User:Lindsay Ridgeway/Reward-based Field Training for Retrievers/Basic Retriever Skills

Basic Retriever Skills

Sit

Field retrievers need to be able to sit on cue in a number of different situations:

While waiting in a blind
After coming to heel on either side at the line, for a marked or blind retrieve
When delivering the bird after a retrieve
During handling, before a "back" or "over" cue
In numerous non-fieldwork situations, such as when arriving at the curb before crossing a street

For the experienced dog, in many of those situations, the context acts as the cue, and the only reason to have a verbal cue is during the training phase. For example, it may not be necessary to verbally cue "sit" during delivery of the bird.

On the other hand, the whistle sit is needed whenever the dog is handled.

Training the Behavior

Many people train "sit" by either forcing the dog into a sit position (however gently), or by luring the dog into a sit with a treat or hand target.

It is unnecessary to train "sit" in either of those ways, and both create additional associations within the learning process that are unrelated to the desired behavior and that the dog will eventually need to unlearn. For example, both of those methods require the trainer to be in close proximity to the dog, whereas field dogs need to learn to sit when cued even from hundreds of yards away.

Here's a way to train "sit" called capturing the behavior:

Take the dog, a clicker and a handful of treats into a small, quiet room with no distractions. A small bathroom may serve the purpose.
To get the dog into the training game, click and treat three times. No behavior is required by the dog to earn these treats.
Sit down and wait. Eventually, the dog will also sit down.
At the very instant that the dog puts his butt on the floor, click and treat.
Use your hand or another treat to lure the dog out of the sit, then repeat.
Once the dog begins to quickly sit again as soon as you've distracted him out of the sit, you can begin clicking and then using that treat to get him out of the sit. This will fill him up more slowly and give you more productive training time.
After the dog is reliably sitting down repeatedly, say "sit" just as he is sitting down several times.
When you can anticipate the moment before he is about to sit, say "sit" an instant before the sit a few times.
Soon, you will see from the dog's alertness that he is responding to your verbal cue. When you see that, try saying "sit" at a moment when the dog was not about to sit anyway. If he sits, the behavior is now on cue. If not, return to the earlier steps a bit longer.
If the dog is sometimes not looking at you when he sits, stop clicking and treating any sits when he is not looking at you. You want to "sit" to be an attention cue as well as having the dog in a sit position.

It is not necessary that all of the above steps occur in a single session. Take as many short, fun sessions as necessary to get "sit" on cue.

Notice that you are not requiring the dog to face you, or sit in heel position, or be near you when he sits. Although any of those criteria might be desirable for some performance dogs, it is advantageous for the retriever to be able to sit on cue in any position and at any distance. Be sure to add enthusiastic praise when he sits at some distance from you.

"Sit" as Muscular Impulse

An interesting phenomenon is that "sit" has probably become a particular muscular impulse for your dog, not a position. As a result, you may find that the dog will not understand "sit" if you cue it when the dog is lying down or in any other position besides standing.

It is generally not necessary for a field dog to respond to "sit" from any other position besides standing, but if you wish for the dog to be able to respond to "sit" from those positions, realize that you will have to train them separately. Some trainers use a different cue, so that "sit" continues to be the name of a muscular movement and can become essentially reflexive without any conscious thought from the dog.

Eliminating Latency

Latency is the delay after a cue is given before the dog responds. Many trainers feel that when a dog is learning a new behavior, he may need time to think about the cue, and the trainer watches for that "lightbulb moment" when the dog seems to figure out what you're asking him to do, that is, what will earn him reinforcement.

However, watching for the lightbulb moment may lead you into a trap. You give the cue and then you wait, and when the dog finally sits down, you celebrate his newfound understanding with special enthusiasm. Even though he eventually does sit down, he has now begun to learn that "sit" does not require an instantaneous movement in order to be reinforced. He may have even learned that the delay increased his reward.

It is a matter of individual style how soon you want to stop clicking and treating for slow responses. One legitimate approach is to never click for more than a split second of latency. You may never get to see a lightbulb moment that way, but you also won't have to extinguish slow latency later. Extinguish means to eliminate a behavior by not reinforcing it.

The key to this is sharp observation. During training, you want a correct response at least 70% of the time you give a cue. If your criteria for a correct response includes split second latency, it behooves you to give cues in such a way that the dog is able to respond correctly at least 70% of the time. If the dog is not doing so, you have raised other criteria, such as distractions or distance, too fast.

Extending the Behavior

A number of extensions of "sit" will be required once the behavior is learned:

Generalizing for location
Distraction proofing
Hold and sit
Whistle sit
Distance sit

Each of these is discussed in the following sections.

Generalizing for Location

After training "sit" in the small, quiet location you used for the initial training, you are likely to discover that the dog does not seem to know what "sit" means in any other location. This is because dogs are brilliant discriminators, often cueing on stimuli that was not intended as part of the cue.

So once "sit" is trained in the first location, retrain it repeatedly in other locations. You may need to follow all the same steps, though it's likely they'll go increasingly quickly each time the behavior is retrained. For example:

Train in the living room.
Train in the kitchen.
Train on the front porch.
Train in the front yard.
Train in the back yard.
Train elsewhere in the neighborhood.
Train at various training fields.

Eventually, your dog will realize that "sit" means the same thing regardless of location.

Distraction Proofing

Just as the dog will initially consider location as one of the stimuli cueing the behavior, he will also take other criteria into account, such as the presense of other people and objects, noise, light, smells, and so forth.

Just as you retrained the dog to sit on cue in new locations, you will also need to repeatedly retrain "sit" as more distractions are added. It is common to think that the dog "knows" the behavior and is inadventently or intentionally ignoring you. A simpler way of thinking about it is that the dog is responding to stimuli, and has not yet learned which stimuli matter and which do not. By retraining "sit" under a series of increasingly distracting situations, the dog can gradually sort out that the only sequence that predicts he will receive reinforcement is whether you cue him to sit and he does so. As simple as that concept may sound, it takes dogs time to learn it, and it's a lesson they have to learn many times before they begin to generalize it for new distractions and other behaviors.

As you are generalizing for location (see above), also begin to add distractions. Eventually, train to the point where the dog responds correctly in even the most challenging conditions, such as around other dogs or with a tennis ball rolling past him.

Whistle Sit

To be edited

Another thing you can do is to get a whistle sit with Deuce facing you. With Deuce at some distance, call him to you. When he gets close, while still facing you, whistle one tweet, then immediately use your current "sit" cue. The instant his butt hits the ground, reinforce. You'll have to decide what reinforcement works best for this. It could be praise, it could be throwing a toy. I like using a click and a treat.

After you've done that a few times, Deuce will anticipate the "sit" cue as soon as you whistle. Of course, reinforce him for sitting on the whistle and not waiting for the verbal or visual cue. Once that seems to be fairly reliable, practice-practice-practice using one tweet on the whistle to cue a sit.

Traditional trainers use force (a heeling stick and/or an e-collar) to train a very fast response to the whistle. Since we're not using force, we need to substitute lots of reps and lots of high-value reinforcement.

Distance Sit

To train the dog sit and face you at any distance, even hundreds of yards away:

To be edited

The way I would approach it is with a barrier of some kind. You don't need to train it with the whistle. You can train it with the verbal cue, then add the whistle later.

So, for example, in a hallway, you'd put a couple of chairs or something else to block the dog from coming through. Then with the dog on one side, and with you, treats, and a clicker on the other side, you'd cue "sit". Initially, you'd be as close to him as ever, so assuming his sit is already reliable, he'd sit and you'd click/treat.

If he has a "stand" cue, use that to get him on his feet, then click/treat again. If not, you could either take a few days to train "stand", or you could hold the treat in such a way that the dog has to stand to get it, or you could toss the treat on the floor behind him so he has to stand to get to it.

After 2-3 of those, you'd step back one step. Because of the barrier, he can't follow you, so again you'll cue "sit". Hopefully, he'll sit. If not, move closer, do a few more reps, then step back again. Continue in this way until you've got fairly good distance, perhaps 6-8 feet. Each time the dog sits, click and rush forward with a treat.

After that, you have a variety of things to do, in whatever order works best.:

You need to fade the barrier. In other words, switch from two chairs to just one (he could get through if he wanted to but you won't reinforce if he does). Then a lower barrier, then a carpet, and finally no barrier at all.

You need to proof for new locations, such as the front yard and eventually the field.

You need to proof for distractions, such as a bowl of food on the floor, or balloons, or family members wandering around. You'd introduce the distractions first at distance that doesn't bother him, and over a period of one or more sessions, move them closer a little at a time. If he becomes distracted, move them further out again, then start working them closer again.

You need to add the whistle tweet as a "sit" cue. To do that, tweet and then cue "sit" several times. Soon, he'll begin to anticipate the verbal cue when he hears the whistle and you won't need to actually say the verbal cue.

You need to build distance. Eventually he'll need to be able to sit from hundreds of yards away. That's not necessary now, but once you've faded the barrier, start working on greater distances. One way is with the dog inside a tennis court and you outside.

The reason that distance sit is difficult for most dogs is because of how it's trained originally. Typically a dog will need to learn a whole new concept of what the cue means, that it's a muscle movement, not a position near you. Hopefully, though, he'll get it pretty quickly. The more reinforcement he gets for success, the easier it will be for him to repeat.

Coming to Heel

Shaping the Behavior

Prerequisites: "touch" (hand target), "sit", and sit while holding an article in the mouth Note: Laddie sat without dropping the article without that being trained; that possibility needs to be integrated into this section.

Work in a quiet, non-distracting room with a handful of treats and a clicker.

Position yourself with the dog behind you and on the side which is not holding the treats and clicker.

Put your hand at your side with the palm facing backwards and cue "touch".

As soon as the dog touches your hand, cue "sit". "Sit" acts as a reinforcer for "touch".

When the dog sits, click and reinforce with a treat. You can feed in position, then call the dog out of position to set up next rep, or toss the treat on the ground to pull the dog out of the sit and set up the next rep.

Work both sides.

Extending the Behavior

Gradually shape improved position of the sit until the dog is sitting directly beside you, facing in the exact direction he was walking as he came to your hand. This will be valuable later for casting.

When he's sitting in the correct position, you can add a cue word or let your hand movement evolve into a visual cue.

Gradually use the cue to call him to heel from other positions than directly behind you. The eventual goal is for him to be able to come to heel when running toward you from in front.

Combine with steadiness training so that the dog is reinforced for coming to heel by the opportunity to retrieve.

Add carrying a dummy, Dokken or bird while coming to heel ("coming to heel" is a phrase that traditionally includes the sit).

Generalize for location and proof for distractions.

Building Motivation

Here's a game that Lindsay Ridgeway used to motivation for the chain of behaviors that make up a retrieve in training Laddie, his Golden puppy, in August 2007:

Swing the dog to heel and cue "sit". Reinforce with food.
Throw the dummy. Dog waits in sit position, coiled and ready to spring.
Release the dog by calling the dog's name.
As the dog runs back with the dummy, swing him to heel and cue "sit". If he's played the game enough, he'll sit with the dummy in his mouth even though he might not be able to keep the dummy in his mouth while sitting as an isolated behavior.
Take the dummy and reinforce with food.
Again throw the dummy as dog waits for release.

If the dog won't hold the dummy while sitting, take the dummy before cueing "sit", or even before cueing the swing to heel. As the dog's motivation for the game increases over time, you may find it's unnecessary to train the hold as a separate behavior, and that one day you'll cue "sit" while the dummy is in the dog's mouth and he'll just do it.

[More material]

NOTE TO LINDSAY. Merge this with the notes above. Note that the section notes above click/treat for "touch", and uses "sit" to reinforce "touch" from the beginning of the training. That is a preferred approach, but other material in the section below should also be retained.

A field dog comes to heel at the line before a marked or blind retrieve, and again to deliver the bird or dummy after the retrieve.

A prerequisite to this training is the dog responding to a hand target (see the section Hand Target in the chapter "Foundation Behaviors").

To train coming to heel and sitting:

Go to a quiet location with the dog, a clicker and a handful of treats.
Click and treat three times in quick succession to get the dog into the game.
Move into a position so that the dog is behind you off either flank, put your hand at your side as a target, and cue "touch". Click and treat.
Repeat, randomly calling for "touch" on either side of you. If the dog is less responsive on one side than the other, practice that side more until he is equally responsive on both sides.
As that becomes fluent, the verbal cue "touch" will no longer be needed. Click and treat with special praise and enthusiasm for the dog responding to the visual cue alone.
Next, cue "sit" at the moment that the dog touches your hand. Click and treat for the dog sitting.
Continue to repeat that behavior, gradually requiring more and more behavior from your dog:
- Instead of the dog being directly behind you, start with him pointed at your hip, then lead him 90 degrees around and into position.
- Gradually increase the size of the turn until you can bring him around from in front of you.
- Gradually ease the requirement that the dog actually touch your hand, as long as he comes to the desired position.
- Click for rough positioning at first, then gradually raise your standards and don't click if he doesn't meet your current criteria. If he is successful less than 70% of the time, you've raised the bar too high and he's not getting reinforced enough to stay in the game and enjoy it. If he is successful more than 80% of the time, you can increase your criteria a bit.
- Again, practice both sides in unpredictable sequence.
Occasionally, the dog will sit before you have a chance to cue "sit", perhaps because he is not quite in position. If he is not too far out of position, click and treat for that anticipation, so that he will learn to sit without the verbal cue. Then gradually raise the criteria so that you only click and treat if he is in position when he goes into his sit.
Because you are training both sides, you may not want to bother to train separate verbal cues for each side. Instead, you can gradually fade your target hand into a more subtle version, such as a wristy spiral movement, and let that serve as your visual cue for "come to heel and sit".

As with training "sit" above, you'll need to generalize coming to heel for location and retrain it under increasing levels of distractions.

Take your time with this, so that coming to heel presents a facile, joyous picture.

Steadiness

A field dog is said to be steady if he can wait at heel until released without any physical restraint. In some venues, it is also impermissible to use verbal cueing.

The two situations a dog typically needs to be steady in competition is when waiting at the line while marks are thrown, and when honoring while another dog runs marks.

Some methods of steadiness training can be de-motivating for the dog, and some trainers prefer to delay training of steadiness. In fact, in the lower levels of competition, such as Working Certificate (WC) and Junior Hunter (JH), the focus is on natural ability rather than training and the dog is not required to be steady. Typically, dogs competing in those events are restrained at the line with a check cord, and no honoring is required by the event.

But reward-based trainers often work to develop control in drive, that is, a high level of motivation simultaneous with self-control, from the beginning. The goal is to develop intense focus while waiting, and explosive drive to the mark when released.

Training Steadiness

Here is a method used by Lindsay Ridgeway for training Laddie, his male Golden, starting in July 2007 when Laddie was 11 weeks old. At that time, Laddie had been practicing tug, "sit", and retrieving for about four weeks.

The Game

Holding the dog in your arms, put down the dummy at arms length from the starting line and get a cookies (such as a meatball from the dog's meal) and the dog's tug toy.
Put the dog on the ground, cue "sit", and feed for that response.
As soon as the dog has eaten the cookie, cue the retrieve. Traditionally, the dog's name is called out, for example, "Laddie!"
Ideally, the dog runs to the dummy, picks it up, and runs back to you with it.
- If he doesn't leave the line, run with him part way or all the way to the dummy.
- If he runs with you but doesn't pick up the dummy, pick it up and throw it a short ways.
- If he still doesn't pick up the dummy, either this is a bad time to train, or he hasn't had enough work with simple retrieving.
- If he doesn't return to you with the dummy, try running away from him to get him to chase you with the dummy.
- If he still doesn't return to you with the dummy, put him on a long line (perhaps 25') so that you can gently reel him in after he has the dummy in his mouth. Do not let him rehearse not returning to you with the dummy. That can be a self-reinforcing behavior that you do not want to have to untrain later.
When the dog arrives with the dummy, take it from him and instantly reinforce. A great way to do that is to bring out his tug toy and snake it around on the ground to get him playing tug with you. Other possibilities include handing him a bite of food or throwing the dummy out for him to run and retrieve again. See Notes below.
Pick the dog up and cuddle him, then repeat from the beginning.

Extending the Behavior

Once the dog is playing the game reliably, gradually increase the amount of time for the dog to sit before you cue the retrieve.
- If he breaks, as he inevitably will in the early sessions, simply pick up the dummy and reposition him at the line, then again cue "sit".
- It's essential that you not cue the retrieve an instant after he breaks, since that reinforces the break. He must wait until you cue the retrieve, or you pick up the dummy.
Once the dog is reliably waiting for the retrieve cue when the dummy is at arm's length, gradually introduce tossing the dummy while he is waiting.
Once the dog is reliably waiting when you toss the dummy, use someone else to act as your "gunner", throwing the mark for your dog while he remains steady.
Next, assuming the dog has been introduced to gunfire, have the gunner fire a blank pistol before throwing the mark.

Notes

If the dog breaks and it becomes necessary to re-cue "sit", do not feed again. That would reinforce the undesirable behavior chain of breaking, then sitting.
Some trainers do not use food in field training. Try the above method both with and without food a few times and see which gives you the most enthusiastic and reliable results with your dog.
Some trainers take the dog's dummy (if he easily gives it up) and with a shout of "hey-hey", throw it again for him, or throw a second dummy, instead of inviting a game of tug. These simple retrieves are called hey-hey dummies. The advantage of tug is that it offers a high-drive game as the reinforcer for a high-energy return, as does a hey-hey dummy, but it also helps to build the dog's relationship with the trainer, which hey-hey dummies do not do as well. If it improves your dog's motivation, you could play tug, throw the dummy, and play some more tug before doing the next retrieve. Reinforcing the return with food is not preferred, since food tends to calm the dog and we want the retrieve to be high energy.
On the other hand, some dogs will stop delivering the article to hand if you do not use food in the early stages of training. While that is the case, continue to use food, and then gradually begin to substitute tug and hey-hey throws for the food. A time will come when the dog will have little interest in food, and may even comically seem to be insulted by an offer of food. It is not necessary that the dog start with that attitude. It will come.
Carefully watch the number of retrieves you do per session. If you see loss of interest, you've gone too long. Stop immediately, and next time, do a smaller number of reps. Some trainers do no more than three retrieves per session, and might well end the session after a single, especially high-quality retrieve. When the dog is then enthusiastically praised, watered, and crated, he can lie quietly and think about the great job he did on that most recent event in his memory, visualizing it again and again and integrating it into his future behavior. This visualizing process, often used by human athletes as well, can be more productive than a series of increasingly less focused, less motivated repetitions of the intended behavior.

Swimming

Another name for retrievers is 'waterdogs'. Most retrievers will take to swimming, especially for retrieving, easily.

Important The worst thing you can do is to throw the dog into the water. The second worst is to coerce him into the water.

Instead of using physical force or any sort of coercion:

Find some fairly clear water with an easy entry and wade in as far as you dare. Hopefully your dog will follow you out there. If not, quit and try again another time.
Once he is comfortable getting his feet wet, throw a dummy a foot or two from shore for the dog to chase.
If you have already trained "give it", you can cue for him to retrieve to hand. Otherwise, you can wade out to pick up the dummy to throw again.
Very gradually increase the distance you throw the dummy into the water.
Quit each session while the dog is still anxious to play. You want him to be in that state of mind until the next time you bring him to the water.

Note

If your dog did not follow you into the water, it's either because he doesn't tend to follow you anywhere, or because he found the water uncomfortable.

If you don't feel that he's afraid of the water, or you don't have the option of wading in, you could begin with a short retrieve.

If your dog doesn't tend to follow you around, that is a problem and should be addressed. See topics elsewhere such as:

Introducing Gunfire

Introduce loud noises such as pans and cookie sheets
Play this game at home, then in the yard: Train noise as a recall cue: bang (softly at first), then feed; gradually, add distance and volume
Play this game at home, then in the yard: Have "gunner" throw cookie sheet, then throw food on the cookie sheet (simulates running toward the gunner on a marked retrieve)
Introduce gunfire at training and/or events. Key concept: The sequence for dogs new dogs needs to be rest or quiet activity, gunshot, instant opportunity for highly enjoyable stimulus such as treat, tug, chasing the trainer, or a thrown toy. Do not have the gunfire occur while the high-value stimulus is in progress, or reverse conditioning may occur.

(Note: I posted a more complete description to PGD list in December 2007. That might be worth copying here.)

Heeling

Heeling can be trained as the foundation for competitive obedience, rally, and musical freestyle, and possibly other canine sports as well, and a dog skilled in heeling can perform the behavior for dozens or hundreds of paces.

For the competitive field dog, heeling plays a much more limited role, coming to and leaving the starting line. As a result, it can be trained as a simple, short-distance chain of behaviors: sit, heel a few steps, then sit again.

Prerequisites

Training the Behavior

Here are the steps to train a simple, enthusiastic heel suitable for coming to and leaving the starting line:

Start by working in a quiet location in the house, with minimal distractions.
With dog at either side, cue sit (C/T).
Lower the hand on the dog's side so that your palm is facing the dog, step forward and cue "touch", and when the dog touches your hand, cue "sit", then C/T with dog in sit position.
Repeat until dog is performing the single step behavior smoothly and confidently.
Repeat several more times and fade the "sit" cue.
Repeat several more times, replacing "touch" with the cue you wish to use for heeling. If necessary the first few times, say the new cue word and then prompt with touch. Many trainers use "heel" as the cue for heeling, but because it sounds almost identical to "here", some use an alternative cue such as "with me".
You may wish to continue using the hand target with your heeling cue. If you prefer to put you hand in a different position when coming to and leaving the line, such as swinging at your side or resting on your belly, repeat several more times with your hand in the position you would like to use. If necessary the first few times, start with your hand in the new position, then prompt with a hand target.
When all of that is in place for a single step, train it for two steps, then three steps, and so forth.

Although dogs traditionally heeled on the left side, that is, opposite the side that the hunter carried his rifle, and although competitive obedience still requires heeling only on the left, it is advantageous for a competitive field dog to be able to heel on either side.

Extending the Behavior

If the dog's heeling is not consistently performed with enthusiasm, introduce that as a criterion for clicking. For example, you might C/T the most enthusiastic 70% of your reps, and not C/T the least enthusiastic 30%. Continue to raise the bar until all reps are at a high level of enthusiasm.
When the behavior is solid in a quiet location with minimal distractions, train it again in a new location with somewhat increased level of distractions. Continue to generalize for location and proof for distractions until the dog can sit/heel/sit even under competitive conditions.