Saturday, December 24, 2011

Luring vs Shaping in Flyball

The trend in dog training in general is shaping instead of luring or removing the lure very quickly. This makes sense to me because if the dog has to think about making the right choice vs following the reward they are more likely to repeat it. Also they are more likely to do what you want when you don't have a treat. Luring really only seems to work if you fade it immediately

In flyball we use shaping a little bit for wall work, but for jumping and box work we basically use luring. For some dogs this works really well. Indigo for example has always had a nice box turn and never went around the jumps. Other than the gutter I use to keep her turn tighter, her runs look exactly the same whether or not she has a jump in front of the box.

Some dogs have lovely turns ONLY with a jump board in. As soon as they get to a tournament, it falls apart. This is basically like the dog that will only recall if you show them a treat.

What people generally do to fix this is wall work, lots of box work with the board in place or reteach the turn from the beginning. Sometimes a clear plexiglass jump so the dog doesn't see it or can't anticipate doing a nice turn or a crappy one.  Rarely do people pull a dog from competition for this, so the dog does the bad turn in tournaments and gets rewarded. So without any different feedback for different turns, the dog has no reason to think it matters.

I've been working on a method for training this type of dog. Here is what I do:

1. Wall work to build muscle memory for a nice turn
2. Figure out the box aid that gives the dog the best turn
3. Always have the handler at the box to tug or not tug depending on the turn
4. Have a 2nd person send the dog on a full run with the correct box aid in
5. If the dog does a nice turn (they probably will) give a big reward
6. Next put in a smaller box aid (a gutter or PVC) and do the same thing if the turn is nice, give a big reward, if the turn is crappy give no reward and try again with a box aid back in.
7. If the turn IS nice, try again with no box aid.
8. Keep alternating between box aid and no box aid

 Continue this process so that the dog continues to get the correct feedback depending on the quality of the turn, making sure that you make it easier if the turn is bad. You can click or say "YES" for nice turns.

I'm curious if any other flyball trainers have tried anything like this?


3 comments:

  1. Hmmmm....shaping, luring. I think the biggest issue, is at least how I see it is dogs who need the muscle memory. Trainers tend to lump intricate behaviors together, some dogs get it while others need it broken down. Some dogs you can skip steps some need to stay there longer. A prop suitable for each dog is key,and then as much as I hate repetitive work, the majority of dogs need tons to build muscle memory. Then they need to work distance, my biggest peeve is the trainer who gets a nice turn close and exoects the same 60 feet from the start line. Goodness, how could any turn hold up? Start slow, work close, build good behavior, reward it and be willing to pull a dog who won't hold the criteria. Dob't over run the and expect an awesome turn when they are exhausted. I could jeep rambling, just my minority view!

    ReplyDelete
  2. I know you've seen my quest to find the right prop for Koira's turn to improve. I'm hoping we finally have it. The plan will now be to run her with that prop up close for a while, then start backing up to eventually doing full runs with the prop. If her turn stays nice, we will start replacing the solid jump with a PVC jump in the same place. Then fade slowly to something less, like a gutter, then eventually no aid, is the goal.

    As for shaping versus luring, I am currently holding to the same opinion I know many obedience trainers have- shaping is a great tool, but if you want something precise, it is not always the best method to get there. Since I want a good turn from the start, I will likely always lure it at least somewhat, either with a touch stick or with a hand. But who knows, maybe training changes will come.

    ReplyDelete
  3. Yeah I didn't mention muscle memory. In a lot of dogs tons of muscle memory works really well. But there are dogs that don't get it based on memory and I feel like the flyball training experts don't have a good method for dealing with turns that fall apart in a tournament setting without props.

    ReplyDelete