Has anyone found that natural horsemanship training messed your horse up? How does this happen?

[Ajierene]

@scislandsprite

However, I am a little confused. With using a leg aid and always pairing it with a clicker, wouldn’t it eventually move into R+ territory since the horse would be moving forward for the reward, not to alleviate the leg aid? Very much a hypothetical situation since that’s not the most common method but would the leg aid be counter conditioned over time with consistent R+?

You can look at it that way - especially if you have a forward horse. When I kiss to my horse out on the trails to move to trot or canter, she does so readily and learned that cue quickly. When going to the canter, I adjust my seat to direct which lead I want (if any), but no real leg cue and she’s happily off. EDIT: since we are talking about leg cue, the same would be/is for her - little bit of leg and she’s off.

Heading home with her…that’s sometimes another matter - she LOVES to adventure and I am often having discussions with her about how I don’t have all day to go crashing through the woods or down the road or off to wherever. We have to go back eventually.

A horse I ride for a friend, much less forward to the leg cue will likely always lean towards R- as he much prefers the walk…or stand…or you know what, let’s just go back home and chill.

[Scribbler]

@scislandsprite

I definitely agree that a significant amount of training is beyond one quadrant.

However, I am a little confused. With using a leg aid and always pairing it with a clicker, wouldn’t it eventually move into R+ territory since the horse would be moving forward for the reward, not to alleviate the leg aid? Very much a hypothetical situation since that’s not the most common method but would the leg aid be counter conditioned over time with consistent R+?

People combine mild R- training with rewards and that is perfectly all right.

There is no reason it has to be purist unless you are wanting to make a point about R+

I have found clicker training pure R+ to be really effective for tricks and groundwork on my mare. When she’s in the zone she will back up on voice command or with a wiggle of my finger. In the stall I usually have to touch her with my finger (light pressure). If we are having a bouncy event while handwalking, then I am going to use every ounce of my strength to keep her from knocking me over.

One of the puzzles to me is, what do people think the optimum outcome of NH or R+ is going to be? Horses have moods and opinions, and the more athletic they are the bigger those moods and opinions can be. You are never going to reach a place with any healthy horse where you are guaranteed that their moods and opinions are going to line up with your desires 100 per cent every time. Things that work nicely in an arena don’t work that way outside. You can do all kinds of liberty work in an arena but no one is going to walk a horse down the road without a halter and lead rope. (but you can train a dog to heel in public without needing a leash).

Anyhow, I think these training methods can implicitly sell a fantasy that somehow you can make a horse completely compliant to your every wish. And it’s not true. Working with horses is constant negotiation. And that’s fine! They have brains and needs and desires and instincts, they problem solve and are curious. That’s what makes them fun. We try to meet their needs as best we can but there are times we need to say, no here’s what I need.

[sascha]

@Scribbler

I heard a story years ago about a dedicated R + trainer that sat in an arena space for over an hour at night waiting for her horse to choose to exit the arena back to his field. She didn’t want to force him. I don’t recall how it ended.

That’s hilarious.

And yeah, nope, that won’t be me. We occasionally get dangerous weather and I don’t care how stupid my horse is acting, she’s getting her ass into the relative safety of the barn. The end. I am not willing to risk her life (or mine) because I have an idea in my head that it is terrible to set firm boundaries and expect respect for my space and my direction.

Oh and then let’s add the other end of the weather situation - horses inside for literally weeks due to dangerous ice in the fields they normally spend their day in. How is that first day of turnout going to go once the footing is safe? Are we going to hope and pray that we can get them safely across whatever ice remains on driveways and lanes or are we going to ensure that poor buggers absolutely will not question listening to their handler? I know where I stand on this, but I’d be very interested to know where others stand on it. And I do recognize that some mind bending may be involved - for those with horses that live out 24/7 transport yourself in your mind to a difficult climate and a horse with an injury that needs to stay in for several weeks/months but is otherwise healthy, happy and FIT.

[scislandsprite]

@Scribbler

@scislandsprite

I definitely agree that a significant amount of training is beyond one quadrant.

However, I am a little confused. With using a leg aid and always pairing it with a clicker, wouldn’t it eventually move into R+ territory since the horse would be moving forward for the reward, not to alleviate the leg aid? Very much a hypothetical situation since that’s not the most common method but would the leg aid be counter conditioned over time with consistent R+?

People combine mild R- training with rewards and that is perfectly all right.

There is no reason it has to be purist unless you are wanting to make a point about R+

I have found clicker training pure R+ to be really effective for tricks and groundwork on my mare. When she’s in the zone she will back up on voice command or with a wiggle of my finger. In the stall I usually have to touch her with my finger (light pressure). If we are having a bouncy event while handwalking, then I am going to use every ounce of my strength to keep her from knocking me over.

One of the puzzles to me is, what do people think the optimum outcome of NH or R+ is going to be? Horses have moods and opinions, and the more athletic they are the bigger those moods and opinions can be. You are never going to reach a place with any healthy horse where you are guaranteed that their moods and opinions are going to line up with your desires 100 per cent every time. Things that work nicely in an arena don’t work that way outside. You can do all kinds of liberty work in an arena but no one is going to walk a horse down the road without a halter and lead rope. (but you can train a dog to heel in public without needing a leash).

Anyhow, I think these training methods can implicitly sell a fantasy that somehow you can make a horse completely compliant to your every wish. And it’s not true. Working with horses is constant negotiation. And that’s fine! They have brains and needs and desires and instincts, they problem solve and are curious. That’s what makes them fun. We try to meet their needs as best we can but there are times we need to say, no here’s what I need.

I’m very aware that R- is alright! Nowhere in my replies did I say that everyone needs to use R+ and I think I mentioned a few times that I don’t even use R+ entirely. Honestly the vast majority of people that use R+ in their training do not use 100% R+, at least that I know of (can’t speak for every trainer/owner).

I mentioned in my reply being quoted that my question was very hypothetical and was out of curiosity. I don’t ride with a clicker, the majority of my riding is R-. I just wasn’t able to answer my own question about the counter conditioning since I’ve never counted conditioned while riding before.

[Scribbler]

@scislandsprite

IME R+ just doesn’t go that far for riding because you have to halt to treat

It is effective for standing at the mounting block and for teaching a fast halt. And back and some lateral work under saddle. But anything that requires forward gets hampered by R+

[danhelm441]

@Scribbler

IME R+ just doesn’t go that far for riding because you have to halt to treat

Not that I disagree R+ has its limitations, but it isn’t intended to be all treats. In theory at least you’re supposed to build up such a positive association with the click that the click itself is reinforcing. But on the other hand, my experience with praise and clicks to horses is that they so frequently associate it with stopping for a break or pausing for a treat that, if you don’t fork one over and in fact push them to keep moving forward, they get very distressed/confused and it undermines the positive association you were working so hard to build. I know I can’t be the only person who has said “good pony!” after nailing a transition, only to have the horse immediately drop to a walk/stop thinking praise means break time. It’s even worse with a click because then they’re eyeballing you patiently waiting for the promised treat.

[Scribbler]

@danhelm441

@Scribbler

IME R+ just doesn’t go that far for riding because you have to halt to treat

Not that I disagree R+ has its limitations, but it isn’t intended to be all treats. In theory at least you’re supposed to build up such a positive association with the click that the click itself is reinforcing. But on the other hand, my experience with praise and clicks to horses is that they so frequently associate it with stopping for a break or pausing for a treat that, if you don’t fork one over and in fact push them to keep moving forward, they get very distressed/confused and it undermines the positive association you were working so hard to build. I know I can’t be the only person who has said “good pony!” after nailing a transition, only to have the horse immediately drop to a walk/stop thinking praise means break time. It’s even worse with a click because then they’re eyeballing you patiently waiting for the promised treat.

@danhelm441

IME the power of the clicker is that it can indicate the exact moment the horse did the requested behavior, but the treat has to follow pretty fast. Yes you can build “extension” so for instance maresy will play nose soccer multiple times before I give her a treat now. But I can’t envision clicking say at every transition riding for 10 minutes and then giving a treat. Horses recall just doesn’t last that long

My R+ only friends claim there are amazing trainers online that have taught everything mounted by R+. When I look at these sites I don’t find them entirely convincing. I have a trick horse of my own. I know you can make things look any way you want especially on video.

I have never seen anyone IRL accomplish meaningful mounted work with pure R+ or indeed even produce a horse that is consistently safe to handwalk down the trail. But I have seen lots of people accomplish both these things with tactful R- and some use of treats.

I have experimented with R+ under saddle. Maresy and I are really good at it on the ground. It just does not translate unless you are teaching whoa.

[sascha]

@danhelm441

I know I can’t be the only person who has said “good pony!” after nailing a transition, only to have the horse immediately drop to a walk/stop thinking praise means break time.

Absolutely 100% guilty. It was a bit of a shock the first time it happened with my current because standing around is what you do when it’s sunny out and you’ve had your fill of hay.

[scislandsprite]

I think Hannah Weston from Connection Training rides with mostly R+? Although to be honest, I don’t ride with a clicker so my eye isn’t particularly developed for it under saddle. I believe there’s a few people out there that “blaze” the clicker so the horse doesn’t get a treat every click but again, I’ve never tried it so can’t speak from personal experience.

Part of me has always been hesitant to try it under saddle anyways; I’m probably subconsciously afraid my old Pony Club instructors will appear to scold me for feeding a horse while bridled. Nevertheless, I’m content riding with mostly R- as like many have mentioned, it can be done as mildly as you’d prefer

[Feathered_Feet]

There’s a lot of research (old learning & behavior stuff) demonstrating that intermittent reinforcement produces more robust behavior than continuous reinforcement (e.g., slot machines… they don’t reward every lever pull, only some lever pulls, and that uncertainty is part of why they are so addictive). So if you want robust behavior, it’s actually better to not treat every time the horse does xyz (to say nothing of the fact that then the first time you don’t treat, it can send them into a tail spin ). So, click for all good/wanted behaviors, but don’t treat for every single one - treat occasionally (use your judgement). (Full disclosure: I never quite believed that intermittent reinforcement produces more robust behavior than continuous as a blanket rule, but lots of old white men did lots of old experiments in the 50s, 60s, and 70s to demonstrate this. Personally, I think it’s probably more situation dependent than they’d like to admit, but who am I to argue with the literature?)

And as someone mentioned above, the beauty of clicker training is that the click becomes the reinforcer itself, independent of the treat. So yes, you can absolutely use clicker training under saddle because not every click means stop and get a treat. I’ve used it to great effect training my horse through FEI - I click with my tongue for a nice change, nice pirouette, nice half steps, etc. I’m not sure how other riders feel about me clicking away in the arena but they’ve never commented so But I am often vocalizing to give my horse guidance, whether its with a click or my voice so perhaps it all gets drowned out.

(Side note, I definitely do not do only R+ - I use a lot of R- also, because as someone mentioned above, it’s really hard to do any meaningful work with just R+. And frankly, I think R- is a perfectly good tool when used reasonably - horses use it with each other all the time, so why can’t I? The trick is to set up a learning/teaching environment where your horse knows a little pressure just means “you can safely keep trying answers and I’ll let you know when you try the right one”).

[Eclectic_Horseman]

Food rewards are useful to teach the horse how to learn in the very beginning ( i.e., when I give this cue, I want you to do this, and if you do you will be rewarded.) Then again food rewards are very useful for teaching NEW things-- a bridge reward (vocal, clicker whatever) is useful for more advanced training and it means “yes! That’s it! That’s what I wanted. Food reward coming later.” Sometimes just a kind word or a scratch is enough. Sometimes when being ridden just ending the training session and giving much praise is enough.

When horses have become trusting and eager to please, you don’t need to specifically reward the things that they already know.

[Ajierene]

@Feathered_Feet

There’s a lot of research (old learning & behavior stuff) demonstrating that intermittent reinforcement produces more robust behavior than continuous reinforcement (e.g., slot machines… they don’t reward every lever pull, only some lever pulls, and that uncertainty is part of why they are so addictive). So if you want robust behavior, it’s actually better to not treat every time the horse does xyz (to say nothing of the fact that then the first time you don’t treat, it can send them into a tail spin ). So, click for all good/wanted behaviors, but don’t treat for every single one - treat occasionally (use your judgement). (Full disclosure: I never quite believed that intermittent reinforcement produces more robust behavior than continuous as a blanket rule, but lots of old white men did lots of old experiments in the 50s, 60s, and 70s to demonstrate this. Personally, I think it’s probably more situation dependent than they’d like to admit, but who am I to argue with the literature?)

If you want to look more into this, it is commonly referred to Schedules of Reinforcement and BF Skinner paired this with his Operant Condition studies starting in the late 1930’s. It has been studied by psychologist since (and a bit before, but BF Skinner is considered the father of operant conditioning because of how prevalent and published his works were).

Interesting note - continuous reinforcement has the fastest extinction rate when reward is removed.

Variable ratio reinforcement has both the fastest response rate and slowest extinction rate.

[eightpondfarm]

@Scribbler

(but you can train a dog to heel in public without needing a leash).

i tried it. With my mule, on a long walk in the woods. with a few pears (it was a very good year for pears …they were everywhere in our yard) in my pocket. he followed or walked next to me when the trail widened for the entire time. Now, added, that particular woods and it’s attached pasture is one of the pastures the herd gets rotated into, so it wasn’t novel to explore. AND because they (all the horses here except whichever mustang i have in for training) are on pasture 24/7…the lure to just go off and graze just isn’t there. Just a walk in the woods, a girl and her mule.

[eightpondfarm]

@Scribbler

I have experimented with R+ under saddle. Maresy and I are really good at it on the ground. It just does not translate unless you are teaching whoa.

it KINDA does though. My tuned-in mare will, when effusively praised, turn her head in hopes of a bit of cookie in the ring. Sometimes she actually gets one too. I food reward while i ride. I have a treat pouch on the saddle.

[scislandsprite]

Ooh very interesting, thank you for the information!

Is variable ratio reinforcement alternating how often the behaviors are being reinforced?

[eightpondfarm]

@Ajierene

Let’s say you are teaching the horse to walk towards a cone (as someone stated before as a possible objective). You say “walk” and when the horse takes even one step, reward. But you want the horse to traverse the length of the ring towards that cone and that’s ay 30 paces. So, then you without reward (P-) until the next step in conditioning - say two strides and on and on until it will walk 30 to wherever the cone is. This step is P- because you are taking away the reward to decrease the behavior (walking on stride) but then you are moving into R+ when two strides are taken and you give reward.

Moving a horse toward an object. This is exactly what i’m doing right now with one of my guys. He’s learning to walk ‘on lead’ (which in his case is a 6" piece of twine on the halter’s leadring). In the beginning, he was rewarded for leaning. He was rewarded for coming toward the reward. To get him to take a step, again i lured him. Holding the carrot piece a little further away so he would need to take a step in order to reach it. With accompanying verbal praise at each juncture of course. To get him to take two steps (which did not occur until a couple of sessions later with him) i just held the carrot further away. And, when finally we were walking the carrot in my hand and me moved with him, not moving that lure any further away but eventually moving back toward him and in his reach. Finally, he is moving with me along and around inside the corral. Carrots are in my pocket and not held out as lure. He knows Good Boy means " " He will get a carrot coupled with good boy for going around the water hydrant which rather squeezes him … or through the muddy corner, or in a circle-round the feeder…etc. In other words, the rewards come when he does something new on the halter, more difficult (or so i think?). I’m not withholding a reward in any of these cases am i.

[eightpondfarm]

An error i made recently rewarding each step:
One of the mustangs that i’m aboard now needs to learn how to trailer (so i can haul him to lessons). It’s a ramp load. So…i rewarded the first step (piece ofcarrot&praise). He backed up and of course that was allowed. Lured him again for anothr first step. Step:carrot:praise. Lured for another step, which he rather pounded, and step:carrot:praise. Well, i am pretty sur that this is where ‘we’ misunderstood. Long-story-short…Because he then POUNDED each footfall!

Because it’s a steel trailer i’m not picky. He will eventually lighten up those footfalls…but for now, i think it might just be a good thing. One of the scary things about a trailer is the resounding sound of hooves…and the reverberation feel and echo sound. I am thinking that if he doesn’t object, and thinks this is the Real Way one must get aboard, he’s a bit ahead of the game at this point.

He now boards the trailer thinking that the louder the better. They’re so literal huh?!

[Feathered_Feet]

I did a PhD in this so I am all full up on reading old L&B studies, but yes others who are interested should definitely read up! There’s a lot that’s out there for sure

[Ajierene]

Variable Ration Reinforcement is when you reward at different intervals like on the 3rd, then 4th, then 2nd, interation of the desired behavior.

Fixed Ratio Reinforcement the number is fixe - ie every third time. This has a slightly lower repitition and slightly higher extinction rate.

Fixed and Variable interval Reinforcement give reward at certain time intervals - so every 5 min (fixed) or every 2 min, then 3 min, then 1 min (variable) not matter how many times the behavior is exhibited.

Variable Interval has a similar response/extinction rate as variable Ratio

Fixed interval has a response rate that speeds up closer to the interval time and slows down, speeds up, etc. For example - a food every 5 min no matter how often the lever is pushed. The rate pushes the lever once every 15 seconds from min 0-3, every 10 seconds from min 3-4, every 5 seconds from min 4 - 4.5, every 1 second for the last 30 seconds of the interval, but then goes back to once every 15 seconds after food delivered.

[eightpondfarm]

annnnnnd…you know what you can do is watch the horse and see when a reward is needed. they show you.

edit: how do they show you?
by their interest level.