For five days a trainer offers Bullet every toy he has, and gets nothing. On day six, inside a bout of hand-fighting, Bullet takes a tug in his mouth for the first time. Four days later the game is gone and the trainer stops the session on camera to rebuild it. On a long walk in a public park he lets Bullet go to sniff and says what the footage has been showing all along: for this dog, freedom beats play, 'and I think it might always be'. The dog decided what paid him, and the program went on with that.
into personal play oh there we go first time he took it in his mouth yeah good job dude good job
uh because i had him playing tug really hard before and i want to get that hard tug playing
much about anything else. Treats, meh. Playing, that's all right. It's pretty good. But at the level of play that I've been able to get him to engage in so far, this right here is far superior. And I think it might always be.
The chapter in one paragraph
Pressure and release only teach if the release is worth having, so before the method can press on a dog it has to know what pays him. The dog decides that, not the trainer. The program offers toys, touch, food and freedom and watches which he works for, and it checks the toy, the heat and the handler before concluding that a dog 'doesn't play'. Its default currencies are two. The first is play: a real contest for possession that the dog gets to win, with the command as the way he turns the game back on. The second is freedom: the end of the command and permission to go back to sniffing, exploring or whatever he was doing. Most of an ordinary walk belongs to the dog, and the work is dropped into that freedom. It does not replace it. Food is not banned. It works only while nothing in the room matters more, and a dog paid in food every time learns to watch the hand, so the house position is play first and food case by case. Play can be grown in a dog who does not have it, though the footage is honest that it does not always arrive. Once the job is known, the reward becomes 'might' rather than 'will', and it never drops to nothing.
Principles
The dog decides what counts as a reward. Offer toys, touch, food and freedom, watch which one he works for, and build on that. Petting a dog who dislikes it can punish him instead of paying him.
Said as a rule on filmOn film
So, in that case, it's not a reinforcement. If I pet a dog that hates being pet after they do something, that could be a positive punishment.
right and so that's how it works because it's not necessary every dog kind of chooses what's a reinforcement and what's a punishment
This is why you should always take some stuff with you and try it, right? Because it can totally change your approach. Try tug, try ball, try stuff.
ball yeah we'll give it like a like a three out of ten floor pole we'll give it like a five out of ten tug we're gonna give that like an eight or a nine out of ten
Concepts: The dog tells you the reward
Exceptions
- A failed toy may simply be the wrong toy. Bullet ignored soft toys and then gripped a solid one hard, 'because he's got a much stronger bite' (watch 06:16). Heat changes the answer too (Ramsey, watch 08:57).
- Touch is a reward only if this dog wants it. For a frightened dog who does not trust the handler, petting and praise can be aversive, so they are held back (Kodak, watch 09:30; reward-access-first).
Games are pieces of the hunt: stalking, chase, grab, the fight for possession, the win. Find the piece this dog is buying and give him more of it.
Said as a rule on filmOn film
All of these games Imitate different parts of the prey cycle. That's why dog goes so nuts for them, right?
figure out which part of the play they like the most and use that to its advantage
The way that play works is there's like a cycle so there's six parts of the cycle, right?
Concepts: The dog tells you the reward
Exceptions
- Nobody explains why a dog prefers one piece, or whether preferences stay stable. Stormi's trainer expects hers may shift (watch 07:08).
The reward for listening has to beat the reward for not listening. There is always a reward: the one the handler gives, or the one the dog takes.
Said as a rule on filmOn film
I heard somebody say it was, you have to make sure the reward for listening is greater from the reward for not listening.
was taking a while back and and he said uh make sure that all of your commands the reward for them listening to you is greater than their reward for not listening
psychology of a dog's brain. Their brain usually works with like as far as commands go okay command, do I want to do this or is there something I want to do more?
Concepts: The dog tells you the reward, Food facilitates calm, it cannot create it
Play pays as a real contest for possession, one the dog is allowed to win. The toy is valuable because the handler wants it too. Waving a toy about is not play.
Said as a rule on filmOn film
In dog training, play is God. Like, that is, like, the thing that makes everything else possible. If you can get a dog to play, you won.
He doesn't want me to have it now. Yeah. Right. Why is it important to him? Why is the toy important to him now when it wasn't before? Because he's I want it. I want the toy. And he can tell I want the toy.
It's like um it's like if you're way put a video game in somebody, if the first time they ever play, if you just crush them and then you crush them over and over and over again, will they want to play more? No. You let them win the first time,
We're gonna use that for all the stuff. He learns so much faster. He's so much more cooperative. It builds a lot of trust. And it really fits within the paradigm that we train in.
Concepts: Play becomes the reward - and play is a contest
Exceptions
- The claims that play teaches faster are handlers' judgements, not measured comparisons.
- In protection work the win is priced by the criterion: 'as soon as she bit the whole sleeve and wouldn't let go for a minute, then she got access to keep it' (Dolce, watch 05:33).
The command sits inside the game. The dog is asked for a sit or a down while the game runs, and his reward is that the game turns back on. The out that ends a round is information, not punishment.
Said as a rule on filmOn film
You see as soon as he finally did it, he got the toy. See, now he learned how to turn it back on. Oh, you can turn the game back on whenever you want. When I ask him, you can get your game back.
there's breaking the game kind of like a boxing match and in this Bell nobody's mad that the bell got wrong right because this is a clean interruptions a break you asked him to do some work to turn the game back on and his reward is that the game turns back on
integrate the the play right and so the out is not punishment yeah it's just information
really like proof point so the main thing with teaching it is that everything you teach her should be while you're playing
Concepts: Play becomes the reward - and play is a contest
Exceptions
- The toy can go straight back on 'Ok!', as Apollo the Sharkie's trainer does to keep the out from predicting the end of the fun (watch 03:22).
A game can be ruined. The trainers name the ways: too much pressure inside it, too many behaviours, taking the toy away every few seconds, a command before every throw, a game that is predictable, and a handler who is not really in it.
Said as a rule on filmOn film
about less is more oh yeah and then like one of the other things that like the biggest thing is just like not taking away the tug every 15 seconds like some people want to do where it's like all the time make it a real game
gonna do like uh you know and for every throw you must sit or you must down or whatever that's too much right if you're doing that your dog's gonna be like you're gonna see it in their face
And so you think you never want it to be a predictable game. Games that are predictable, they die.
So, the dogs feel like, they know you're in it. And so if they feel like you're not in the game with them, then it doesn't work very well at all.
Concepts: Play becomes the reward - and play is a contest
Exceptions
- The '45 seconds to a minute' of play per round is one handler's working figure, not a house standard.
Play is a capacity that can be grown. Before calling a dog a non-player, ask whether he is missing a game or missing a relationship with the person offering it. A dog who does not yet trust you will not compete with you.
Said as a rule on filmOn film
training why because ideally for a training scenario you need to have a relationship with the dog you need to have trust with the dog they need to like you and they need to know that you like them right um otherwise how are they going to ever play with you
Maybe they don't want to compete with you directly yet.
do that yeah quit licking my legs so i just think it's lack of exposure like most of the time they learn that as puppies like you give them a bucket of water you let them play in the water and you just like leave it alone type thing um i just don't think he ever had that
Don't just assume that, oh, well, you know, they're kind of like fearful or they're kind of nervous so they won't do it. You don't know that. They might do it.
this is good the play will just get stronger and stronger after this now that he's finally going for the tug i just got to keep raising that capacity for him and his obedience is going to come along so much faster with that
Concepts: Play is a capacity you can grow
Exceptions
- The record is honest about limits. The fast turnarounds on film were dogs who already had a game and needed a person they trusted. None of the dogs building play from nothing reached a lasting toy game on camera (Bullet, Bo, Callie, Egg), and all of them were trained anyway, paid with freedom, sniffing and contact.
- Bullet's trainer, after the tug came and went: 'at the level of play that I've been able to get him to engage in so far, this right here is far superior. And I think it might always be' (watch 10:09).
Release to freedom: pay the dog with the end of the command and permission to go back to what he wanted, whether that is the smell, the tree line or the open field.
Said as a rule on filmOn film
There's actually plenty of positive reinforcement going on here, because what is it that Louie wants to do instead of whatever it is that I ask him to do? Well, what he's doing currently, which is going around and sniffing everything.
release to freedom is king, it is the king, the king of the dog training, release to freedom and the tug, the playing of the tug, it is the king of the dog training
you don't need treats because the positive reinforcers whatever they're sniffing right
than what I was just doing. If you have good timing, you know how to do it. You can train them literally with just negative reinforcement and a release to freedom
much about anything else. Treats, meh. Playing, that's all right. It's pretty good. But at the
Concepts: Release to freedom
Exceptions
- It works only on a dog who wants to be released: 'it sucks when they don't when they're not like receptive to that though because then you're like i gotta do it a different way' (Connor, Deuce, watch 04:12). Nobody gives a test for 'receptive'.
- The same handler also withholds freedom as pressure. With Luna he locks the leash so she cannot explore when she is still fighting a position (watch 18:14). This is carried under conflicts.
The release has to mean something. It is said with excitement, the dog actually goes, and the equipment has to leave somewhere to release him to. A recall is paid by sending the dog back to what he left, so coming back reopens his world and does not close it down.
Said as a rule on filmOn film
And I throw a little party, that makes the release cue more powerful, right?
doing it right is when they get they go from being obedient to being like incredibly amped
of them is if you say okay you want to make sure that he goes and does something because you don't
like that's no good right and so that's one of the problems with like those short leashes um i don't use them that often because of that like i can't reward them the same way that i can with this
Is there anything wrong with you sending her back to where the squirrel was? No, right?
and so then what what is it what is your role end up being in the walk restrainer restrainer exactly restrainer um and so then you're not a source of freedom you're a source of restraint
Concepts: Release to freedom, Meaningful release
Exceptions
- He does not have to come to the handler. The release is given by the word alone, though movement may be added while the dog is still learning the word or seems unsure.
Start with freedom. The first minutes of a session, and most of an ordinary walk, belong to the dog. The work is dropped into that freedom, in short pieces, and each command is paid by sending him straight back out.
Said as a rule on filmOn film
So for the first few minutes, I'm still letting him do his thing. Maybe give him a command or two.
to start with the sniffies. Don't get them out into obedience right away. Let them acclimate
notice that like kind of how we're doing the walk is it's very she's i'm not telling her to do anything um and that's like probably a good 90 of every walk will just be like her doing stuff
say is get used to giving him way more freedom than you think you should
That does not mean that we don't use negative reinforcement or we don't require them to do things, or that we don't create boundaries or penalize things, it just means that there's an order of operations.
Concepts: Start with freedom
Exceptions
- The freedom a dog can be given is bounded by his recall. Poseidon is let wander because his recall and 'let's go' are reliable (watch 06:20).
- Both 'ninety per cent of the walk' and 'the first few minutes' are right, depending on the dog and his skills. For most untrained dogs, start somewhere quiet.
Restraint makes the problem it was meant to stop. A dog kept on a short lead with pops and 'no' builds frustration, and a dog who is never given the freedom takes it by pulling.
Said as a rule on filmOn film
just put a bunch of you know restriction and so you can see what that does to the dog when i'm always restricting it it builds frustration and then they end up freaking out and barking right
created like this resistance we're now like oh well why are they holding me back i want to go
if you don't practice that uh them being in that window on a regular basis they won't make the right choices because they've never had to try
but with the flexi and having training There's not any confrontational nature. You're not a resistor or a restrainer of the dog.
Concepts: Start with freedom
For a nervous dog, the freedom to investigate is part of the treatment. A dog who is investigating cannot be in fear or panic at the same time. Exploration is ranked the top stress reducer.
Said as a rule on filmOn film
This is a big secret. If you let the nervous dog go and just like step around and do its thing, it's open up nerves, which for her is going to be big.
once you get them investigating then they can't be in these other mindsets these other modes so
environmental exploration right that's king and then negative reinforcement right with the release the freedom right back into that environmental exploration and then physical exercise those are the top three like stress reducers
You can see she's got a much calmer attitude, and I didn't achieve it by suppression. I achieved it by giving liberty, giving freedom, allowing her to go and enjoy the scenery and get habituated and get used to the idea of being around here
Concepts: Start with freedom
Exceptions
- The emotional-systems account is one trainer's framing, given for one dog. Nobody defines 'decompression' or says how to tell it has happened.
Food is used sparingly, if at all. It works only while nothing in the room matters more, a dog paid in food every time comes to watch the hand and not the person, and a dog trained on treats may wait at home to see the food first.
Said as a rule on filmOn film
so we used to use treats a lot more we don't use them anymore because they're the most reliable if you have the treat or if the dog doesn't want something more than the treat right but they're like it becomes insanely unreliable if the dog even slightly cares about something more than the treat
gross and i don't like that look that i see on a dog's face when you turn up treats and they come up to you and you're like hey like kodak and they're like yeah and they're just like they're not even looking at me yeah they're just looking for the food
It starts becoming not a friendship and more symbiosis. It's like a bank.
Like, if there's one time I give her a treat, and let's say she goes back to her owners and then she just doesn't do it and she just looks at you, but she knows it, she's like, well, where's my food? I need to see my food first.
because you really can't train without some kind of reward right so we do use like negative reinforcement which is a lot of pressure but the dogs really aren't going to enjoy themselves if it's all pressure and nothing positive
Concepts: Food facilitates calm, it cannot create it
Exceptions
- Playing with food is different from giving food as the reward every time. Treat chasing 'can be excellent if done correctly', and lures 'definitely' have a place when necessary.
- Angus 'does really good with treats', so food became more of his training (watch 04:25). Cosmo got a lure 'for certain dogs' (watch 01:44). Christian uses food as a moving chase with Apollo the Bully (watch 01:20). One handler keeps food hidden until the dog answers the word and gives it about one time in five (Rocko, watch 08:50).
Once the dog knows the job, reward it some of the time, not every time: 'might get it' keeps him keener than 'definitely will'. The reward must never disappear altogether.
Said as a rule on filmOn film
So then he's really charged up to do it because he's like, oh, I might get it, and might get
um but either way you don't stay on a reward them every time for everything forever because
right so think about it this way it's like a battery all of your commands and all of your um
it's like maybe a 50 50 thing where he's like sometimes i get a treat sometimes i don't
Concepts: Might beats will
Exceptions
- The '50 50' is Connor's suggestion to one family, not a house ratio. Kelton's explanation, that dopamine peaks in anticipation, is his gloss and is not tested on film.
What owners are told
Find out what your dog wants by offering several things and watching. Don't write him off as a non-player after one dull throw: change the toy, change how you play, and try touch and sniffing too. Make play a real contest, let him win often at first, and ask for a command while the game is running, paying it by starting the game again. When you release him, say 'okay' as if it means something and let him go. On most of an ordinary walk let him be a dog, on a long line or retractable, and drop commands in when you need them. When you call him away from something, send him back to it. If you use food, keep it out of sight until after he answers, use it now and then, and never let the reward stop altogether.
On film
This is why you should always take some stuff with you and try it, right? Because it can totally change your approach. Try tug, try ball, try stuff.
teach them there. We generalize in all of those scenarios. And if you'll notice, like for the past probably five minutes, I haven't put Phoebe in any commands. And this is what we want our clients to be doing too.
to go sniff a random tree that's not on the sidewalk that's off the route well ask yourself what exactly is the harm what's preventing me from letting them do that
treats for instance say okay and then after you say okay then give the treat and i wouldn't i
What trainers know
Run a reward assessment before teaching: ball, tug, flirt pole, hands, touch (head against body), release to sniff, water, movement, food. Score each out of ten. Match the toy's firmness to the bite. Your own energy drives the dog's. Test the prey-cycle fragment by withholding the win. With a nervous dog, use prey moving away from you, not a toy pushed at him, and try personal play in case it frightens him. Play possession games by strategy, never rip the toy away, act beaten, and raise your level as he improves. Build the out on low-stakes objects first. Keep play at least as long as work. Leash pressure can be introduced inside tug so it predicts play (Kali, Kane). Treat play level as something to monitor: if the game drops, stop the obedience and rebuild the game with bonding time (Bullet, watch). Release to freedom pairs negative reinforcement at the dog's level of understanding with a big release and a long break to process (Kali, watch 11:08). Charge the release, and say 'okay' before stepping away, never with it. A long line or flexi, not six feet. On a break the dog was never told not to greet, so don't correct it. Open the session with nothing, thread the work in, release straight back out. If food is used, keep it hidden until the dog answers, deliver it after the release, at random. Lures fade fast (Cosmo).
On film
make it clear use just the right amount of negative reinforcement guide them depending on like their current understanding of what you want them to do and then a big release to freedom and a nice long break afterwards so that they can process and retain the information
introduce like a little bit of pressure while i'm playing with him so that he doesn't take the pressure the bad thing
it um i think i need to spend a little bit more time just chilling with him and bonding
i don't remember but he reacted to my words instead of the treat he's off so but i only do this like once every five times
In the dogs
Bullet
Five days of toys ignored, then a solid tug gripped hard on day six and lost again four days later. The trainer's reading is that freedom far outranks play for this dog, 'and I think it might always be'. Christian's diagnosis is that Bullet probably never played as a puppy. He was trained on release to freedom throughout.
On film
Okay, he doesn't care about any of the toys yet.
Yeah, he likes a much stronger toy because he's got a much stronger bite this dog
Kodak
On the first outing play is ruled out as likely to panic him. Days later a ball is offered anyway, and four minutes in the session plan is thrown away. From then on everything is taught through the ball. The family had written on the intake that he loves his balls.
On film
sometimes with dogs like this, like if I tried to play with him right now, oh, oh, that would be panic inducing for Kodak.
all we're gonna actually do today um based on what i'm seeing now we are literally literally just gonna we're just gonna play ball
Zeus
The fastest assessment on film: a heavy tug, 'this is what he loves'. The sharper reading comes a day later. It is possession he wants, so possession games carry everything, and giving the toy up becomes the lesson. He learns he can turn the game back on.
On film
This is what he loves. He loves to play tug.
No. Out. It's not yours yet. You will get it as soon as you're down, buddy.
King
The game taught as a subject. A senior trainer takes over a junior's flat tug game and turns an ignored toy into one King guards, by wanting it himself. On maintenance walks the first minutes are King's own, and the release to sniff is 'also very rewarding' next to the ball.
On film
He doesn't want me to have it now. Yeah. Right. Why is it important to him?
it's it's the strongest reinforcer for King King cares about that more than he cares about anything else
Deuce
A powerful dog everyone expects to live for the tug turns out to come back for pats. His file also holds the handlers' own account of how a game gets ruined, and the moment the house technique is named and argued about: 'release to freedom' with the phrase called 'dark'.
On film
he likes actually like he'll play tug um his preferred method of reward though is just like love which is funny because you would expect the ratty to really wanna
Luna
Sniffing is her reward and treats are deliberately not relied on. Connor calls letting a nervous dog step around and explore 'a big secret', and names a quiet park as the venue. With the same dog he also locks the leash so she cannot explore when she fights a position.
On film
Sniffing around is her reward and she loves it. The thing is, I don't really rely on treats
fantastic opportunity if you are trying to work on nerves with confidence. A nice quiet park. It's probably your best bet.
Lola
Kelton gives her family the case against food ('like a bank') and the handler's choice between being a restrainer and being a source of freedom. Her recall is built so that coming back reopens her world. On a long line, eighty per cent of a session is 'nothing'.
On film
Turns out that what it does, it makes it transactional, it's like a bank, like I'm like a bank and
a session like this i might be out here for 10 15 20 30 minutes or whatever and like 80 of it is what i'm doing now which is starts with an n nothing
Dutchess
Touch as a first-class reward. Tug was tried and honestly reported as a weak motivator. A rub on the face and time off the lead are what she works for, and 'it's pretty cheap'. Her go-home carries the fullest statement of why the business stopped relying on treats.
On film
every dog learns differently and i don't believe there's like a one one to all method for every dog so for her if her reward is just like going out and sniff and then like rub on you and that's what motivates you then that's all you need to do it's pretty cheap
Bo
The honest negative. Kelton works at building Bo's ball drive for months and it never lands, so the release carries Bo's training instead: 'I worship at the altar of the release command'.
On film
much I'm trying to get him to care about the ball more let's go good job dude
worship at the altar of the release command you have a solid release command and things will never ever never ever be clearer to your dog than that
Kalypso
She refuses every treat for two private lessons. In the third, the trainer stops offering food, lets her find a squirrel, calls her off it and sends her straight back to it.
On film
Is there anything wrong with you sending her back to where the squirrel was? No, right?
Dolce
Play suppressed as a puppy and reopened on purpose. The chase part of the game is built with a flirt pole off the possession game she already wants. In protection work the win is priced by the criterion: the whole sleeve, held.
On film
we're just trying to open up more for the chasing part so that we can get her to really enjoy that part of the game and then we can transition to like sleep, work, and things like that.
Roxy
A treat trial run on camera, stopped with 'that's enough treats' because the trainer is thinking of the day she goes home. The reward is re-tested by context: the pool on a hot day, freedom on walks. Her file also holds the case against short leashes.
On film
Eh. Roxy. I didn't need to use the e-cog. Down. Good girl. Oh, good girl. I think that's enough treats. Treats can be a good thing, but treats can also be a bad thing.
Angus
Angus learns badly under pressure and well with treats, so more of his training moves to food. The trainer names the cost: a dog who watches and nudges the hand. His story ends where the footage ends.
On film
Angus does really good with treats. So I'm using treats for more and more of Angus's training
Where the trainers differ
On film
Block them away, that way there's like no freedom there. Doing a release of freedom, that's one of my favorite methods for teaching.
And so it's kind of like a dual reinforcement.
On film
There's actually plenty of positive reinforcement going on here, because what is it that Louie wants to do instead of whatever it is that I ask him to do?
On film
better way to say that because i don't like saying it that way nope um i don't like being like we're gonna take away his freedom like it just sounds so dark
technique you can do with the e-collar. It's called release to freedom. And I have to find a better way to say that cuz I don't like saying into the camera
On film
negative reinforcement release to freedom is king baby it's king and play is god
On film
You can see she's got a much calmer attitude, and I didn't achieve it by suppression. I achieved it by giving liberty, giving freedom,
On film
that I'm taking away people's freedom. But what's happening is that I do the e-collar. If she's still trying to fight me, I will lock the leash in place to where she can't go out and explore.
On film
what I don't want to do is start right now at this moment I don't want to start immediately integrating like too much obedience into the play because of his like super crazy nerve starting point
I'm not giving her any formal commands yet. Here. I just want to see if I can get her interested in the game a little bit stronger.
On film
That's the easiest way to train a dog is if they like to play catch or they like to play tug or something like that
On film
Um because the play session is the way that often if a dog is uh capable of like opening up and playing with you, um if not, we just have to go straight to the walk at first.
Well, she likes freedom, so we're gonna use that for right now
On film
My last couple of sessions with her have been play out in, like, a field. And what I discovered is that, I mean, like, her primary form of reinforcement is just, like, a release of freedom.
On film
that she loves the pool so that's a reward
As a course unit
The learner can run a reward assessment, play a possession game that pays a command, deliver a charged release into freedom, structure a walk so that most of it is the dog's and commands buy freedom back, and move a known behaviour from every-time to sometimes reward without letting it reach zero.
Lessons, in order
- The dog tells you the reward
- Play is a capacity you can grow
- Play becomes the reward - and play is a contest
- Release to freedom
- Meaningful release
- Start with freedom
- Food facilitates calm, it cannot create it
- Might beats will
- Proof it against his own currency
Drills
- The reward assessment, as run for Kane and Beo: offer ball, tug, flirt pole and your hands for a couple of minutes each, animated, and score each out of ten. If nothing lands, change the object and your own energy before concluding. Also try release to sniff and touch.
- The miss test, as run with Rocko: with the best game, make him miss a few times and watch whether his arousal climbs. That tells you whether chase or possession is the part he wants.
- Play-command-play, as run for Madam and King: a live tug game, then the out (drop the toy yourself if needed), one command, then 'okay' and the game restarts at once. Vary how many rounds pass before the next command. Sometimes give the toy straight back with no command.
- The release walk, as run with Bullet, Lola and Kalypso: on a long line, let him choose for a few minutes and note what he chooses. That is his wage. Ask for one simple behaviour, and the instant he complies say 'okay' with energy and send him to the thing he chose. For recalls, call him away and send him back.
- The first-minutes rule, as run with King and Beo: arrive and do nothing. Let him out, let him pee and sniff, with no cue at all.
- Variable schedule, as given to the March Zeus's family: once he knows it, give the reward after 'okay' about half the time, at random, and never stop entirely.
Assessment
- Your dog walks past the ball and ignores the soft tug. What two things do you change before deciding he doesn't play?
- When your dog does what you asked, what turns back on, and does he know he turned it on?
- Your dog ignores treats outdoors and keeps pulling to a hedge. You call him and he comes. What do you do next?
- Outside, with a squirrel in view, what reward do you have that beats the squirrel?
- Why does a dog paid every single time become less reliable, not more?