08/09/2025
A great read about many misconceptions about training- specifically food reward based training.
๐๐๐ญโ๐ฌ ๐๐ซ๐๐๐ค ๐๐จ๐ฐ๐ง ๐ฌ๐จ๐ฆ๐ ๐จ๐ ๐ญ๐ก๐ ๐๐จ๐ฆ๐ฆ๐จ๐ง ๐ฆ๐ข๐ฌ๐๐จ๐ง๐๐๐ฉ๐ญ๐ข๐จ๐ง๐ฌ ๐๐๐จ๐ฎ๐ญ ๐ซ๐๐ฐ๐๐ซ๐๐ฌ ๐๐๐ฌ๐๐ ๐ญ๐ซ๐๐ข๐ง๐ข๐ง๐ ๐๐ง๐ ๐จ๐ฉ๐๐ซ๐๐ง๐ญ ๐๐จ๐ง๐๐ข๐ญ๐ข๐จ๐ง๐ข๐ง๐ ๐๐ฌ ๐ ๐ฐ๐ก๐จ๐ฅ๐.
I will preface this that I used to believe in and say most of these things, this is not a dig at anyone, there is a significant lack of knowledge on operant conditioning in the Horse world it leads to a lot of confusion and defensiveness.
โ๐ก๐จ๐ซ๐ฌ๐๐ฌ ๐๐ซ๐๐งโ๐ญ ๐๐จ๐ ๐ฌ, ๐ญ๐ซ๐๐ข๐ง๐ข๐ง๐ ๐ฐ๐ข๐ญ๐ก ๐๐จ๐จ๐ ๐๐จ๐๐ฌ๐งโ๐ญ ๐ฐ๐จ๐ซ๐ค ๐๐จ๐ซ ๐ญ๐ก๐๐ฆ.โ
This one is a doozy, it implies that dogs are one of the only animals that learn from food rewards which is far from the case and has been displayed consistently throughout anecdotal experience as well as studies on numerous different types of animals.
Food rewards are a high value reinforcer for many animals. Horses are trickle feeding animals, which means their regular daily time budget involves eating for the majority of their time.
They are highly food motivated and their bodies are built in such a way that in some ways, they are actually more built to consume food throughout training than dogs are.
There has been numerous studies done on horses pertaining to rewards based training, and there is a consistent trend showing the efficacy of it.
Regardless of whether or not people want to train with food rewards, there is no denying the fact that it is an effective way to train horses when it is done correctly.
Spreading misinformation to try to justify the way you want to train exposes an internal conflict.
If you cannot be comfortable using pressure and release without trying to convince yourself that training with rewards is inherently inferior, it might be worth contemplating why that is.
โ๐ ๐ข๐ฏ๐ข๐ง๐ ๐๐จ๐จ๐ ๐๐ฒ ๐ก๐๐ง๐ ๐ญ๐๐๐๐ก๐๐ฌ ๐ก๐จ๐ซ๐ฌ๐๐ฌ ๐ญ๐จ ๐๐ข๐ญ๐.โ
This is not true, food can actually be used to teach horses safer behaviours around food. What teaches horses to bite is the application of the training, environmental stress, being hungry, resource guarding among other things.
This is the equivalent to trying to claim that pressure and release teaches horses to run away from humans. You can for sure use pressure and release to make horses avoidant of humans to the point where they want nothing to do with them, but this would be an indicator of incorrect use.
If horses are becoming food aggressive and pushy in a rewards based system, the fault lies with the application of the reinforcement, not food as a reinforcement as a whole.
It should be noted that the people who do this are rightfully angry when people use poor application of pressure and release training as an example for how it always works, so they should not be creating a double standard that they then direct at rewards based trainers.
โ๐
๐จ๐จ๐ ๐๐ฅ๐ฐ๐๐ฒ๐ฌ ๐ฆ๐๐ค๐๐ฌ ๐ก๐จ๐ซ๐ฌ๐๐ฌ ๐ฉ๐ฎ๐ฌ๐ก๐ฒ.โ
The truth is, many horses do not live species appropriate lifestyles. They are meant to be consuming forage on a constant basis, when they are unable to do this, it creates stress behaviors.
Lots of people try to train horses with food when they have gone hours without access to food prior. The discomfort that this hunger causes is rightfully going to make them more anxious around food.
Horses can start developing stomach ulcers in as little as four hours without access to forage.
So, imagine how many horses must feel if they are operating with an ulcerated an empty stomach, and then are offered food.
This can be addressed by ensuring that horses have near constant access to forage, feeding a flake of hay or a small meal of soaked feed before training and using low value rewards.
When training with food, it is imperative to use low value food because really sugary high value treats will make horses more stressed and pushy plus they also should not be fed in large quantities.
Forage based rewards, like hay pellets, are safe to feed in larger quantities and low value.
Increasing chew time can also help manage anxiety. So, some people may want to opt for chopped hay over hay pellets.
Also, reward in small handfuls of forage, not just a singular pellet or treat. It is much more likely to increase grabbing if the horse is barely getting a mouthful and does not feel satiated.
If you watch how horses consume food in a natural setting, they are typically taking full bites, not just a tiny little nibble or two.
Itโs also incredibly important to teach a bridge signal. A bridge signal is usually an auditory cue (the clicker in clicker training, for example) that signals the correct behaviour has been performed and reward is forthcoming.
It marks the exact moment the horse has done the right thing, providing clarity.
It bridges the gap between the correct behaviour being performed and the reward being fed.
Without one, there is a lack of clarity and anything the horse does between the desired behaviour and actually being fed the reward can end up accidentally being rewarded.
This makes it a lot easier for someone to mistakenly reward pushy behaviours.
So, with rewards based training, a bridge signal is a necessity for clarity.
โ๐๐จ๐ซ๐ฌ๐๐ฌ ๐๐จ๐งโ๐ญ ๐๐จ๐ฆ๐ฆ๐ฎ๐ง๐ข๐๐๐ญ๐ ๐ฐ๐ข๐ญ๐ก ๐๐๐๐ก ๐จ๐ญ๐ก๐๐ซ ๐ฐ๐ข๐ญ๐ก ๐ซ๐๐ฐ๐๐ซ๐๐ฌ, ๐ญ๐ก๐๐ฒ ๐ฎ๐ฌ๐ ๐ฉ๐ซ๐๐ฌ๐ฌ๐ฎ๐ซ๐ ๐๐ง๐ ๐ซ๐๐ฅ๐๐๐ฌ๐.โ
This is greatly oversimplifying Equine communication.
It also implies that horses perceived training done by humans the same way that they do with communication from other horses, which is not true.
Horses are well aware that we are not horses.
No matter what people do to try to model, Horse behavior, it will never be received the same way because we are not horses.
Horses also donโt ride around on each each otherโs backs, load each other into trailers or do quite literally anything that humans do with horses.
With that said, horses do use positive reinforcement in developing bonds with each other.
Mutual grooming is a very common bonding activity that horses will do and this is an example of positive reinforcement.
Both horses get mutual enjoyment from the grooming and their bond grows as a result.
But, regardless, we shouldnโt be making training decisions based off of how Horse is communicate with each other.
I also see a lot of people trying to justify physical punishment like hitting horses based off of the fact that horses will kick and bite each other.
This fails to acknowledge the fact that horses have a diverse and subtle language that typically involves lots of warnings before any physical punishment contact.
Healthy herds are not frequently kicking and biting each other.
Horses also speak the same language and can have far more clear communication with each other so people cannot liken themselves to being a horse.
โ๐๐ฌ๐ข๐ง๐ ๐๐จ๐จ๐ ๐ซ๐๐ฐ๐๐ซ๐๐ฌ ๐ข๐ฌ ๐๐ซ๐ข๐๐๐ซ๐ฒ.โ
Nope. A bribe precedes the behaviour, the reward comes after.
By the time the horse is getting a reward, theyโve already performed the desired behaviour.
But if we want to use this as a reason to drag R+ training, with this logic pressure and release would be coercion.
A reward is simply a reinforcer for a desire behaviour. You are effectively โpayingโ your horse for offering the desired behaviour.
No correct application of R+ involves solely luring horses with food aka โbribing.โ
โ๐๐ ๐ฒ๐จ๐ฎ ๐๐จ๐งโ๐ญ ๐ก๐๐ฏ๐ ๐๐จ๐จ๐, ๐ญ๐ก๐ ๐ก๐จ๐ซ๐ฌ๐ ๐ฐ๐จ๐งโ๐ญ ๐ฅ๐ข๐ฌ๐ญ๐๐ง.โ
Also not true. With R+, you shape behaviours and build duration and at that point can reward intermittently.
While new behaviours are being learned, rewards need to be more frequently but once the behaviour has been conditioned, you can reward intermittently.
Rewards build a pattern of pleasant outcomes following certain behaviours and the behaviour itself can become reinforcing due to that history.
In all honesty, you can build more duration between reinforcers with R+ than you can with R- (pressure and release) because the reinforcer isnโt built into the cue with R+.
Horses trained with R- need to be frequently cued with pressure and then the release of it.
You donโt ever โphase outโ the pressure with negative reinforcement. Itโs always used. People donโt seem to view this as a training weakness so why is alternative logic applied to R+?
โ๐๐จ๐ฎ ๐๐๐งโ๐ญ ๐๐๐๐ ๐ ๐ก๐จ๐ซ๐ฌ๐ ๐ ๐ญ๐ซ๐๐๐ญ ๐ญ๐จ ๐ฌ๐ญ๐จ๐ฉ ๐๐ ๐ ๐ซ๐๐ฌ๐ฌ๐ข๐จ๐ง ๐จ๐ซ ๐๐๐ง๐ ๐๐ซ๐จ๐ฎ๐ฌ ๐๐๐ก๐๐ฏ๐ข๐จ๐ฎ๐ซ.โ
R+ training focuses on the WHY behind the behaviour and addresses the underlying cause. You arenโt going to reward behaviours you do not want to see, so of course you donโt feed food following aggression.
You do, however, look at the likely causes of aggression which generally relate to fear, anxiety, chronic stress, pain, inadequate management and moreโฆ
When you start addressing the causes, the behaviour will either reduce in intensity dramatically or go away completely.
You can also โcounter conditionโ unwanted behaviours by rewarding a behaviour that is โincompatibleโ with the unwanted one. For example, a horse cannot bite at the person leading them and face forward with their head in a neutral position at the same time.
If it is more reinforcing to face forward and the reasons why behind the biting in the first place are being addressed, the horse is more likely to defer to that behaviour.
Complex behaviours like aggression or extreme fear behaviours that are dangerous are best addressed by dealing with the cause if you want to create safety and predictable behaviour.
Focusing on suppressing the behaviour without considering why it exists can lead to the behaviour becoming more unpredictable and extreme.
โ๐๐๐ฅ๐ฅ๐ข๐ง๐ ๐ข๐ญ ๐ง๐๐ ๐๐ญ๐ข๐ฏ๐ ๐ซ๐๐ข๐ง๐๐จ๐ซ๐๐๐ฆ๐๐ง๐ญ ๐ข๐ฌ ๐ง๐จ๐ญ ๐๐๐ข๐ซ. ๐๐ซ๐๐ฌ๐ฌ๐ฎ๐ซ๐ ๐๐ง๐ ๐ซ๐๐ฅ๐๐๐ฌ๐ ๐ข๐ฌ๐งโ๐ญ ๐๐๐.โ
โNegativeโ in operant conditioning terms refers to the removal of a stimulus.
Negative reinforcement is โnegativeโ because the removal of pressure is what reinforces the behaviour.
It may be more clear to call it subtractive reinforcement.
โPositiveโ in operant conditioning terms refers to the addition of a stimulus.
Neither of these terms refers to โgoodโ or โbad.โ
โ๐๐๐ฐ๐๐ซ๐๐ฌ ๐๐๐ฌ๐๐ ๐ญ๐ซ๐๐ข๐ง๐ข๐ง๐ ๐๐๐ง๐ง๐จ๐ญ ๐ฎ๐ฌ๐ ๐๐ง๐ฒ ๐ฉ๐ซ๐๐ฌ๐ฌ๐ฎ๐ซ๐ ๐ฐ๐ข๐ญ๐ก๐จ๐ฎ๐ญ ๐ข๐ญ ๐๐๐ข๐ง๐ ๐ฉ๐ซ๐๐ฌ๐ฌ๐ฎ๐ซ๐ ๐๐ง๐ ๐ซ๐๐ฅ๐๐๐ฌ๐.โ
This isnโt true. Not all pressure is an aversive pressure where the release of it is the reinforcement.
A lot of pressures are neutral. For example, gently resting your hand on your horseโs shoulder. They may not enjoy the sensation but they also arenโt bothered by it.
Itโs still pressure but the release of it isnโt reinforcing anything โ the pressure has no meaning until itโs linked to something.
Some pressure is also pleasant, like scratching a horse in an area they like. This would be positively reinforcement.
So, positive reinforcement can use neutral pressure to link to a previously learned behaviour, like following a target. It can then be put on cue so the pressure is used to cue a behaviour.
Once the horse understands the cue, it will look similar to how traditional pressure and release looks, but how it was taught is different.
The primary reinforcer is the food reward.
Rewards based training doesnโt mean you never touch your horse, use equipment or apply pressure.
It isnโt some magical telepathic method that works off of straight vibes and no physical touch.
It just means that the physical cues being taught are not being taught by the use of aversive pressure.
The bottom line is that R+ is highly effective with horses and many other animals.
A lot of the problems people attribute to rewards based stem from a lack of knowledge, lack of understanding and/or incorrect application.
It is fine if you donโt want to use rewards in your program, but you donโt need to spread misinformation.
Both R- and R+ can be utilized ethically in a program but Iโve personally found that rewards lead to my horses being more motivated in training and with a much more pleasant outlook towards training.
Regardless of how you train, it is important to understand how your reinforcers work and have a basic grasp of operant conditioning.
Lack of understanding of how the application of your training methods actually are received by the horse can make training much less effective or impact ethicality.
It should be viewed as a non-negotiable for anyone working with animals to understand operant conditioning and be able to explain the function of HOW theyโre training.
Unfortunately, the horse world is behind in this regard.