Excrutiatingly Stubborn Dog

[Vickie]

Why do people over think simple concepts?

I think it's a form of procrastination for most

[Aidan3]

I have seen negative fallout many times between dog and handler, but the dogs have been subject to some heavy abuse and cruel treatment or have been handler sensitive dogs massively over corrected for their temperament level, but personally I think it's an over exaggerated aspect for the marketing of positive training methods for the most part perhaps.

What is your definition of "fallout", Petsitters?

This is not a judgment on you, but on all of us, I would be very surprised if you were able to identify most fallout, much less link it to training. It's just not that simple. If you develop some negative emotionality, avoidance, disinterest, loss of focus, "low drive" or ambivalence in a dog, how do we link that to single events, or worse, to recurring patterns of action by a dog owner over time?

It is extraordinarily difficult outside the lab. In the lab we can compare samples and hold all other variables constant. In the lab we can attribute effect to cause, in the real world we can rarely do that (although intuitively, we do it all the time, mostly incorrectly). In the real world there are too many factors, we make inferences where there are none, we see the things we are conditioned to see and miss the things we are unaware of.

No doubt everyone has seen this awareness test on YouTube? If you haven't, take a moment to watch, if you have, you will know what I am talking about (those who have seen it - NO SPOILERS!)

Edited May 30, 2011 by Aidan2

[Steve K9Pro]

K9: Thought I might throw a few thoughts into this one.

C: Personally, I was heartbroken when I finally saw the fallout I had caused in my previous dog with corrections. Absolutely devastated. It took me nine years to realise and I only did because I raised an animal without much compulsion at all and saw the difference. I tried to fix it and I couldn't. Old habits die hard. I don't consider it a scare campaign to try to communicate to others just how horrible it was when I realised what I'd done. I wouldn't wish that on anyone. That dog worshipped the ground I walked on and no one but me could see what I was so upset about, but that didn't help me at all. It can be very subtle and insidious.

K9: And this just may be the way that you conducted the training. something specific with that dog that it may have displayed those behaviors you see, and I accept your other dog doesn't display these signs, but 2 dogs in one household isn't really enough to solidly make an informed decision.

I will say though that fall out is real, can certainly happen and still does today, but because it can happen doesn't mean it always will with every dog and every trainer.

As for Paul Hutton, I don't know him but he is like many I do know, they have a method that works for them, at a point perhaps in their career that they don't want to start again with something new, and are sick of being bashed by people who say if you correct a dog it will ruin it. I am sure he gets good results with what ever method he uses, but it is fair to say that, dogs have evolved in my short life time and we need to as well.

C: I think we absolutely need to work within the temperament of the dog, but I don't think this needs to rule out rewards.

K9: 100% agree, rewards should never be ruled out, I firmly believe rewards are one of the foundations of good behaviour.

A: Steve is a brilliant trainer, but how does this support your argument that "in the pet dog training market that a balanced combination of all training methods works best"? You are comparing a rare talent against a lowest common denominator, and using the opinion of this forum (one of thousands around the world, and not a representative sample) to support your claim. Some forums recommend Cesar Millan, others recommend Susan Garrett, others recommend Uta Bindels. It depends on what people are training for, and who is the brilliant trainer closest to them. For example, if you ask on this forum who to see in WA, you will more than likely get the recommendation of one of several no-compulsion trainers.

PS: Steve's link I posted supports my argument Adian, so Steve is wrong in that case?

C: I think the point Aidan is making is that great trainers are the best because of their talent, not the methods they use. Given the best trainers do not all use the same methods, then the argument that the methods those trainers are using is what creates their success not only appears to be inaccurate, but I think diminishes the skill of those trainers.

K9: I think the best trainers have instinct and optimism and can create both of these assets in the dogs they work with. Some may use a treat, some a prong collar, doesn't matter to me which as long as that person is not solely the one or the other due to politics.

A: Steve is a respected friend of mine, but I do not agree with some points in that article even though I "have a tool box with many options" and agree with many of the other points he makes.

K9: Thanks Aidan, likewise, I will refer people to other trainers because I believe they can help, "how" they help is not important to me, just that they do, and Aidan does, so the hows, whats and whys are of no real importance to me.

I also have been asked to write a book by many including a few serious offers, but just like the article in question, my opinion slowly changes it's slant over time & I feel the book will be outdated by the time it is read.

A: However, your claim was made on the basis of popular opinions from this forum and the history of results from Steve and Mark. This ignores every purely positive trainer who also has a brilliant history of results, the opinions of every other forum in the world, and the fact that sometimes this forum recommends purely positive trainers if they are in the area.

It also ignores the fact that Mark and Steve do not agree on many things.

K9: lol very true about Mark and I.

Edited June 7, 2011 by K9Pro

[PetSitters]

Also, choosing to use a quadrant that involves compulsion does NOT always mean the trainer 'did not want to build foundations'. And compulsion and positive reinforcement are no mutually exclusive for some trainers.

Hey, I didn't mean that. I said I didn't think claims that building a reward system was too difficult and complicated for some dogs was a reason to not use rewards with them. It's a cop out. I don't think that means there is no reason to not use rewards and I do think it's the owner's call and I don't really care what call they make. But if they think they can't build a reward system because the dog is too hard, I challenge that.

I don't know how many times I have to point out that I have used punishments for it to sink in, but I'll point it out again. I find it rare that I really want to suppress a behaviour. I don't think it's conducive to my aims with my dogs to regularly suppress behaviour. I am well aware that that is my personal choice and I do not think that all positive reinforcement trainers should have the same aims. I thought I made it clear that I think it's a personal choice and doesn't really matter in the scheme of things.

Corvus, some dogs are extremely difficult to build a reward system strong enough to support distractions and it amounts to how dedicated someone is to achieve this and most are not with difficult dogs so when people want certain behaviours suppressed or extinguished they need to make a choice, often compulsion is the better option for their situation.

My point to the discussion is when people hire a trainer who will not apply and teach complusion methods with a difficult dog owned by people not dedicated enough to follow motivational routines to reshape their dogs behaviour, the dog ends up permanently in the back yard and the owners can't be bothered with it other than feeding.

The dog may not have suffered a leash correction, but it hasn't been for a walk for 2 years, so which evil is worse, a bit of compulsion training on the leash or a lifetime in the back yard is what I am talking about?

A trainer is hired to train a dog in my opinion not pedal a method and I think they should be obliged to assess the big picture to deal with the behaviour effectively and take into account the owners ability to carry out instruction in the best interests of the dog's future and shape the dog's behaviour so the owners receive some pleasure from the dog.

In the case of the Rotty on a harness I mentioned as an example was the result of a purely positive trainer who effectively relegated the dog to the back yard as the owner couldn't control it on leash because "they don't believe in prong collars"?. If they used a full box of tools with the ability to assess the situation with this dog and owner properly, they could have trained with a prong as I did in the first place and reunited with this owner the pleasure of walking his dog for some quality companionship and most importantly provide exercise for the dog.

[Steve K9Pro]

PS: What I disagree with regarding purely positive trainers is limiting resources to train on the one size fits all basis because not two dogs in temperament and drive are the same.

K9: I agree that this does happen, but I don't think it happens just on one side of the fence, I see many trainers do this from both sides of the "religion".

I also think that training without compulsion is the only way to achieve certain things with certain dogs. All of my Training in Drive programs are non compulsive, there are no correction collars, e collars, no harsh verbal tones,. nothing at all that will apply any type of aversive to the dog other than removal of the reward.

At times drive has attained a level of destructiveness or self destructiveness that remaining positive is almost counter productive to the outcome, not primarily the outcome of the dog but the outcome of the situation. That includes the handler, dog and environment.

Edited May 31, 2011 by K9Pro

[corvus]

This is not a judgment on you, but on all of us, I would be very surprised if you were able to identify most fallout, much less link it to training. It's just not that simple. If you develop some negative emotionality, avoidance, disinterest, loss of focus, "low drive" or ambivalence in a dog, how do we link that to single events, or worse, to recurring patterns of action by a dog owner over time?

I'd be surprised if I were able to identify most fallout as well.

[corvus]

My point to the discussion is when people hire a trainer who will not apply and teach complusion methods with a difficult dog owned by people not dedicated enough to follow motivational routines to reshape their dogs behaviour, the dog ends up permanently in the back yard and the owners can't be bothered with it other than feeding.

Ohhhh, owned by people not dedicated enough to follow motivational routines. Well that's quite specific, isn't it? Why didn't you say!

A trainer is hired to train a dog in my opinion not pedal a method and I think they should be obliged to assess the big picture to deal with the behaviour effectively and take into account the owners ability to carry out instruction in the best interests of the dog's future and shape the dog's behaviour so the owners receive some pleasure from the dog.

I am still not convinced that this a methodology problem rather than a trainer skill problem.

In the case of the Rotty on a harness I mentioned as an example was the result of a purely positive trainer who effectively relegated the dog to the back yard as the owner couldn't control it on leash because "they don't believe in prong collars"?. If they used a full box of tools with the ability to assess the situation with this dog and owner properly, they could have trained with a prong as I did in the first place and reunited with this owner the pleasure of walking his dog for some quality companionship and most importantly provide exercise for the dog.

I have zillions of anecdotes about dogs that would actually manage to get out of the yard if their owners would stop smacking them every time they growled at another dog. Methods and tools are just things we pick up to help us train a dog. They can be used and misused by anyone. I don't think stories about a PP trainer that couldn't do this or a correctional trainer that couldn't do that is very indicative of the effectiveness or lack thereof of one method compared to another. I think it's fair to expect that a good trainer will be able to achieve good results with the minimum side effects regardless of what method they use.

[PetSitters]

My point to the discussion is when people hire a trainer who will not apply and teach complusion methods with a difficult dog owned by people not dedicated enough to follow motivational routines to reshape their dogs behaviour, the dog ends up permanently in the back yard and the owners can't be bothered with it other than feeding.

Ohhhh, owned by people not dedicated enough to follow motivational routines. Well that's quite specific, isn't it? Why didn't you say!

A trainer is hired to train a dog in my opinion not pedal a method and I think they should be obliged to assess the big picture to deal with the behaviour effectively and take into account the owners ability to carry out instruction in the best interests of the dog's future and shape the dog's behaviour so the owners receive some pleasure from the dog.

I am still not convinced that this a methodology problem rather than a trainer skill problem.

In the case of the Rotty on a harness I mentioned as an example was the result of a purely positive trainer who effectively relegated the dog to the back yard as the owner couldn't control it on leash because "they don't believe in prong collars"?. If they used a full box of tools with the ability to assess the situation with this dog and owner properly, they could have trained with a prong as I did in the first place and reunited with this owner the pleasure of walking his dog for some quality companionship and most importantly provide exercise for the dog.

I have zillions of anecdotes about dogs that would actually manage to get out of the yard if their owners would stop smacking them every time they growled at another dog. Methods and tools are just things we pick up to help us train a dog. They can be used and misused by anyone. I don't think stories about a PP trainer that couldn't do this or a correctional trainer that couldn't do that is very indicative of the effectiveness or lack thereof of one method compared to another. I think it's fair to expect that a good trainer will be able to achieve good results with the minimum side effects regardless of what method they use.

I don't think the individual dog and the owners situation is taken into account enough sometimes when a trainer decides what training approach/methods to apply to reshape the behaviour the owner requires. A good trainer in my opinion should be open to all methods with a personal inspiration to gain the skill of determining which training method is best suited to a particular situation and dog and with forethought to believe the owner is capable of conducting the chosen routines to achieve the desired results. Method peddling either purely positive or compulsion applied to every dog and situation is an error in the making as not every dog has the same temperament and drives.

Purely positive trainers and the example given of the guy in New Zealand to me are "method peddlers" given a dog to train that doesn't respond well to the methods they pedal are more likey to stuff it up. I much prefer the concepts of Steve Courtney, Mark Singer and others who don't method pedal and operate with a tool box full of tricks to suit the widest varity of dogs and situations. I don't think there is any place for a commercial trainer apsiring to be a good one should rule out methods to make their training concepts "religion" based

Edited May 31, 2011 by PetSitters

[PetSitters]

I have seen negative fallout many times between dog and handler, but the dogs have been subject to some heavy abuse and cruel treatment or have been handler sensitive dogs massively over corrected for their temperament level, but personally I think it's an over exaggerated aspect for the marketing of positive training methods for the most part perhaps.

What is your definition of "fallout", Petsitters?

This is not a judgment on you, but on all of us, I would be very surprised if you were able to identify most fallout, much less link it to training. It's just not that simple. If you develop some negative emotionality, avoidance, disinterest, loss of focus, "low drive" or ambivalence in a dog, how do we link that to single events, or worse, to recurring patterns of action by a dog owner over time?

It is extraordinarily difficult outside the lab. In the lab we can compare samples and hold all other variables constant. In the lab we can attribute effect to cause, in the real world we can rarely do that (although intuitively, we do it all the time, mostly incorrectly). In the real world there are too many factors, we make inferences where there are none, we see the things we are conditioned to see and miss the things we are unaware of.

No doubt everyone has seen this awareness test on YouTube? If you haven't, take a moment to watch, if you have, you will know what I am talking about (those who have seen it - NO SPOILERS!)

My interpretation of fallout beyond most people's immediate response I would imagine to be a damaging bond between handler and dog, fallout to me is anything suppressing desired behaviour as a side effect of training errors. I think fallout that can only be determined in a lab environment unseen by the naked eye is getting a bit over techo and too miniscule to be overly concerned about in general obedience training IMHO.

[Cosmolo]

That video was so cool Aidan- i was so pleased with myself that i had got it right but totally missed the 'thing'.

[Cosmolo]

WRT building a reward system being too hard- in my experience its not that its too hard so shouldn't be done. Its that in addition to building the reward system the owner NEEDS some immediate suppression for some reason.

Area demographics come into play here too- it is very different rehabilitating a dog aggressive dog in an area with a lower dog population vs one where the owner steps out onto a narrow street and encounters a dog too close, too fast. Trainers do need to take this into account as well.

[Aidan3]

I think fallout that can only be determined in a lab environment unseen by the naked eye is getting a bit over techo and too miniscule to be overly concerned about in general obedience training IMHO.

In that case I haven't explained myself very well

It's not that the effect is miniscule, it's that it's difficult to attribute problems the dog may have to something that you have caused. It could be a very large effect, but you might blame genetics ("weak nerve"), for e.g, or "low drive", or a critical period - not realising that it was something you did in training or behaviour modification. It happens every single day.

There are literally thousands of reactive dogs out there who, while having some genetic predisposition, were made that way through handling. Even more dogs who lag, forge, become over-excited, lose focus easily or whatever - because they have been made that way. That's fallout too.

Or you might simply miss the problem. I watched a person make a dog aggressive under instruction using corrections. Both instructor and handler were convinced that they were doing the right thing, but to anyone watching it was an astounding display of abuse through incompetence. They were seeing what they were looking for and missing everything else (see the YouTube video I posted for a remarkable demonstration of missing obvious things when we're focusing on other things).

The point about the laboratory was that only under true experimental conditions can we say with any confidence that we are correctly attributing cause and effect, but if you can't really see what I'm on about with cause and effect, that's OK, it's not a big deal. We attribute effect to cause intuitively all the time, it's human nature, but we're frequently wrong. We blame genetics, drive, nerve, critical periods, focus, softness, sharpness etc when we don't really know for sure that those things are responsible.

[Kavik]

Area demographics come into play here too- it is very different rehabilitating a dog aggressive dog in an area with a lower dog population vs one where the owner steps out onto a narrow street and encounters a dog too close, too fast. Trainers do need to take this into account as well.

Completely agree with this! I had much more success in improving the behaviour and it was easier and less stressful with my dog aggro dog when I moved to a less populated area.

[PetSitters]

I think fallout that can only be determined in a lab environment unseen by the naked eye is getting a bit over techo and too miniscule to be overly concerned about in general obedience training IMHO.

In that case I haven't explained myself very well

It's not that the effect is miniscule, it's that it's difficult to attribute problems the dog may have to something that you have caused. It could be a very large effect, but you might blame genetics ("weak nerve"), for e.g, or "low drive", or a critical period - not realising that it was something you did in training or behaviour modification. It happens every single day.

There are literally thousands of reactive dogs out there who, while having some genetic predisposition, were made that way through handling. Even more dogs who lag, forge, become over-excited, lose focus easily or whatever - because they have been made that way. That's fallout too.

Or you might simply miss the problem. I watched a person make a dog aggressive under instruction using corrections. Both instructor and handler were convinced that they were doing the right thing, but to anyone watching it was an astounding display of abuse through incompetence. They were seeing what they were looking for and missing everything else (see the YouTube video I posted for a remarkable demonstration of missing obvious things when we're focusing on other things).

The point about the laboratory was that only under true experimental conditions can we say with any confidence that we are correctly attributing cause and effect, but if you can't really see what I'm on about with cause and effect, that's OK, it's not a big deal. We attribute effect to cause intuitively all the time, it's human nature, but we're frequently wrong. We blame genetics, drive, nerve, critical periods, focus, softness, sharpness etc when we don't really know for sure that those things are responsible.

I do understand Aidan exactly what you describe, and of course the big question, are these behaviours made which is possible for sure. but I guess to what extent is the essesnce. Aggression can be triggered and escalated in some dogs from corrections and own one like that myself, but having said that, he is highly food and toy motivated and biddable and to date have never found a reason to consider a physical correction with this dog

[JoeK]

What is happening most of the time is people make the booking for trainer when the dog hes out of control which is good becuase the owner realise something needing to be done for the dog, so is good beginning. The problem with the purely positive training when the dog behavior out of control is often not the best method to get rid of bad behavior fast enough and the dog still practice the bad behavior which is no good for some dogs and not fixing him properly.

So you needing best a trainer with full box of tools and knowledge of many training method not just couple of the tool and one method is not rocket science for understanding that the more tool and more knowledge of training method the better of course. So when you getting trainer for problem dog that only do one method then the method he do maybe not best method for fixing particular dog.

Sometimes the fallout is good to suppressing bad behavior, doesnt matter if he fell out if the behavior is getting fixed like if dog is scared of correction and stop aggressive lunging is good result on the pet dog.

Joe

[sas]

Area demographics come into play here too- it is very different rehabilitating a dog aggressive dog in an area with a lower dog population vs one where the owner steps out onto a narrow street and encounters a dog too close, too fast. Trainers do need to take this into account as well.

Completely agree with this! I had much more success in improving the behaviour and it was easier and less stressful with my dog aggro dog when I moved to a less populated area.

Same here.

[Mrs Rusty Bucket]

Sometimes the fallout is good to suppressing bad behavior, doesnt matter if he fell out if the behavior is getting fixed like if dog is scared of correction and stop aggressive lunging is good result on the pet dog.

The trouble with fallout - is that it suppresses the good behaviour. Ie it's what you didn't intend to train, by the dog associating the wrong behaviour with the correction. And given the potentially traumatic nature of correction - it maybe more difficult to fix the wrong association made using punishers, that it is to fix a wrong association made using rewards.

If you're using adversives to stop a bad behaviour and that works - then it isn't "fallout". It's a semantics/word definition thing.

Egs of fallout (ie what dog learns is not what owner intended):

Scolding the dog for barking. Dog thinks owner is joining in and barks more.

Scolding and hitting a dog for growling at another dog. Dog learns to stop growling, but will attack directly instead. Ie teaches the dog to attack with no warning.

Scolding and rubbing a dogs nose in it for toiletting inside the house. Dog learns to hide when it goes to the toilet. So owner doesn't find it. And dog may just learn to hide from owner.

Alpha rolling a dog. Dog learns to fear handler and may opt for the direct attack without warning especially if alpha rolled for growling.

Scolding and jerking by the neck, a dog that barks at something unusual in the park. Dog learns to associate pain with the something unusual. And may attempt to attack the something in order chase it off to avoid the pain.

[JoeK]

Sometimes the fallout is good to suppressing bad behavior, doesnt matter if he fell out if the behavior is getting fixed like if dog is scared of correction and stop aggressive lunging is good result on the pet dog.

The trouble with fallout - is that it suppresses the good behaviour. Ie it's what you didn't intend to train, by the dog associating the wrong behaviour with the correction. And given the potentially traumatic nature of correction - it maybe more difficult to fix the wrong association made using punishers, that it is to fix a wrong association made using rewards.

If you're using adversives to stop a bad behaviour and that works - then it isn't "fallout". It's a semantics/word definition thing.

Egs of fallout (ie what dog learns is not what owner intended):

Scolding the dog for barking. Dog thinks owner is joining in and barks more.

Scolding and hitting a dog for growling at another dog. Dog learns to stop growling, but will attack directly instead. Ie teaches the dog to attack with no warning.

Scolding and rubbing a dogs nose in it for toiletting inside the house. Dog learns to hide when it goes to the toilet. So owner doesn't find it. And dog may just learn to hide from owner.

Alpha rolling a dog. Dog learns to fear handler and may opt for the direct attack without warning especially if alpha rolled for growling.

Scolding and jerking by the neck, a dog that barks at something unusual in the park. Dog learns to associate pain with the something unusual. And may attempt to attack the something in order chase it off to avoid the pain.

Yes,this can happening when the handler is not knowing properly when to apply correction is same as smaking the dog when he come to you after bad behavior is not how is done, but the falling out from incorrect application is handlers fault not the method. Is like extinction method on the purely positive they doing at the dog park and what happens is you ignore the bad behavior and rewarding the good, so when the dog attacking other dogs in the park you ignore when he attack and reward him when he doesnt is good method that one yes?

So I taking big nasty mongrel to the dog park for practice the extinction training and when he nail everyone elses dog and they yell to me to get the mongrel out from park, I tell them no, we do extinction training from the purely positive book and ignore the bad behavior to prevent the falling out then everyone should be accepting my training method yes?

Joe

Edited June 3, 2011 by JoeK

[bedazzledx2]

JoeK I really don't think you understand the complexities of positive training. Management plays a very big part as does not being permissive. An aggressive dog should never be let loose in a dog park, positive or not. There are always consequences but to say positive training is all about ignoring the bad behaviour and rewarding the good is over simplistic.

Sometimes the fallout is good to suppressing bad behavior, doesnt matter if he fell out if the behavior is getting fixed like if dog is scared of correction and stop aggressive lunging is good result on the pet dog.

The trouble with fallout - is that it suppresses the good behaviour. Ie it's what you didn't intend to train, by the dog associating the wrong behaviour with the correction. And given the potentially traumatic nature of correction - it maybe more difficult to fix the wrong association made using punishers, that it is to fix a wrong association made using rewards.

If you're using adversives to stop a bad behaviour and that works - then it isn't "fallout". It's a semantics/word definition thing.

Egs of fallout (ie what dog learns is not what owner intended):

Scolding the dog for barking. Dog thinks owner is joining in and barks more.

Scolding and hitting a dog for growling at another dog. Dog learns to stop growling, but will attack directly instead. Ie teaches the dog to attack with no warning.

Scolding and rubbing a dogs nose in it for toiletting inside the house. Dog learns to hide when it goes to the toilet. So owner doesn't find it. And dog may just learn to hide from owner.

Alpha rolling a dog. Dog learns to fear handler and may opt for the direct attack without warning especially if alpha rolled for growling.

Scolding and jerking by the neck, a dog that barks at something unusual in the park. Dog learns to associate pain with the something unusual. And may attempt to attack the something in order chase it off to avoid the pain.

Yes,this can happening when the handler is not knowing properly when to apply correction is same as smaking the dog when he come to you after bad behavior is not how is done, but the falling out from incorrect application is handlers fault not the method. Is like extinction method on the purely positive they doing at the dog park and what happens is you ignore the bad behavior and rewarding the good, so when the dog attacking other dogs in the park you ignore when he attack and reward him when he doesnt is good method that one yes?

So I taking big nasty mongrel to the dog park for practice the extinction training and when he nail everyone elses dog and they yell to me to get the mongrel out from park, I tell them no, we do extinction training from the purely positive book and ignore the bad behavior to prevent the falling out then everyone should be accepting my training method yes?

Joe

[Mrs Rusty Bucket]

Yes,this can happening when the handler is not knowing properly when to apply correction is same as smaking the dog when he come to you after bad behavior is not how is done, but the falling out from incorrect application is handlers fault not the method.

I agree with this, Joe.

It's my opinion that it is harder for a hander to get the punishment/correction method right. To get the timing right. That's why it is spectacular when you watch some of the top trainers, because their timing is so good for all types of reward and punishment combinations. But I'm not that good so I keep the corrections to a minimum. And I don't want scolding my dog to be a first option (because some of the training seminars I go to don't allow), so I try to practice other things. I have been known to say "Uh" at my dog when she goes to lick toast crumbs off my plate without permission. Sigh. It does stop her.

so when the dog attacking other dogs in the park you ignore when he attack and reward him when he doesnt is good method that one yes?

LOL. Nope. No Way. If dog distracted and unhappy about other dogs in the park, I'd work at a distance where the dog was paying attention to me and not the scary dogs he wants to fight, and reward him paying attention to me.

And then gradually I'd work on the edge of his distraction level ie moving closer to the distracting dogs then away then closer bit by bit... And we wouldn't go in the park until he could reliably pay attention to me, without being distracted or upset by the other dogs at all. Ideally I'd do some training with a grovel dog like mine, as distraction dog - to get the aggressive dog used to the idea that he doesn't need to attack all other dogs.

Ie I wouldn't ignore him attacking. I'd limit his opportunity to do so - to zero. And I wouldn't scold him for showing upset, because that's good feedback for the trainer. Even the dogs that have been trained not to growl before they attack, first show tiny signs of discomfort. They fidget, their hackles go, they fixate on the other dog (ie get distracted). All this you want to pay attention and it means you're working too close.

If dog wanted to eat me, I'd invite an expert in people aggressive dogs to help. That's way beyond anything I want to deal with.