only when we start seeing the dog make the next criteria)

In what training scenarios are clickers helpful.

Clickers are timing aids, they are helpful for quick fleeting behaviors (such as nose touches) where it is often not possible to deliver R+ before the dog has commenced another behavior. Hence a conditioned reinforcer (clicker) is necessary to mark the behavior.

Short term fixes to motivation issues try to salvage the situation during now (i.e during this training session). List some of these fixes.

DT22 1) upgrade / rotate / novelty of the motivator 2) if there is competition, can we either remove the competition or leave the area 3) Premack, use the competition as a motivator.

Long term fixes try to address long term issues so that this is less likely to occur during future sessions. List some of these fixes.

DT22. 1) Use deprivation (i.e establishing operations ; regulate the economy) and saturation (i.e abolishing operations to make the competition boring). 2) Proof gradually - Control the situation strongly in the beginning so that over time, the dog begins to believe that you control everything.

What are the rules of Tug?

DT28. 1) Use a start cue. I.e invite the dog to start and he only grabs when invited. Otherwise cancel. 2) Use a stop cue: Dog lets go on cue otherwise cancel play 3) Never let a dog put teeth on human. If ever it happens. say "Owww" and cancel playtime. 4) Use reinitiation of game as a reinforcer and cancellation as punisher.

How do we know if a dog's "unwillingness" to perform is a motivation issue (WHY) or if the dog is merely confused by what he should do (What)?

If the dog will take a freebie, then it is probably a WHAT problem. If not, then it's probably a WHY problem.

Define : Split

Insert an extra step that is mid-way in difficulty between two existing steps in a training plan.

List the criteria for PDS rules.

PUSH = 4 or 5 out of 5 STICK = 3 out of 5 DROP = 2 or fewer (drop immediately at 0 for 3 or 1 for 4) Besides 0 for 3 or 1 for 4, make the PDS call at the END of the set rather than trying to manufacture more winners by fudging around within the set.

What 4 things do Bob Bailey stress when it comes to the execution of animal training?

RATE, CRITERIA, TIMING, MECHANICS

How are Reinforcement Rate and Criteria related?

Rate is inversely related to Criteria. The higher the criteria (ie the more difficult the goal is), the lower the reinforcement rate will be.

What is Bob Bailey's mantra which lists the items that a good trainer should obsess with?

Rate, Criteria, Timing, Mechanics

Define : Stick

Remain at the current step in the training plan (do more repetitions)

What does Dr. Susan Friedman have to say about how dogs use behavior?

"Behavior doesn't just flow like a fountain. Behavior is a tool (that) animals use to produce consequences.""

What are the 3 high level steps in Making an OC Training Plan.

1) Define the terminal behavior i.e what EXACTLY a dog will have to do under what conditions. 2) Break the terminal behavior into its parameters 3) Build an incremental plan starting with what the dog can do now and ending with the terminal behavior. Remember to include splits!

List the order of parameters to prioritize during production training. (3 D's)

1) Distraction 2) Distance 3) Duration.

Dog's guess much more than trainers imagine. What are their guesses based upon? (4 items)

1) Aggregate (cumulative ) R+ History. Which behaviors get reinforced the most. 2) Recentness (which gets me R+ lately?) Usually the behavior that was trained most recently 3) Preferred behaviors (self - reinforcement). Sometimes dogs will guess the behaviors that they enjoy the most themselves! 4) Order of events.

When should we split?

- If the current step is too easy ( e.g if the dog is 5 for 5) and the next step is too difficult (e.g if the dog is 0 for 5). - After a Drop - Push - Drop go ahead and split.

What best practices should we employ when clicker charging? (3 things)

1) Click then feed (in that order) 2) Be aware of blocking / overshadowing by pouch reaching, bag crinkling or commencing delivery of R+ to the dog. 3) Vary the time in between click.

What 2 things should we do if we have a dog with a motivation problem

1) Close economy on motivators 2) Saturate competition.

What should we do when a dog keeps guessing? I.e makes a lot of mistakes or just waits, or never performs on the verbal but waits for the signal every time. (4 items).

1) Ensure that we are clearly separating verbal and signal. Verbal - Pause - Signal 2) Help the dog with discretionary sticks until he starts jumping the prompt. 3) Impose a cost: Cue errors with a word like "Too Bad" + no R+. Over time, he learns that "Too Bad" predicts decreased R+. 4) Another cost: If you give the verbal, but he needs the hand signal to do the correction action, impose a cost by praising only but no R+.

What are the usual culprits that slow reinforcement rate? (3 points

1) Not adhering to PDS rules 2) Not training fast 3) Mechanical / technique issues that are slowing us down.

What 4 areas are key to good Mechanics during training?

1) Quiet Body 2)Timing 3)Prompt & Fade 4) R+ Storage and Delivery

What things could cause WHAT problems. I.e failure of the dog to understand what to do next.

1) Under-training. The trainer presumes that the dog actually knows but the dog actually doesn't. It is a collision between the trainer's expectation and the actual strength of the behavior. 2) Failure to generalize. Not only does the dog need to know what to do, but he must know what to do in a variety of situations. A dog that knows how to down-stay only at home is handicapped in his ability to do so when it counts outside the home. 3) Poor inter-cure discrimination.

Addressing Motivation: When giving a command to a dog, it helps to imagine the dog asking these two questions : (note that the order of the questions is important!)

1) Why should i do it? 2) What is the behavior? Note that Motivation (WHY) always comes first. Only if the dog is motivated by the reinforcer will he then proceed to step 2.

"Reinforcement isn't an instant, it's a process" - Bob Bailey. At which 2 opportunities do trainers get to Pay a behavior?

1) With the secondary R+ (e.g. clicker or whatever the dog attends to that tips him off) 2) With delivery of primary R+

Define : Training Plan as it relates to dog training.

A series of steps to train a behavior to train a behavior that gradually increase in difficulty. Starting from what the dog can do right now and ending with the terminal behavior.

Define : Biddable

An old term used to describe how easy a dog was to motivate.

During training, we must convince the dog that we control what he wants. What happens if we cannot control these things?

At the early stages of training, avoid training during times that we don't actually have control (or at least avoid cueing behaviors at those times . If we do this enough times, the dog eventually assumes that we always have control. Later on we can actually get compliance even if we don't have control because he he doesn't test the system as much anymore.

How do we use Toy motivators?

By using fetch and tug games.

What are HYDRAULIC Motivators.

DT 19. Motivators whose motivating power waxes and wanes over time. For example a dog is less likely to work for more food after having a very heavy meal. Whereas a dog which rarely gets a special treat will be extremely motivate to work for it.

What would be described as OPTIMAL CRITERIA for each step of the training plan? How do we quantify it?

DT 39. Optimal criteria keeps the dog interested so he doesn't quit, but not so easy that we don't make progress. We quantify it through the rate of reinforcement: For beginner animals, 10/min (i.e every 6 sec or around 8-12 per min). . For keen dogs or advanced training dogs, it is possible to lower the rate.

What Bob Bailey quote reminds us about the ramifications of CC creeping in when practicing OC.

DT 55 - "Pavlov is on your shoulder" Reminding us to be careful of aversives

What 2 options for a dog that is unable to be lured into a down from sit.

DT 74 1) Plan B (Shaping) 2) Backdoor Sit (Sit from Down)

What are the 2 options for "Stiff" little dogs that won't be lured into a down

DT 74 1) Plan B (Shaping) 2) Tunnel Down

What is the Optimal Spacing between Training Sessions?

DT 93 1) Meyer & Ladewig (2007) found that dogs trained once a week learned faster than those trained daily. 2) But as pet dog trainers, taking pragmatic concerns into consideration. Just train when we can or whatever the owner will actually do.

Do motivators have a hierarchy? If is it possible for one motivator to interfere with the efficacy of another motivator?

DT20. Yes and Yes! If hunger and eating is currently the top motivator for a dog, then another dog trying to hump (i.e have sex with) our dog will interfere with him trying to eat!

If an unmotivated dog will NOT take a freebie, what does that imply?

DT21. It means either 1) We've trained too long, put him away and try again later 2) We've good competition. Use establishing and abolishing operations to deal with the competition. 3) He's tired of the motivator. Try an even stronger or a different class of motivator.

If an unmotivated dog will take a freebie, what does that imply?

DT21. It means that it is either 1) problem with our rate 2) we have a WHAT problem.

Describe how we can exploit Position Feeding for 1) Short term 2) Long term

DT77 1) Short term : we can use Position Feeding to set up the next trial or to encourage the behavior we are trying to build / reinforce. 2) Long term : repetition or patterns may result in biases (For example the dog orienting to the feeding hand or training location). THe trainer can exploit these to build the terminal behavior. "Is there a particular bias we can exploit and build with my R+ delivery that may be helpful to the cause?"

What are the 3 likely causes of a treat motivated dog that gets it wrong during training. i.e a WHAT problem.

DT83 1) Dog has just not gotten enough training regardless of what the owner presumes 2) Failure to generalize. The dog needs more repetition in different environments. (Described by Karen Pryor as NEW TANK SYNDROME) 3) Poor inter-cure discrimination. The dog can't tell the cues apart. They typically perform well when focused on one single behavior, but they get confused when you mix up the regular order or try to add compound behaviors.

List some common parameters

Distractions, Degree and Type of Prompting, Duration

True or False : "The click ends the behavior"

False. After the click, there are still opportunities to exploit position feeding to either reinforce the current behavior or to set up the next behavior. for example if we are training stand from sit, if a sit happy dog sits after standing, re-stand her to pay.

Define : Drop

Go back to the previous step in the training plan. i.e drop criteria.

Define : Training Criteria

It is EXACTLY what the dog needs to do to get paid in painful detail. e.g stay for 10s with trainer 5ft away in living room with no distractions.

What Training Is (As opposed to what it is NOT)

It is a strengthening a behavior. Training is at very best, will alter the probability of behaviors occurring on cue. As trainers, we must realize that we will never ever get 100% correct response from the animal, but we can merely increase the chances that the animal gets it correct.

Define : Push

Move on to the next step in the training plan . i.e raise criteria.

What is the importance of having a quiet body.

Not having any unnecessary extraneous movement by the trainer allows the dog to most readily catch the movements that do matter (i.e let the hand movements stand out from the background), such as prompts, cues and R+ delivery.

What happens if we lose the dog midway through training? How do we know if its a rate problem or just has no motivation?

Offer a freebie. if it is a rate problem, he'll take the freebie. If he doesn't take the freebie, he's not motivated. Either stop training or find another more potent motivator.

When it comes to addressing Motivation, what is a sure sign of an incompetent animal trainer.

One who tries to address "WHAT" (i.e The dog is simply confused about What is the behavior expected) before addressing the "WHY" (i.e , Why dog doesn't want the reinforcer) or gives gobbledygook explanations about motivation. If you don't have control over what an animal wants, you can't train him.

Is motivating an animal during training a Moral Issue or a Technical Competence issue?

Technical Competence Issue. Animals know how to spend their behavioral dollar and they won't waste energy doing something that won't benefit them. Be wary of animal trainers who don't talk frankly about how they motivate the animals and instead go into moral explanations about things like "the desire to please" or "the dog has to RESPECT you as a pack leader".

What training is NOT! (i.e. misleading philosophies about animal training)

The following are misleading thoughts that we can have about training. 1) "He KNOWS what to do" 2) "When will he REALIZE?!" 3) "The dog must be made to UNDERSTAND" 4) "I think he's GOT IT" These become painfully obviously wrong when we think that the dog has accomplished the above but yet still give wrong responses after he allegedly "UNDERSTANDS"

What does it mean if we find ourselves doing discretionary sticks a lot?

There is a mis-match between the training plan and our comfort level. Change the training plan to suit our comfort level.