Attn Michael and other successful owners: My recalls suck.

[GlassOnion]

I don't feed my birds any millet and save it for training. The problem is that when I call Apple, he will come only if he feels like it. I know he loves millet as his favourite treat, yet he won't fly to me for it unless he is in the mood. How can I make him more motivated to recall? This is proving much more difficult than call, *click*, treat, repeat.

[Michael]

1) Food Management
2) Variable Ratio Reinforcement

First you have to get pretty good recall in the first place using Continuous Reinforcement. Every recall, click and reward. Try to avoid recalling the parrot when you think it won't come. If you've already done a lot of recalls, parrot looks distracted, you know parrot doesn't want the reward, etc, don't even bother calling because it solidifies that it won't have to always come. Don't do too much other training in the same session either because that dilutes the motivation.

Try to vary the treats and use alternative rewards from time to time such as toys, petting, and attention. These are unlikely to be very effective with a cockatiel so it's up to you to figure out how well they can work. Train in a low distraction environment. Practice the bird flying from a low distraction place such as a Parrot Training Perch. Also never show the treat to encourage recall (that's luring and the bird won't come when it doesn't see it). Hide it in your hand and even if you don't have it, pretend you have it the same way. Use a target stick if it doesn't know to come.

If you can get 5 solid flight recalls using this method (and even not any more) most training sessions, then you can apply variable ratio reinforcement. You want to have a solid history of recall preceding this so you don't end up erasing it but if you are sure the bird knows what to do, this won't undo your efforts (or only temporarily). Begin by not giving a treat randomly once out of 3-5 recalls. Just click, give the bird attention, let it catch its breath but don't give it the treat. Progressively decrease the number of times you reward out of the number of recalls but be sure to keep this random and unpredictable. You will find that the bird will recall with the same or even better reliability (if you do this right and don't advance it too quickly) rather than worse. The parrot will learn to come whenever you call for the mere possibility of getting a treat (or something good) rather than only when it wants the treat. It's more likely to want the treat if it gets rewarded less frequently as well.

For an experiment, I got Kili to fly 60 recalls (over a mile of distance) for a mere 4 seeds. Incredibly her reliability was the same as 1 treat per recall (which would have otherwise been just 4 recalls)! The point is that if the parrot can do so many recalls for training and practice (not to mention great motivation for exercise), then the likelihood it will recall to you when you just want it to come for you, the chances are much higher. If it is used to recalling on a ratio of 1 treat for 10 recalls, then the 9 times you recall it for the hell of it vs the 1 time you sneak it a treat you hid away for that random time, it will give you the chance to have reliable recall without it looking like it's all for food. Of course if you can convince your parrot to come to you merely for scratches or attention, that's even better but I think that's much harder with budgies/cockatiels. And with any parrot, there's only so many times it will come for those things. But with the presence of some element of hunger, it will recall reliably for the possibility of getting a treat as long as the hunger is still there and recalling is the only way of attaining it.

[GlassOnion]

Okay, thanks for the read. My question is, if my bird isn't even motivated by a 100% guarantee in getting a treat, why would he be motivated by a possibility of being rewarded?

[Michael]

Because it becomes sort of like a game. A lottery of sorts. I think it's the feeling of the more you fail, the harder you must try kicking in. But there MUST be an occasional pay out or they'll give up entirely. However, variable ratio reinforcement is more resistant to extinction (if you always reward and then for a few times you stop rewarding, the bird may give up trying). The other benefit is that you can make fewer treats go a longer way before the parrot is no longer motivated by treats. So if you can reliably get the parrot to do 5 recalls for 5 treats, maybe not more, but 5 is always possible for a training session, then training over time to stretch it to 60 recalls for 5 treats is possible. Each time the bird recalls it still receives a secondary reinforcer click (and attention, petting) and the possibility of getting a treat. The bird knows that if it sits around and doesn't recall it won't get the treat and there is no other way to get the treat, then it has to try every time in case that's the time it will get it.

This doesn't solve a situation where the bird won't come 5 times for 5 treats, it just makes you able to get more recalls per day because the bird takes longer to earn its treats and fill up. This is why I said evaluate food management techniques first (I'm presuming the behavior has been trained already). So assuming the training has been done correctly and you're certain the bird knows what to do, it will be about experimenting with food management to improve motivation for the treats to get the recalls. The simple fact is that flight takes way more energy than the typical tricks so likewise it requires greater motivation to get it to come. See about reducing meals a little bit or stretching the times between meals, etc. Be smart about it. See this other post I made about experimenting with various levels of food management. The good thing is that once you've done a ton of training in this manner and the bird has a strong habit of coming to you off the bat, then less food management is required and higher ratios can be achieved. However, some limited level of hunger (such as time since last meal) and continuing some level of variable ratio reinforcement is required.

You can use other things to reward as well. If I see my parrots looking anxiously at me posed to fly, I can tell they want to be with me, so I recall them and don't give a treat. They just want to hang out with me. This is great and it works some of the time. It's just if you want reliable recall, stick with food cause it works most of the time.

[GlassOnion]

Alright....thanks.. we'll see how it goes!

[Erithacus]

I rarely reward my tiels with food. I usually give them a pet or praise them. If they come they will come. Sometimes they just ignored me if they found something more interesting to play with or if they were very comfortable with the place they were sitting on. On good days, they would respond 80% to 100%.

[zazanomore]

Einstein seems to be more motivated by praise then by millet.

We've reached the point that where I can tell him to step up, and he'll say "good boy!".

Could you use his favorite toy?

[laducockatiel]

I have a method that I tried and it has proved easy:

What you need to do is get the selected treat, put it behind your hand and start with your hand close to the bird. Your bird will jump on and eat the treat. Every time you do this, move your hand away gradually. Eventually, when he flies to you from a long way away, then you can take the treat away amd cos he is used to flying to you.

Hope this helps