The cost/benefit of animal training

Updated December 2023

This would qualify as abuse as defined below.

When planning to teach an animal to carry out any new response, there are two questions to ask before even getting started.

  • What is the cost/benefit of teaching the behaviour? Should it even be taught in the first place?
  • If so, how should it be taught?

Let’s look at the first of those two questions, as it will help decide whether a behaviour should be trained at all, and identify potential situations where we simply shouldn’t bother.

Oh, and I address the question of how to chose the best training method in this blog post.

Training new behaviour could in some cases cause suffering to the animal.

Suffering, in this example, is one type of “cost.” There may be other costs too, but for now I’ll settle for things like distress, discomfort or injury. So, for each prospective behaviour to be trained, one might assess whether there’s potential suffering involved.

Also, whether there’s any benefit to training the behaviour. Those benefits may include skills that help the human, skills that help the animal, or simply building rapport and connection between the human and the animal, even simply having fun together.

And so, we could imagine four potential outcomes of training a specific behaviour.

cost-benefit matrix
Trained behaviours and animal husbandry events may be useless, abusive, useful, or an ethical dilemma.
  • If there are no benefits to training the behaviour, then there are no reasons to train it.
  • If there are benefits to be gained from training the behaviour, there will be an ethical dilemma if the animal may suffer from learning or performing the behaviour.

This blog post is not about how to resolve the ethical dilemmas of situations where there would be both benefits and suffering involved. That type of situation could range from easy decisions such as whether to teach cooperative training for veterinary procedures, to difficult decisions such as whether to experiment on animals.

If you’re interested in those ethical / philosophical questions, and learn more about your own thoughts on the matter, check out the “animal ethics dilemma” test here.

I’m surprised by how often the cost/benefit question seems to be overlooked. It’s crucial in deciding whether to go ahead and train a behaviour or perform certain animal husbandry techniques.

Let’s look at a horse example.

The “cost” of hyperflexion

Apparently, there has been a trend in the last 20 or so years to ride dressage horses with heads bent increasingly beyond the vertical angle.

horse side view
“Normal”, vertical, head position versus hyperflexed head position (beyond the vertical line).

In terms of costs, could hyperflexion cause distress, anxiety, or injury to the animal?

I’ll be frank.

I’m not a horse person.

I’ve never seen the extreme position (hyperflexion) being trained live.

I don’t know the names of the bits, reins and whatnots. Never been to a live horse show.

But there’s been a lot of discussion about this in the media, so I spent some time on Google Scholar to educate myself about potential harmful effects of hyperflexion.

What I found was disconcerting.

Hyperflexion may involve an acute stress response as measured by an immediate increase in the stress hormone cortisol.

If trained coercively, hyperflexion is uncomfortable (as indicated by an increase in tail-swishing, head tosses and attempts at bucking). It also increases fearfulness, and the researchers warned that such horses could potentially be more dangerous to ride. In the same study, if given the choice, the horses preferred having their head in a regular position (with the nose line at or above the vertical) rather than in hyperflexion.

Perhaps most disturbingly, 80% of dressage and jump horses develop injuries in the neck, thought to be connected to the exaggerated hyperflexed postures (Weiler, 2006, 2007). However, it seems these findings have not been generally accepted by the international horse community, perhaps because Weiler’s findings haven’t been published in any peer reviewed journals (as far as I can tell). His thesis is in German, and I can’t find him referenced on Google Scholar.

This is the problem with science. Unless a study is published in a peer reviewed paper, it’s as if it doesn’t exist. From what I can gather in the snippets of quotes I’ve found, Weiler studied 1000 horses and found anomalies (inflammations and abnormal bone growth at the poll) amongst the majority of dressage and jump horses, but not in other types of horses. To me that would suggest that there is a correlation between riding style and injury, and inferring that one might cause the other isn’t exactly a long shot.

Hyperflexion (head bent beyond the vertical line) causes acute stress and possibly also long-term injuries. Image: Wikimedia

OK, so there is a risk of suffering in horses when exposed to hyperflexion. Not all horses all the time, perhaps, but all these different lines of evidence indicate that it could potentially be a very serious problem. As discussed in this blog post, assessing discomfort and pain is difficult in dressage horses, since many wear nosebands and double bridles that close the mouth so that the telltale signs of discomfort (yawning, open mouths and thrashing tongue) disappear. With that standard equipment, the horse thus has fewer ways of communicating that he’s not at ease.

However, a study comparing head- and neck positions of horses during warm-up and the subsequent competition found that not only was the head-position more than 5 degrees further behind the vertical during warm-up, but horses also showed about 30% more conflict behaviour during warm-up – a clear indication of discomfort. We’ll get back to the disturbing observation that horses are ridden differently during warm-up than the actual competition.

Parts of figure 2 from Kienapfel et al., 2021 looking at 49 starters in two Grand Prix tournaments in Aachen. Despite regulations, only one of the horses were ridden with the nose line in front of the vertical.

So, in the matrix above, hyperflexion would rate either as abuse or as an ethical dilemma depending on whether there are any potential benefits associated with the technique.

If hyperflexion is potentially bad for horses, why are they trained to perform like that?

The “benefits” of hyperflexion.

Benefits explain why we do things. Benefits somehow makes life better for someone, people or animals.

What are, supposedly, the benefits of hyperflexion? How does it make life better for people or animals?

I find four lines of argument hypothetically in favor of hyperflexion – but I don’t think they stand up very well to scrutiny.

Actually, they’re potential costs.

  • Do horses ridden in hyperflexion get better scores in dressage competitions?

Well, I find two studies saying that they do, and another saying that hyperflexion was actually penalized with lower marks in the lower competition levels, and there was no difference in the higher competition levels.

So it seems that whether hyperflexion is rewarded during competition depends on the individual judges – and since international regulations require the nose line at or in front of the vertical, it may even be penalized. An unpredictable outcome – not a clear-cut benefit; sometimes even a cost.

  • Do people think hyperflexion in horses looks nice? More attractive?

No. When asked to select a representation of their ideal horse, 93% of participants in one study chose images of horses with heads above the vertical line – that is, only 7% preferred the hyperflexed “look”.

I’m thinking that this is perhaps one reason why riders choose to ride with the head further beyond the vertical during warm-up than the actual competition: they know that the audience would disapprove.

I could be wrong, of course. Again, I’m not a horse person and I’ve never been to any horse competitions, but if the whole warm-up ring isn’t open to the public, this would set my alarm bells off. Just like my alarm bells went off at the abuse recently uncovered at the Helgstrand training facility, where most training occurs behind locked doors, out of the public eye.

And as appalling and infuriating as it may be that this double-standard hypocrisy is apparently widespread in the horse world, it’s not the topic of this post.

Back to whether people think hyperflexion looks nice; the answer is no, so this would be another cost, not benefit.

  • Do horses become suppler after stretching the neck – does hyperflexion improve posture and gait?

Not according to this article – hyperflexion puts an increased load on the forelegs, and the author Jean Luc Cornille calls it a “failure of Olympic dimensions”. Scientists warn that there is risk of injury if the horse is ridden too long in hyperflexion. Again, cost, not benefit.

  • Is hyperflexion a way to ensure a submissive horse?

Apparently in some parts of the horse training community, “submission” more or less equals cooperation. A submissive horse is apparently one that is supple, relaxed and responsive, as far as I can tell.

Not what I as an ethologist would call submission, but I argue this point and discuss the dominance concept elsewhere.

What I will point out is that there is a huge risk of learned helplessness during lengthy hyperflexion (Call et al., 1998) – that the horse has no control over unpleasant or harmful conditions, and that their actions are futile – they become helpless, and shut down.

Indeed, learned helplessness can interfere with learning and performance.

What does this mean? It means, that rather than performing better in dressage competition, hyperflexion could have the opposite effect. Indeed, cooperation can be achieved through other means – even with bitless bridles, and this argument is, just as the others, cost and not benefit.

So, hyperflexion is stressful and potentially damaging to horses, there is no clear competitive advantage to hyperflexion in dressage, the general public actually prefer seeing horses that are not hyperflexed, it’s arguable whether posture and gait is improved, and it may cause learned helplessness which interferes with learning and performance…

Cost, cost, cost, cost and cost. I don’t find any benefits.

So, I must conclude that hyperflexion in horses is generally abuse.

I could be wrong, of course. Maybe some of the “benefits” that I dismissed and reframed as costs actually has some value to them, at least for some horses sometimes.

In such cases, hyperflexion would perhaps be an ethical dilemma rather than abuse.

In such cases, one would have to weigh the costs and benefits of hyperflexion for that particular animal. Might costs be reduced? Benefits maximized? Is it worth the risks to the individual animal’s health?

That discussion is beyond the scope of this blog post. I’m painting a broad picture here, not discussing details and exceptions.

But I am curious.

Why are people training their horses to hyperflexion in dressage competition, given that there’s no obvious benefit but a pile of costs?

What’s the point?

Why do they expose their horses to hyperflexion, which is uncomfortable and stressful to the horse, and could cause long-term injury? Did I mention that their breathing is affected?

Judges are not automatically inclined to reward hyperflexion with better scores, and international regulations forbid it. Many of the so-called benefits are actually hidden costs.

Then why?

Why do people train a behaviour that is abusive – has no benefit and causes suffering?

Perhaps just because.

Because someone else did it. Someone influential. Saying “this is what you should do”.

Social conformity can make people do the inanest things, as shown in this video.

So, we can see how the practice of hyperflexion can continue to propagate despite being not only useless but also detrimental to the horse – despite being abusive – simply due to this phenomenon of social conformity.

After all, we’re social creatures, and we conform to the group, and fall prey to a large number of cognitive biases.

I suppose snapping out of this type of situation isn’t easy. I would expect practitioners of hyperflexion to get defensive when confronted with these findings. After all, nobody wants to hear that they’re abusing their animals, and we’d expect Cognitive Dissonance, the Semmelweiss effect, the Backfire effect and Reactance, four such cognitive biases, to be triggered in response.

Perhaps they might (if they were reading this) challenge my lack of understanding of horses, or the fact that I’m referring to research that hasn’t been published in peer-reviewed papers.

Or perhaps argue that the researchers didn’t know the first thing about horses either.

Or that there are benefits that I don’t know about.

Overcoming one’s cognitive biases is tremendously difficult, but being aware that one has them is a first step, and a rational approach might perhaps help avoid those cognitive biases from being triggered. Now, I’m not sure whether this blog post will help anyone overcome their biases, but the point I’m trying to make here is not emotional but rational: should hyperflexion be trained at all?

Look at the cost/benefit of your particular situation and ask: is it worth it?

The cost/benefit question is important, as it will help you decide whether to train a behaviour or carry out an animal husbandry intervention at all.

That decision is also a moving target, so adjustment is needed as we learn more.

A few years ago there were hardly any data indicating that hyperflexion was particularly detrimental to horses – there were no known costs in terms of suffering. In fact, in an FEI expert meeting in 2006 it was stated that: “in experienced hands there was no apparent abuse, improper welfare or clinical side effects associated from the use of hyperflexion”. Note that the papers I’m citing in this blog post are all more recent and reach the opposite conclusion.

Also, Rollkur (the coerced version of hyperflexion) has recently been banned by the North American Western Dressage Organization, and all the bad press will probably influence the judges of dressage competitions, further reducing the potential benefits. I mentioned that three scientific papers disagreed on whether hyperflexion was penalized or beneficial in dressage competition: in data from 2008 it was beneficial (Lashley et al., 2014), in a 2014 paper it was penalized (Kienapfel et al., 2014) – but in the Kienapfel paper from 2021 it was again rewarded. So it could be that in some circles, what used to be a benefit is now a cost in terms of dressage points – but not all judges have boarded that train, apparently.

They’re not on board with international regulations – yet. Hopefully public opinion will help make that shift sooner rather than later.

So, answers to the cost/benefit question change over time. In 2006, hyperflexion would not have been “abuse” or even an “ethical dilemma”, but rather it would have been classified as “useful” based on the knowledge available at the time.

A sobering thought. What are we doing that is “useful” today but will be “abuse” in ten years?

A paradigm shift is painful. I hope that practitioners of hyperflexion will take a few deep breaths and consider these rather recent findings when deciding how to go about training the next horse, or prepare for the next competition.

It’s never too late to change. That’s how we grow – and the cost-benefit question will help us develop, question norms and find new solutions in the world of animal training and husbandry.

But – the cost/benefit question was just the first of two important questions.

The second one is: how should a behaviour be trained? That’s the topic of this blog post!


I teach about animal behaviour, learning and wellbeing. Sounds interesting? Sign up to get notified whenever one of my online courses opens for admission, and I’ll also keep you posted on free trainings, silly experiments and new blog posts!


Caspar et al. (2015). Human Preferences for Conformation Attributes and Head-And-Neck Positions in Horses.

Christensen et al. (2014). Effects of hyperflexion on acute stress responses in ridden dressage horses.

Hall et al. (2008). Is there evidence of learned helplessness in horses?

Kienapfel et al. (2014). Prevalence of different head-neck positions in horses shown at dressage competitions and their relation to conflict behaviour and performance marks.

Kienapfel et al. (2021). Comparison of different head and neck positions and behaviour in ridden elite dressage horses between warm-up and competition.

Lashley et al. (2014). Comparison of the head and neck position of elite dressage horses during top-level competitions in 1992 versus 2008.

Rhodin et al. (2009). The effect of different head and neck positions on the caudal back and hindlimb kinematics in the elite dressage horse at trot.

Weiler (2006, 2007) – interviews in Horses for life Magazine.

Von Borstel et al. (2009). Impact of riding in a coercively obtained Rollkur posture on welfare and fear of performance horses.

Zebisch et al. (2014). Effects of different head–neck positions on the larynges of ridden horses.

32 Replies to “The cost/benefit of animal training”

  1. I have zero respekt for the trainer/ businessman whose methods were exposed recently, but it’s not just one bad apple, it’s a culture. It’s heartbreaking to see that nothing much has changed for the last 40 years. I still get embarrassed when I think about how we treated horses in the bad old days, partly because we didn’t know any better and partly because it was unthinkable to question the trainer’s methods. I tried once and I became invisible, pretty tough for a teenage girl who just loved horses and spending time with friends.

    1. Very saddening to see how pervasive that culture seems to be – and particularly the double standards! Showing one image to the audience, while something else entirely is going on behind locked doors! Hopefully the public reaction to these revelations will make judges actually step up and penalize hyperflexion – if indeed the sometimes-higher-scores is what maintains that behaviour..?!

  2. BTW: The is a number of Rollkur critics and they do put pressure on the FEI. Judges and riders should confirm to FEI rules against Rollkur.

    The FEI’s response, however, is laughable: They just deleted their Rollkur rules. ‍♀️‍♀️‍♀️

  3. “However, it seems these findings have not been generally accepted by the international horse community, perhaps because Weiler’s findings haven’t been published in any peer reviewed journals”

    Nahhh… it’s because riders don’t read. Virtually no rider refers to science for information. Knowledge is mostly passed down by word of mouth. It’s all about tradition and trust in the instructors or gurus.

  4. The noseband makes it more difficult for the horse to express pain from the bridle but the new pain scales developed for medical use might be used also for ridden horses, for example Equine Pain Face. It is difficult to differ pain from stress even if you use a pain scale but in this case both pain and stress are important to consider. I agree it doesn´t seem to give better scores to use hyperflexion but also it doesn´t give better scores to not use it. Probably it is not only used as a trend but as a quick fix, as it takes a lot of time, skill and perseverance to educate a horse and a rider.

    1. Good point about the importance of considering both pain and stress! (Pain I’d say often leads to a stress response, but stress doesn’t’ necessarily need to be pain-related)

  5. I love this course, I will never finish it, going from link to link ……now getting stuck in social conformity and how it just takes 1 other person to not comply with conformity to change so many other people’s attitudes. But ” straying from the herd” releases cortisol ( Loretta Breuning , PhD, )so it feels like a real threat . We human mammals can choose to run with a different “herd” …….
    You are training me good , I’m having fun SEEKING information and my partner is having fun earning stickers for putting on her glasses (she’s nearly 70) . Incidentally, both of us would have gotten up and asked the receptionist what the beep was about……..That comes from being used to being different

    1. haha, yes that video is soooo interesting! 🙂 so glad you’re enjoying the Getting Behaviour course, Maren! 🙂

  6. I read this article and disagree with the idea, that teaching some behaviours is useless and useless behaviours should not be trained. I don’t think thare are “useless” behaviours. Why?
    1. Teaching an inadvanced animal “useless” behaviours teaches it the rules of this game
    2. Teaching “useless” behaviours by inexperienced trainer teaches tham how to train
    And both of the above are achieved without the pressure that the desired behaviour will serve some purpose later
    3. As you wrote elsewhere – training is enrichment and
    4. Positive training builds positive relationship
    There may be bahaviours that we shouldn’t teach because they may be dangerous (teaching foot targeting a horse using a shoe as a target) or counterproductive with regards to other goals (teaching the dog to open the frigde). These I would call more than useless.

    1. Good point about the “useless” category being very small – perhaps I wasn’t clear enough about that. If the behaviour serves a purpose for either the human (learning how to train) or the animal (learning the rules of the game) then it is, per definition, useful…! 🙂

  7. There is an excellent DVD if you can find it called” if Horses could speak” it details why hyperflexion became popular. Judges did mark very high for it, despite their denying doing so. it looked much more exciting to spectators ,therefore increasing attendance and attention to the sport, getting bigger sponsorship and cash prizes. that is why it was reinforced once it was started with certain riders as mentioned in an above comment. In the video, they referred to the horses as “leg movers” versus classical dressage being “back movers” back movers actually build up the horses’ biomechanical soundness similar to a human who does Pilates. The Horses are healthier into older age. Hyperflexion breaks the horse down instead. they push Young horses way too early, which saves money in not having longer upkeep costs. Horses become unsound and undergo Neurectomy or other leg surgeries, and still are euthanized at younger ages. I was glad to read in the comments above that the sport is moving away from this. But it is still definitely a problem. it is referred to in slang as crank and spank, meaning the horse never gets a release and you’re just forcing the horse into the bridle with spurs and whip instead of having it use self carriage. Self carriage in classical dressage is something that becomes pleasant for the horse if it is using his body in a way that is not natural to it but with a rider on his back creates a weightless center with the rider at the apex of the upward curve of the spine of where the back has gradually increased muscle structure to support a rider. which is obviously not what horses were made to do, but something that we can help them do better because we do ride. In correct carriage, the horse feels release in the springy efficiency of the hind end powering the front and creating a reverberating rhythm that is harmonious for that particular horse and its rider.

    1. Interesting if it was marked high – even if judges denied it..! Hopefully that awareness is changing as the discussion continues…

  8. Gah! I got so caught up in this unit and following the trail of seeds (links to articles). Being in the equine world it was so thought provoking and enraging at the same time! So much of what I see I have to turn my head away and the reluctance of some people to even consider a better way. I have literally seen people knee their horses in the gut for not standing still at the mounting block and no one bats an eye to this! Why is so much force and punishment accepted in the horse world?!

    I am curious if you followed through with an article debating what animal submission looks like from an ethologists point of you versus what submission looks like say in the horse world.

    1. 🙂

      From the ethological perspective, dominance interactions are all about priority of access to resources.

      Say that two animals, who know each other, arrive at a resource at the same time.

      The dominant animal may show some type of aggressive behaviour, but more often the subordinate may show some type of submissive behaviour. Not “submissive” as in “I submit to your will” but as in “I will not contest you for this resource”. To me that’s the main difference between the ethological interpretation of the word and how I see it used in some animal training communities.

  9. In my early years of horse ownership with a feisty mare, I was encouraged to use more restraints, more equipment, more insistence. Thinking that these horse folks knew more than me I blindly followed their advice…how I regret that now. However, it wasn’t too long before I realized that things were getting worse, so off came all of the “dodads” and after searching for more humane treatment and knowing better to do better, I had the horse of my dreams. So sorry for her that it took a long time.

    1. I think you’re not alone in this frustration. I invite you to let go of that guilt, and enjoy the relationship you have now. Without you, she would have had a completely different life!

    1. I don’t see a place to edit …
      Rolker is trained in a N+ (P+) context as the pressure never has the intention to release. If the horse makes himself ‘lighter’ in the reins but still gives the rider the ‘feel’ they are wishing to achieve, the lightness is not because of the human releasing pressure.

  10. I also meant to say that the perceived ‘benefits’ of Rollkur, apart from slowly grinding a hot horse into submission and stopping a spooky horse from spooking, as he can’t see well enough with his head down there to see things to spook at, are that it can encourage an artificially extravagant front leg action. Horses often react to pain and irritation in their mouth and nose area by striking out with their front legs, as exhibited by horses suffering from a poorly understood condition known as headshaking, which is thought to probably be caused by pain from the trigemenal nerve (like trigemenal neuralgia in people). ‘Encouraging’ this sort of front leg action in this way can result in extraordinarily impressive front leg action in talented horses but it is totally incorrect as the horse is not working through the back and carrying more weight behind, the hind leg action does not match the front legs and it is all absolutely contra to the whole ethos of the purpose of dressage training but it did, for a while, impress judges into giving huge scores. Thankfully those days seem to be behind us now!!!

  11. I was very interested to read this blog, thank you very much for it!

    As an avid dressage follower for several decades now (I’m a amateur clicker dressage trainer myself) I’m pretty sure that the reason Rollkur became fashionable is because Anky von Grunsven and her trainer, Sjef Janssen, started using it with their horses and, for a time, were just about unbeatable (because/in spite of it, who knows?) and it very sadly became a widely accepted (in many circles) and fashionable training method. According to the rules judges have always been supposed to mark horses down who go behind the vertical but seemed to forget to do so for a long time. For some reason Sjef seems to have been enormously influential in getting the powers that be to overlook the fact that this training method is inherently abusive and allow it to continue, even to the point of protecting riders from cameras when warming up at competitions. The very powers that were supposed to be protecting dressage horses at competitions failed them completely.

    Then, fortunately for horses and for dressage, the hero arrived!!! Carl Hester and his pupil, Charlotte Dujardin, came along and showed the world that with correct, classically based training methods their horses could take on the world and win, presenting a picture of harmony, ease, grace and partnership that everyone recognised and admired. Now, at last, judges and commentators are marking down and criticising riders who ride with forceful, heavy hands and driving seats and horses who drop behind the vertical, who have incorrect movement and problems with the contact (with the bit and the rider’s hands).

    Carl Hester is leading the world back to classical, bio mechanically correct training methods because he has proved their worth in competition and everyone is now aspiring to be like him and Charlotte, thank goodness!!! So often it only seems to take one really influential person to change the world, for better or for worse. Carl is definitely the best thing to happen for dressage horses in a very long time!!!

    1. Helen, thanks so much for giving some background to this phenomenon. Glad to hear that you feel that there is a change of tides…! 🙂

    2. That’s so good to read, Helen, thanks for posting it.
      This is a very interesting blog. I’ve heard about Rollkur and it’s good to learn more about the costs and benefits, ethological view plus of course how things are developing on the discussion front.

  12. The ‘social conformity’ video was extremely useful (and funny and disturbing, I might add) in illustrating how vulnerable we are to social pressure. Even with the welfare of our human and nonhuman animals in mind we can all fall prey to ignoring the cost/benefit of a behavior.

    A very interesting blog, Carolina. Looking forward to part two.

    1. Chris! Glad to see you here – I’ll take this opportunity of thanking you again for the train ticket that you so graciously gave me in Denmark. 🙂
      I agree that the conformity video is disturbing. The behaviour is perpetuated into a completely new group – and nobody questions it..!

Leave a Reply

Your email address will not be published. Required fields are marked *