So now that we know how to make a good training plan and why that is important, we also need to know how we can make the best of it.
People often do not remember or care to keep a record of their training. This is interesting since to make progress, you should be aware all the time of the success rate in training. How can we otherwise know whether we should keep up the good work or change something radically. If the dog does not meet the criteria we have set in a couple of training sessions, we should check our training settings and criteria and make changes on our exercise. A good rule of thumb is that 80 per cent or 2/3 of the repetitions should be successful before the criterion can be raised.
Usually we trainers and handlers tend to have expectations on our training and really want the training to be successful. When focusing these successes, we often do not notice the small mistakes there but value the good parts of the training. We might end up praising an average performance. On the other hand, if we expect the dog to fail, we may judge it based on mistakes and ignore its steps forward. This is why everything should be written down when training.
Well – how is it possible to keep your eyes on the dog, possibly keep the reward/s in your other hand and the leash in the other? Maybe you also have a clicker. How do you manage to mark down all the data?
There are several ways to keep record. The easiest way is to have someone to watch you train and mark down every successful and unsuccessful repetition. Having an assistant also often helps you find out the possible mistakes you do during the training that you yourself are totally unaware of. Even more efficient it is if you take a video of your training session, watch it through in between every set and write down the repetitions yourself. Then you also see how the exercise really went and can make the needed adjustments for the next round.
But if you do not have anyone to help and setting up a camera is inconvenient, you can decide a number of the repetitions you will do in that session beforehand. Sets of 3-5 repetitions are the kind you can still usually remember during the training and then you can write them down straight after.
One way is to set a time limit and count only those repetitions that meet the criteria. Depending on what you are training, and how long does one repetition last, you may want to set a limit where you assume the dog has enough time to do certain amount of repetitions and count how many of them needs to be successful ones. Let’s say you work on “sitting on a heel position on a command”. You set the time to 30 seconds and evaluate that a dog is capable of doing 10 repetitions in that time including the time it takes from you to deliver the reward and reset the exercise. If you have 30 seconds and 10 repetitions, 80 per cent of that is 8 repetitions. Now you can take 10 pieces of food in your hand or treat pouch, start the timer and start the exercise. Stop the exercise when the time is up and count the rewards you have left. If there are 2 (20%) or less pieces, you have achieved the needed success rate to move to the next criteria. If all of the pieces are gone before the time is up, you should adjust the time and the amount of repetition for the next round.
“Using the amount of treats to count the amount of succeeded repetitions.”
If you do detection training, where the dog searches a specific area, the number of repetitions is not the issue. Then you can set a time limit or the number of efforts the dog has and write down how many hits and controls the dog alerts in that time. After the time or the amount of tries is up, you count the hits. If you have succeeded in 80 per cent of the hits in given time, you can change the criteria again.
The biggest benefit on collecting data is keeping on track of how your training is proceeding and the possibility to go back and review the whole process. You can analyse what were those harder exercises that you should have cut in smaller pieces and what criteria you could have increased faster.
When you work with scent detection, and especially if you train working dogs, the well collected data is part of back up, if anyone ever doubts your work or the way your dogs work. Collecting all the important facts like the amount of scent, used distraction scents, the blind and double-blind exercises and so on, will help you show the skills your dogs.
Do you have good examples of training logs? Feel free to share in comments!
How many times have you trained a dog and got frustrated when it fails over and over again? In how many of those times did you have a written training plan that you also followed?
We usually go to the training field or training facilities and have an idea of something we want to train. Maybe something that did not go very well last time or something we want the dog succeed in. We then decide the theme of the training and start working with the dog. This happens although we know how important it would be to plan something beforehand. Still not many of us writes the training plan down, or analyses those details that need to be taken into account during the training session.
I was visiting SWDI again this January for a 7 days. We had a detection work shop with one theme for the whole week: training protocols. We were given a protocol for detection work that we were supposed to follow for that week. Everyone read it through first time in the lecture room. Very clear and coherent plan with well explained, detailed training steps. Before we took any of the dogs in the training facilities we read the first step over again and set up the needed equipment for the first round.
When we arrived in with the dogs, one at a time, the set up was ready and the dog could be sent to work immediately after entering the room. Assisting handler wrote down all the successful repetitions, possible missed hides, false alarms and the number of resets. Every dog did about 3-6 repetitions before they were taken out for a break again.
The time we had for the training was used effectively, because every time a dog was present, the handler knew exactly what she or he was supposed to do and what was expected from the dog. If the dog failed 2 times in a row, the trainer had to think what should be changed and then reset the exercise. If the dog succeeded 3 times in a row, the trainer could move on.
This method worked with all of the dogs with different backgrounds and learning history. When we proceeded step by step following the plan we could see huge progress with every dog by the end of the week.
What should you take into account when making your own training plan, then?
First you need to decide what your goal is and write it down with all substantive details. For example, if you are training a “sit” command, you need to describe what is the situation that you aim for, what is the environment, duration, distance and the position and station of the dog. If you were starting training “sit” for agility competition, where the dog needs to stay put in the start and wait for the command to start running, the goal could be something like “Dog needs to sit for at least 10 seconds in the start, front paws behind the line, facing the course while the handler walks away from the dog, closer to the first obstacle. Dog may leave only when it hears the command from its own handler.”
When you have your goal, you divide the training in to sub goals and important elements.
At SWDI detection work shop we had a task to make a progression plan for remote scent detection.
The idea was that the handler should stand outside the doorway and send the dog to the line ups two times. You flipped a coin to see which one. There could be one target scent in the line up or not. If there was no scent, you needed to call the dog back. You would get minus points if the dog went for the wrong line up or did a false indication.
To train our dogs for this, we made a progression plan. It looked like this:
Training step 2.4 (Handler standing in the doorway. Send the dog left/right on line ups placed in a 45 degree angle from the door 2,5m to the first object. Dog on a leash.) of remote scent detection on line ups. (Video by M.K.)
We had one only day to complete the whole task: planning, training and competing against other teams. We had three dogs competing in our team. We used about half an hour to write down the whole plan and had the rest of the day for well planned training.
In the next video you can see one of our teams dogs competing. The right side line up is just down from the camera. Dog is sent to the left one. No hits on this run.
Making this kind of training program and following it systematically, you will get stronger behaviours and results much sooner than just training and modifying the training plan while you are working on it. Without a plan there is a big chance you either raise the criteria too soon and the dog fails because of it, or you stay on the safe ground where the dog succeeds but does not proceed.
Do you have good training plans that have shown to be effective in your scent work? Feel free to share them n comments.
“A goal without a plan is just a wish.” – Anonymous
Thank you again for a spectacular training week that again gave lots of new ideas
Jessica, Tobias, Jens, Petter, Team Stefan Löfven and all the other great course mates!
I talked about motivation in the blog Motivation – keep it up! earlier. Since the motivation is the key to keep the dog interested in the work it is supposed to do thoroughly and intensively we should always pay attention to it.
In a study which dealt with Explosive Detection Dogs (EDD) Irit Parnafes-Gazit argued that “[…] despite the reports in the literature that suggest the importance of sniffing rate for successful olfactory detection, I suggest here that the influence of the dog’s motivational level on its detection performance dominates that of the olfaction variables (i.e. sniffing frequency). Under motivated situations the dog will recruit all of its resources in order to fulfill the assigned task successfully, by fully focusing its attention on the olfactory cues available for detection of explosives.” Thus, without motivation it is useless to use a dog for detection work.
In another study Gazit et al. tested highly trained Explosive Detection Dogs and how their motivation changed when they were training on a pathway where there were never any targets to find. Already after a few training sessions the dogs showed significant decrease in search behaviour and started missing newly placed targets. When the dogs were taken to train on another path similar to the clear one, their motivation was again higher as they had no prior expectations to this environment. And finally they tried to gain the dogs’ motivation on the first path by doing 12 training sessions per day where there always was one target per session. However, they failed to regain the level of motivation that the dogs had when working in the new environments or in the environment they knew the dogs would find targets.
This is a very common thing to happen and an effective way to decrease the motivation of a detection dog. This holds especially for the dogs that search for target scents that are rarely found in real life searches. As it can be seen in the study with the EDD dogs searching on the pathways, it is very difficult to gain the motivation again when it is once lost in a specific surroundings.
To keep the motivation up, in some cases you can hide the target scent in the working environment and let the dog to find it regularly. Some target scents, however, can be too dangerous to handle or otherwise impossible to leave unattended in the working environment.
A study of Porrit et al. tried to solve this problem. It explored whether another target scent could help a dog to keep up the good motivation for search, despite the fact that the main target was usually not to be found in the search area. One of the main conclusions of the study was that a dog can be taught various target scents and it is irrelevant how often the main target scent/s is found in relation to the motivating training scent. As long as the dog knows the main target scent/s, it will react to them and indicate them during the search even if it mainly finds only the training scent.
When teaching a dog the training scent, it should be taught separately from the main target scent. By doing so the training scent is added the scent memory of the dog. Usually the training scent is taught first when the dog is just learning the detection work. When the dog has learnt to search and discriminate the training scent, it is later easier to teach it to react also to the main target scent/s.
And this is where the KONG-toy comes in the picture. You may have noticed that the KONG-toy and pieces of it has been mentioned in this blog multiple times and it has been used as the training scent for some exercises. Why it is especially a KONG-toy? The reason is that it is made of patented material that is not used for anything else than these KONG products. To make it clear, this is no advertisement of KONG products and the enterprise is not sponsoring this blog.
The point here is that you can use any scent you like as a training scent, but you need to understand the possible side-effects that might bring. If you for example train with some spice, a tennis ball or another kind of toy, the possibility of them or some ingredients of them appearing in a context that you did not plan to be there is bigger. That means, that in the working environment there is a bigger chance the dog might find the training scent and it is harder for you to control them. If a dog indicates a kitchen cupboard or someone’s bag, it is harder to verify if there is a small amount of a spice or some other material the dog has learnt to indicate, than if there is a KONG toy or some other unique material.
KONG as a training material is popular among many dog training professionals for some other reasons too. It is easy to cut in pieces, but it will not crack even when handled in small pieces. It is possible to boil it to decrease the distraction scents. The red colour is easy for a human eye to see, but dogs can not distinguish red and that way they can not look for the pieces with their eyes so easily. And of course, you can use a KONG toy as a reward, which increases the motivation of a dog to find the scent even more.
The final days of the course with SWDI were about scent detection. (You can find the blog about the first days here.) The composition of the group changed a bit after the tracking days and we had to say goodbye to some of the great dog teams we had just got to know. The detection course brought along some new handlers and dogs from Sweden, France and Germany. Our training facility also changed from the Lindesberg town to Grimsö Wildlife Research Station that was located in the middle of the beautiful Swedish nature.
First we started focus training with the dogs favourite toy. The purpose of the exercise was to get the dog to stare focused at the toy until it was released to take it or get another reward. The idea behind this exercise was to train the dog to focus on the target (a toy, but also later the target scent) and ignore any distractions there might be. So the dog was supposed to keep its eye on the toy and not even look at the handler or anything else, no matter what. The dogs made progress very fast if the trainers were consistent when increasing the criteria.
The next step was to train scent detection with a small article on a small area. There were small drawers, card board box pile and a pile of tyres where you could easily hide those small pieces to be found. It was important to build up the expectation for the dog to find something in that area by making the exercise very easy in the beginning. As soon as they understood to search on that specific area, they had more patience to stay there and search for even smaller pieces and working more thoroughly.
After small area search we moved to line ups. The line up blocks were first introduced to the dogs by one block at a time. As soon as the dog got the expectation [that it was possible to find a target scent there], more blocks were added. We worked mainly on rows of 6 blocks. The exercises varied: the distance of the blocks was changed (making the row that way shorter or longer), the dogs were kept on and off the leash, the handler changed her walking speed or she just stand aside when the dog was working independently. And of he amount of target scent was varied and different distractions added.
Again, the dog was supposed to ignore the handler, no matter what s/he did. One way to make the handler “disappear” from the training situation was to use another person to reward the dog. The handler did not even need to know which block the target was to minimise the impact of the handler on dog. The handler could either walk along the line up with a steady pace or stand aside. Either way the dog could not take any hints of the target scent by watching the handler. If the dog decided to go with the handler, it missed the opportunity to get rewarded. The reward would only appear if the dog found the target scent and stayed with it. Very rapidly the dogs got the idea and ignored the movement of the handler and stayed focused on the line up.
It is important not to get stuck with any routines or manners. Everything should be varied.
You should vary the place you put the hide in: the hight, the depth, the position in the search area, the amount of the scent, etc. And addition to that, you as a handler should also never get predictable. You should train where you stand or sit, how you move or talk, and what happens outside of the search area when the dog does not see you. Other things to vary could be for example size of the search area, surfaces, lighting (darkness), distractions and so on.
Unfortunately my scent detection course ended too early, because I needed to get back to work. I missed many information-packed training days with this talented group, but I have a feeling that this was not the last time I visit SWDI.
One essential thing with every goal-oriented training is a well thought training plan with regular check ups and testing. This is one of the things that was stressed on these courses. At the SWDI they suggested to have training plans for three months which are tested every three weeks. “What you cannot train in three months, you cannot train at all.”, they said. If some problems become visible in testing, you should change something in the training plan to get progress.
I prefer to base my knowledge on facts. What I liked especially in these courses was, that the facts were introduced based on research and available studies made on dog behaviour – and they were taken into account when planning the exercises. We got a lot of information and articles that explain for example how important it is to collect data on your training and to test your dogs skills on regular basis. We also learned how the dog handler affects her dog , how to make use of the play when training, and what is good and bad stress for the dog and how it affects learning. Some of those articles have been already referred in this blog earlier, but there are a lot more to refer to in the future.
The days were really intense and full of theory, practise and good conversations. If you ever have an opportunity to take part in SWDI training, I would strongly recommend on attending.
Thank you for a great week Jessica, Jens, Tobias, Petter, Marie, Mats and Emma!
Also many thanks for all our course mates! It was great to meet you and train with you all!
And last, but not least, thank you Teemu for an awesome travel company!
A couple of weeks ago there was an excellent tracking camp organised in Sweden by SWDI (Scandinavian Working Dog Institute). The camp was followed by a detection training week arranged for professional dog teams. I had a great opportunity and privilege to take part in the training. Although I promised to share some of the more sophisticated motivation techniques with you in this blog post, I would like to share some of the highlights of this training camp first. We will get back to the motivation after these two posts, I promise!
For the first three days we focused on hard surface tracking in a small town called Lindesberg. There were total of eight professional dog teams from Sweden, Norway, the UK and Finland, mostly from police, customs, army and private security services. Some of the dogs were experienced with scent work, some of them were about to do tracking from the beginning.
The training started by teaching a dog to react on an article. The article could be basically any small object which not so easy to see from the ground, most often a small coin or a piece of a kong toy. The exercise was much the same as introduced in one of the previous blog posts “Starting the scent training“.
The next step was to teach the dog to search for an article in a triangle shaped area scented by its handler. The reason for this is to teach the dog that the article can be found only in the scented area. This will help the dog later to stay better on the track. The scenting of the area was made by the handler stepping and possibly touching the ground a bit with his/her hand. Then the dog was brought to the area. Besides the scented area, the dog got no additional help – there was no food, added water or rubber from the shoe soles rubbed to the area. Just the scent that came naturally from the handler.
It should be noted that the surface we trained on was not free of distraction scents. Because it easy for dogs to discriminate their handlers fresh scent on the ground with some older human scent distractions, it is better to courage them to do that right from the beginning.
There was only one article placed in the scented area on asphalt that was approximately 1,5 meters wide and long. The article itself was about 2 x 2 mm (double the size and you would be mocked that you are using an article that is a size of a football!) If it seemed too hard for the dog to find the small article, you could decrease the size of the triangle or place in the area some more articles. It was important to get the dog to succeed and build up the expectation for the dog to find something valuable in the scented area. The idea of this exercise was to get the dog to sniff intensively and use its nose instead of eyes, which it still could do with a coin.
When the dog had learnt that the article could be found only in the scented triangle and was able to stay inside of it, it was time to do the first tracking exercises. By making a triangle and walking out of it from one of its corners, the dog was naturally lead out of the are. When there was no articles in the triangle, the dog would eventually follow the scent on the track and find an article on it. Bit by bit the length of the tracks were increased and the amount of the articles varied.
As soon as the dog got the idea of tracking, it was time to make turns and surface changes: from asphalt to grass, from grass to gravel, from gravel to asphalt and so on. It was essential to make all possible variations of the surface changes ibecause some changes could be harder than others for the dog to solve.
To increase the success of surface changes, an article could be placed right after the surface change. It was also possible to make the track more interesting to the dog by doing something different than just walking while laying the track. You could touch the ground with your hand or stand still for a while in a one spot, make a lot of turns and vary the distance of the articles.
The dog was of course rewarded for all the articles it found, but it was important to reward it also when it was on the track and working well. The reward in these cases was usually a toy thrown in front of the dog to keep the dog facing the fresher track.The reward was mainly an intensive and interactive play with the dog. However, for some tasks that required a calmer reinforcement the food was occasionally used as a reward.
If I compare the way SWDI teaches the tracking with the traditional way of doing it (placing pieces of treats to the footprints), everything seemed more simple. Maybe not for the handler, but especially for the dog. It was important to keep the task straightforward: If the surface changes, the articles are still kept the same size to keep the accuracy the same. The dog should be using only the exact track (max. about 20 cm off the track) and not go with the wind or scent plumes, to make it easier to stay on the track all the time. Handlers movements and guiding are faded off quite soon, to make the dog independent.
If the training is unambiguous to the dog, it will understand the point more quickly, it will succeed better in a given task and it will become more self-confident and independent. Sounds easy, but often we train too many things simultaneously, and confuse the dog.
Did you know, that there is no research done on how long the scent lasts on different surfaces after it has gotten there? There are many books that give you “facts” about it, but it is all based on opinions. No one has ever made a study on that yet, which is quite surprising, taking into account how long people have used dogs on tracking.
Some Nose Work
Between the two courses with SWDI, we had one day off. To make the best out of it, we headed with my travel companion Teemu to Borlänge to meet some Nose Work professionals of Dala Hundservice.
Before attending a weekend long Nose Work course we had some great training sessions at the fire departments training facilities. It was great to see how assertive and talented scent detection dogs work in a place with variable surfaces and unfamiliar scent environment with no problems. Nose work is a dog sport for all the dogs out there and it is supposed to be fun to train and compete with any kind of dog. It was amazing to see how a dog team, trained for only Nose Work scents work as any highly trained working dog. Even though Nose Work is “just a dog sport”, you could not tell the difference with a talented Nose Work dog team and a talented professional dog team.
Part II about scent detection will follow soon…
Success is a result of hard work. And as we all know, to work hard we all need motivation. This is no different for dogs.
How to motivate a detection dog? And how do you keep the motivation up?
To not to kill dog’s motivation, it is important to build the wanted behaviour in small steps and not to rush. For a detection dog this means teaching a target scent and boosting the dog’s will to search for it. If the target scent is not naturally interesting to the dog (like it is often the case), the dog needs to be rewarded with something it really wants. For most dogs play or food work well as a reward, but for others we need to be more creative. It is essential to understand that it is the dog that defines what is rewarding for it. We can not decide that for it.
Evidently, one criterion when selecting a working detection dog is how easily it can be motivated to perform different tasks. Practically any dog, however, can be trained to search as long as the handler knows how to make training motivating.
First, the target scent should be the best thing the dog knows and this should be stressed especially in the early training. In the beginning the exercises should be easy and the dog should succeed every time. By arranging the search environment and hides in a way that the dog finds them independently, builds up its motivation and will to do the task again. If the dog always succeeds without the handlers help, it learns that it can make it on its own and it will not ask help. The more fun the dog has searching and finding, the more motivation it has to find the target scent even when the search area is sometimes empty later in training.
Second, you should never end rewarding your dog even when it has learnt the task. When a dog is rewarded for finding the target scent time and again, searching itself will become rewarding for the dog, based on Premac’s principle – since the finding and alerting is rewarding, the behaviour before that becomes rewarding too. Admittedly, in some cases the learnt behaviour becomes so self-rewarding that the dog wants to work just for the work itself, but in most of the cases the dog needs to be rewarded for its effort regularly. For example tracking is very natural behaviour for a dog. If the dog likes it and can do it for its job, it might like the tracking itself more than the reward in the end.
The scent detection tasks vary depending on the target scent and the environment the dog is working in. If the scent is something that can be found in multiple places and the dog often gets to find it, it also gets rewarded often. And the more often it gets rewarded the higher its expectations are to find the scent the next time its working. If we take for example a mold dog. Mold is something that can be found in many buildings. If not in the structure, at least in the drains. So the mold dog finds the target scent in most of the cases and also often gets to alert in multiple locations. So whenever a mold dog is off to work its expectations are continuously high and so its motivation is easy to keep up.
A video of a highly motivated hunting labrador training is found behind this link. (Trigger warning: a dog carrying a dead rabbit.)
But what about bomb dogs in a highly secured building where the risk of bombs is high but the probability of them existing is low? A dog may be working continuously in a same environment without ever finding anything. Even the best detection dog with the highest motivation and drive will at some point realise that there is nothing to find. And if the dog never finds anything (and is never rewarded) its motivation to work will drop.
These were the basic principles that can be used when building a dogs working motivation in general. More sophisticated techniques at that the professional dog trainers use will be discussed more in detail in the next blog post.
To keep the dog interested in its work it needs to find the target scent often enough. It depends on the dog how often is enough. When teaching an animal a strong and durable behaviour the key to success is variable reinforcement. It means that the reward is not available every time but the handler creates a pattern (a reinforcement schedule) that determines when the dog has a possibility to find the scent and get the reward. In scent detection the variable ratio is usually used to teach a dog to search larger and larger areas. The hides are there somewhere but the dog does not know if those can be found soon or later. It only has its hopes up for finding it soon. Reinforcing with variable ratio keeps the task unpredictable and exciting, when the dog can never know when the target scent is there to be found. This is important to remember in every stage of the training – including very experienced dogs.
When a handler makes hides for the dog to keep it motivated there are some things that should be paid attention to:
Patterns: We people tend to build up behavioural patterns. In scent detection training that means that we often make hides in similar places (like in drawers or behind a radiator) or walk the same route when making the hides (like walking first to make the hide and then walking around for distraction) or avoid making hides in places where it is hard for us to put it (like up, where one would need a chair to reach). Dogs pick up our patterns very quickly. They also often recognise patterns that we ourselves do not even realise having, like the earlier blog post describes. If a dog is being motivated by training hides, there should be many different people making them so that the possible patterns get broken and the searching stays interesting to the dog. Otherwise the dog may not work well when the pattern is missing, which usually is the case in real life searches.
Handlers signature scent: If the training scents are repeatedly handled or stored by the dogs own handler, they will get contaminated with his/hers individual odour. If the handlers odour is continuously simultaneously present with the target scent the dog might start searching for a scent mix (target scent + handler) instead of the original target scent. If this happens, the dog might not alert the plain target scent anymore. It is important to change the target scent often and use also other handlers training scents to avoid this problem.
Guiding: When we train our own dogs and know the locations of our training hides we have the tendency to get frustrated if the dog does not find it as soon as we expected. It becomes tempting to guide the dog just a little bit so that it would find the hide and we could reward it. But if the dog learns that we help it to do its work it will start observing us more and slowly becomes dependent on our guidance. The more the dog relies on our guidance, the more inefficient its independent searching develops and the worse the results in a real search are.
Handlers motivation to work is almost as important as dogs. If a handler is stressed or unwilling to work or lacks trust to his/her dog, the dog will sense the handlers mood and it will affect its work. Not necessarily immediately, and it depends again on the dog how easily it will react to its handler, but it is important to be aware of the possible effects. Some dogs may react by feeling a pressure to alert which leads to false alerts. Some dogs may skip alerts even when they have found the hide and some may actually start avoiding the searching entirely, if they connect handlers bad mood to the work it is doing.
For us handlers it is always encouraging to see how our hard work and training pays off: when the dog alerts the target scent in a real life search and finds what we were looking for. Especially when we are not expecting anything to be found. That is the goal we are aiming at and which gives us handlers the ultimate motivation to keep on training.
(If your dogs motivation suddenly drops without an obvious reason, always consult a vet first!)
Our dogs act to get our attention to communicate with us. They may bark at the door that we would understand to let them out, or scratch the cupboard door to get some treats or just sit next to us that we would pet them. A dog uses these behaviors that have produced the wanted outcome before and hopes us to react to these behaviors. It is a dog’s way to communicate.
In scent work it is common to teach a dog an alert behavior, a final response, for instance sitting, scratching or “freezing” to the hide. This behavior is also called an indication behavior. The point of indication behavior is to give a dog a mean to tell its handler that it has found and narrowed down the spot of the target scent. But how important it is for a dog to know and use a specific alert behavior?
A couple of weeks ago there was a discussion in Facebook group “Nose Work” concerning
alert behavior in dog sport called K9 Nose Work. It was based on a blog post arguing that there is no need for a taught alert behavior to understand when the dog has found the target scent: it is enough if a handler knows how to read the dogs body language to pick the right spot of the target scent and call it an alert. The blog post argued that if the taught behavior is not strong enough for the dog, the dog may be insecure at a find and even leave it unalerted. Therefore a handler should rather learn to read his/hers dog well enough to pick up it’s natural indication behavior at the hide and call it a find even without the taught alert.
When talking about the sport of Nose Work it is probably enough to read the dogs body language, at least in the first class competitions where there are no control scents or other intentionally placed distraction scents that a dog might get interested in at the search area. When the dog gets interested in a scent it is presumable that it has found the target scent. But what about when there are distractions? Another dog may have peed in the search area? Someone may have dropped smelly treats there? And what about in upper classes where they intentionally put control and distraction scents to the area?
It is natural for a dog to be interested in some scents and some scents are more interesting to some breeds than others. In general, sniffing dogs are usually especially interested in control scents that are used continuously for training. In Nose Work those would be for example cotton swabs, blotter paper, stick-on floor protectors and so on, whatever material is used with the target scent repeatedly. When a dog finds the control scent from the search area it might sniff it more intensively than the other places that area. It might look very similar to the targeting sniffing behavior at the target scent. If the handler misinterprets this as a sniffing behavior for the target scent and calls it an alert, the team will lose important points. But if the dog has a strong taught indication behavior, which is in stimulus control, it should not indicate. No matter how interesting it finds the distraction.
The other problem that may occur from the lack of a taught indication behavior is mistargeting the scent. If in the search area the air flows move the scent plumes to a different place where the actual source of the scent is located, the dog may try to find it from the place of the scent plume and not the actual place of the scent. Depending on the search area and the scents volume, the plume might be strong enough for the dog to behave like it is close to the target scent trying to indicate the spot. There is a risk that the handler misinterprets this behavior and again calls an alert in the wrong place. If the dog is used to pin pointing the target odor it usually can differentiate whether it is targeting to the plume or the source of the scent, especially, if the source of the scent is small enough to pin point. If it tries to find the source from the plume but does not succeed, next it may try to find it elsewhere. In a competition situation, where every second counts and a handler is usually a bit stressed, this could be a place of false call of alert. But if the dog has a strong taught indication behavior, the handler can trust the dog when it is not alerting. And the handler can even help the dog find the source by analyzing the air flows in the area and directing the dog the other way.
Some dogs’ style of searching is very inconspicuous. Especially in Nose Work, which is a sport designed for all kinds of dogs, the searching behaviors vary. Where other dogs look like going crazy when finding the target scent, some just sniff it a tiny bit longer than other places and even an attentive handler might miss if no final response exists. Usually the dogs get back to that place where they spotted the target scent even if the handler misses it the first time, but it costs in the Nose Work competition multiple valuable seconds.
Nose Work, like many other dog sports, is a hobby that is mainly supposed to be fun both for the dog and the handler.But things get more serious when the dog does sniffing for its living. Should the handlers then always need an alert behavior of not?
Depending on the scent a dog is detecting, the need to teach alert behavior varies. If we need a dog to pin point an exact source of the target odor the alert is necessary. As an example we can use search and rescue (SAR) dogs that need to show the exact place of the missing person in a rubble. The dogs need to alert the exact place where the human scent is the strongest. They need to keep on indicating the place of the strongest scent source as long as the rescue people have located the missing person to start the rescue operation. It is not enough for them just to be interested in a specific place and sniff it. The handler needs to be 100 per cent sure that the dog indicates only the living human scent and does not indicate anything else that it might find interesting. There can be lot of interesting smells in a collapsed building for a dog.
When finding small amounts of target scents like individual bugs, it is again important to find the exact location of the target scent, the bug, to know where to spread the pesticides. Thus, it is not relevant to know where there is some of a bugs scent but to find the place where the bugs are in the very moment. Likewise, a drug detection dog which is used to search vehicles in customs needs to give a clear, objectively verifiable signal at the target scent.
Some working dogs do not necessarily need a specific alert behavior due to the nature of the work it is doing. For instance in customs the detection dogs may just stand aside of a gate when people pass it when arriving to a country. The dogs stand still and keep on sniffing the air flows that come in with the arriving people. If they pick up the target scent they will walk in to the crowd to find the source of it and follow that person. Then the dog has done its work and the handler stops the person and asks him to step aside for a further inspection. However, although these dogs do not have a specific alert behavior (besides the following) for this specific task, they are still often taught to indicate their findings on luggage and post packages by some alert behavior.
There are no short cuts to a reliable alert. It takes a lot of work and practice from both handler to learn to read the dogs behavior at a hide and the dog to learn a strong, reliable alert.
First, the target scent needs to be very clear to the dog. It needs to know what is the exact target scent and what are the control and other distraction scents that it is not supposed to react on. Part of the training is to teach the dog not to be interested in anything else in the search area than the specifically taught target scent. Most of the distractions are easily taught for the dog to ignore but you can never prepare for all the possible distractions that a dog may find naturally interesting. Thus, to avoid possible misinterpreting on dogs searching behaviour, it should be taught a strong behaviour chain where the target scent triggers the alert behavior before the dog even thinks. Every time the dog finds the target scent it should react to it with no hesitation.
Second, it is useful to know your dogs body language and behavior so well, that you can see when it is on the scent and close to the target. But its gestures should not be its only way to alert, if you want to be sure and avoid any possible misinterpretations.
In the end both are important to master: understanding the dog body language and the taught, strong alert behavior. When the handler reads the dog well and the dog does the alert behavior as an almost unconscious reaction when finding the target odor, the team is effective and they can rely on each other. It is possible to cope with just another, but for the best result, it is best to have them both.
You can teach an old dog new tricks but the lazier trainer you are the earlier you should start. At its best, the training starts already in the puppy pen.
If you are lucky to get your puppy from a breeder who is willing to put a little effort on them while they are tiny, there are some things you can ask a breeder to do already in the puppy pen. The studies show that the puppies’ sense of smell develops as early as in their mothers womb. So for example the food that the mother of a puppy eats might affect the scents puppies later get interested in .
A few systematic training programmes are developed to activate the senses of puppies . One of the most well-known is Bio Sensor-programme, also called “Super Dog- programme, rather widely used by advanced working dog breeders. It is based on stimulation of the early neurological system of the pups in order to improve their performance later in life. There are opinions for and against of its effectiveness, but firm scientific evidence of its effectiveness is still lacking.
Another quite similar programme, ESI – Early Scent Introduction, is developed to stimulate the sense of smell of a puppy. In ESI different kinds of smells are introduced to puppies in the same time period as the Bio Sensor is applied. Multiple different scents are expected to develop puppies’ sense of smell so that they become more accurate in the actual scent work later. No specific studies exists on the effectiveness of ESI, but several studies made of other mammals show how important it is to get used to new scents soon after birth (see e.g. Fillion & Blass, 1986; Kaplan, Cubicciotti & Redican, 1979; Muller-Schwartze, Muller-Schwartze, 1971)
Sometimes you hear people say that a puppy needs to be at least six months or even a year old before you can start training them. But since dogs learn all the time and everywhere, a puppy would have all that time to learn things on its own way. And then all of a sudden an owner changes the rules and a dog is expected to follow them from that day on. Doesn’t sound fair, does it? So why not start training on the day the puppy arrives and keep the same rules from the day one?
Food is usually very valuable for puppies. Especially, if they have had siblings as their litter mates, they may have fought to get their share. Thus it is easy to use food as reinforcement. Most puppies eat kibble when they move to a new home. Have you ever thought how many individual kibbles they eat per day? Have you ever thought how many reinforcements that means? That’s quite many! There is no universal rule saying that a puppy should eat from a bowl. Instead of giving away “free food”, you can ask the puppy to work for it. No matter what kind of food the puppy eats, you can always split a daily food amount in to multiple small portions to be used in training. Indeed, you can use all the food for training. Using it as a reinforcement, the value of the food stays high.
The house rules are important to learn in the beginning, of course, but there are things you can do to prepare your pup for scentwork too. Dogs learn really quick when they are young as they have not much learning history yet. The behaviours that they learn while being a small puppy get very strong, if you just remember to reinforce them every once in a while. So teaching a young puppy to use it’s nose early on may pay off later in scent work.
You can start teaching with a training scent or the “real scent”, the one that you already know the puppy is going to need to learn later for its work. Using a training scent first may be easier to handle than the real one. And if any problems occur along the way, the problems will not be connected to the real scent. Remember though that every time you reinforce a scent, it becomes important to the dog. The more reinforcements it is followed by, the more important it gets. So using a training scent means, that the dog will react to it later, even after you have taught it the real scent. It is possible to teach the dog to ignore the training scent, but it takes a lot of hard work to make it so reliable, that the dog never reacts to that anymore. So choose the training scent carefully, so that it won’t affect the dogs work later. One of the common ones to use is a kong toy or a piece of it, as that rubber material does not exist anywhere but in those toys.
After you have chosen the scent to work with, you can start the training just as you would do with older dogs. One possible way was introduced in an earlier blog post Starting the scent training.
If you are using a scent that needs to be covered, you can start by using cans or other containers in scent discrimination line up to keep the puppy from touching the target scent. Hiding the scent for actual searching can also be done with the containers in the beginning
For tracking, scent discrimination and searching, the ways to teach a puppy are the same as with older dogs. However, puppies do not control their body as well as older dogs. Also their bones and joints take time to grow strong, so the working environment should be safe, yet still diverse to let them get used to different kinds of surfaces.
Dogs can also learn from each other. By seeing another dog search and find, they might learn some things that we have not even thought of. This is called model learning. So in case you have an older dog, that does not get disturbed by a puppy sticking its nose wherever the older dog does, then you might benefit this in training.
Below is one way of starting the search training. In this video you see a 9 weeks old labrador pup learning to get interested in an object. It gets rewarded every time it approaches the object on the floor.
The next step for this pup was to learn an alert. In the video the pup is supposed to get down and point the object with its nose. There are multiple ways to teach alerts. In the video it is done by shaping the environment.
Adding distractions is one of the most important things to remember while building a strong behaviour. In the video below you can see some rather big distractions that came naturally in the training environment. Adding distractions bit by bit is crucial to keep the learning process going on and to prevent errors. In the video the puppy reacts to the distractions a little, but they are not disturbing it too much, as it can continue the exercise rapidly after that.
When the alert is built, you can continue to searching or tracking exercises. It is easy for the pup to start the new task in the same environment where it has worked before. In this pups case, the first searching exercises were done inside. When the searching in going smoothly in one environment, it is time to change the environment and to get successful exercises in many different environments with all kinds of distractions that can occur later in work.
The size of the training area is also important to consider. A small pup sees the world differently from adult dogs. There are multiple new things for them to experience every day, that seem ordinary to us. When choosing a place to train, consider the distractions and set the distance from the target scent so that it is first easy for the pup to find. You can also use multiple targets, to increase the pups chances to succeed.
In the next video you see the same labrador pup, now 11 weeks old. It’s doing a search in a skate park. Its work is intense and solid even though it takes time to find the object.
When training a puppy, one must remember to keep the exercises short. Do rather many short training sessions per day than only one longer one. Puppies learn fast, but they also need a lot of sleep to process the things they have learned during the day. They also need playtime and time to explore their environment. Time to be just puppies.
Try to keep the exercises interesting for the pup and plan them so that there are many opportunities to succeed. It is the way to build up the confidence and durability. To encourage independent work, let the puppy solve the problems by itself. Avoid helping it to find the target or get over obstacles. Create the learning environments so that it has possibility to figure out the right way to act by itself. If the exercise seems to be too hard to accomplish, call the pup away, reward for a good try and make it a bit simpler next time.
When the dogs explore their surroundings, they discriminate scents all the time. They know who of their favourites or enemies has just walked by the street, and they can smell where their owner has been while gone. It has been said that a human can smell pizza and think “Mmm…Pizza!”, but a dog can smell pizza and think “Mmm…pepperoni, garlic, tomato sauce, mozzarella cheese, wheat, olive oil,…”.
When we speak scent discrimination we typically mean a specific exercise where a dog is sent to a set up that is called a scent line up, a scent wheel or a scent wall. These are different types of set ups that are used to train dogs to find and indicate a specific target scent. These set ups are not made for dogs to learn to discriminate scents in general, but the scents that are important for their future job as detection dogs. For the dogs (and other animals) working in laboratory circumstances, the scent discrimination is their main working tool and they don’t necessary need to be trained for searching in other surroundings.
There are three typical discrimination exercises 1) reinforcing the target scent, 2) discriminating the target scent from other scents and 3) stimulus control.
Reinforcing the target scent is an exercise where only the target scent is placed in one or more spots of the setting. The dog searches for the target scent that can be placed anywhere in the setting. No distraction scents are used in this exercise, but the target scent can be made for instance stronger or milder. The other spots can be controls or they can just be left empty.
Control scent is usually a scent that is often or always present with the target scent, in practise and/or in real situations, but by itself is not relevant. One typical control scent is the scent of rubber gloves. If the gloves are used while handling the target scent, the dogs might start reacting to the scent of the rubber gloves. Since it is problematic that a dog starts to alert rubber gloves, it is useful to make that one of the control scents early in the beginning.
When the dog has learned the target scent, discrimination exercises can contain any of the other scents that might distract the dog while its doing a search or tracking. These could be for instance food, scent of human beings, scents of other animals or a toy. When these distractions are made non-profitable by scent discrimination, it is easier for a dog to ignore them while working too.
The last but an important use of scent discrimination is stimulus control. When the dog is trained long enough and thoroughly enough it should only alert the target scent. Stimulus control in scent work means that the dog alerts only the trained target scent/s, only when the target scent/s are present and never when it is not. Some trainers may even want a dog to alert only when it is allowed to work, after the search cue, but that is rather rare since in most cases there is no harm if the dog alerts the target scent without the cue. When training stimulus control, the target scent should be taken off the setting once in a while to test how the dog reacts. If the target scent is not present in the setting, the dog should not alert. It should either keep on searching or give an alert to “clear” depending on how its been taught.
Scent discrimination is an effective way to teach a dog the target scent and make sure that it only reacts to it and nothing else. It gives an opportunity to keep reward density high and make simple but informative searching exercises for the dog.
But what could go wrong then?
There are couple of common training problems that occur with scent discrimination with these kinds of set ups. If the final goal is to do searching or tracking and not only discrimination in laboratory circumstances one should remember that scent discrimination is only a tool to teach the dog the basics and to test its skills regularly. It’s something that should be done alongside with search and tracking training and not something that one should train constantly. When the dog knows the basics, it should focus more on the job itself. And here comes why:
One big problem with scent discrimination is us, dog trainers. As discussed in my previous blog post (How do you know, if the dog knows?) dogs read us people like open book. No matter how hard we try not to influence their work, if we know where the target scent lies, it’s extremely likely that the dog can read it from our body language. It is often seen how the handlers train scent discrimination by standing by the scent line or scent wheel. Or maybe even sit by the cans to change the order between the repetitions, like in this video clip.
It should be remembered that as long as the handler is in the sight of her dog the dog can read the handler! The more often this happens, the more the dogs behavior gets reinforced for reading the handler. Thus, the dog may look like it is working very well on a line or on a wheel but it really has no clue.
Another problem occurs when the dog is sent to a real search. It starts asking for help from the handler very easily since it’s used to seeing the handler and getting help from him/her. The dog doesn’t learn how to work independently and how to work in distance. This on the other hand makes the handler want to pressure the dog by either showing directions or by giving a command to search. And if the handler starts helping the dog, thus intervening its work, the dog learns that it can always ask the handler and the vicious cycle is ready.
To avoid the problems with scent discrimination, the set up should always be made so that the dog doesn’t pay attention to the handler while searching. or Blind or double blind exercises may also be helpful. The handler should either only walk the dog trough the set up with even pace or stand aside, away from the set up, while the dog searches independently. And if possible, another person should be present to mark the right behaviour for the dog. If the handler needs to work alone, an automated rewarding machine is a great help! That leaves the dog to focus on the task and the machine, and leaves the handler out of the sight. It gives no signs to the dog what so ever, no matter how much the dog tries to ask for help or read the body language of a trainer.
If the dog needs to do searching or tracking as its main task, it should not work with scent discrimination set ups for too long before starting to train independent searching. If the scent is strongly connected to a set up, it is harder for the dog to learn how to work in a bigger space without that familiar set up. To be efficient, one should first reinforce the target scent and then immediately start hiding it bit by bit to teach the dog to seek its way to it. After the dog knows that it is supposed to find the scent by sniffing and not only look for a container, it makes it easier for it to learn the scent discrimination set up as well.