Reverse Progression/Backward Chaining


#1

Here’s an article about the value of using a “reverse progression” approach to training in sports.

[quote]Rushall, B. S. (1996). Some practical applications of psychology in physical activity settings. In K-W Kim (Ed.), The pursuit of sport excellence Vol. 2 (pp. 638-656). Seoul, Korea: Korean Alliance for Health, Physical Education, Recreation and Dance.

A common characteristic of most physical activity instructors is the skill element progression taught to beginners. Textbooks frequently provide photo sequences starting at the initiation of the skill and ending at the “finish” position. This seems “logical” and is readily justified on the grounds that if a skill is not initiated properly, it will not be completed correctly. However, if the literature on instruction of non-verbal species is examined, it will be concluded that teaching progressions do not commence at the “start” and finish at the “end.” Rather, the first element instructed is the last aspect of the behavior, the “terminal” element. Progressions of instruction are called “backward chaining” when elementary movements are required in a sequence and “backward shaping” when a single two-phase motor behavior is taught.

The assumption that a logical start-to-finish instructional progression is the best can be debated. In forward progressions, students normally learn the early elements of a sequence reasonably well. However, as a chain grows longer, activities have to be performed before the new element can be acted. The imposition of behaving before attempting a new element eventually interferes with learning. When chains are long, the success of implementing new elements becomes quite difficult and errors increase, a result that needs to be avoided if efficient and effective learning is to occur. The interference caused by the activity prior to the attempted control of a new skill element is a major weakness with forward progression instruction.

Forward progressions also invoke other behaviors which, as skill complexity increases, become detrimental to both the learning process and eventual performance. Learners often construct “mental check-lists” of instructed skill elements to ensure proper technique, resulting in cognitive control becoming an established part of covert behavior in the skill execution. Even in a simple skill such as a golf swing, such thought intrusions interfere with learning and skill development. In the length of time before a new element is acted, performance of that which has already been instructed intervenes, which promotes forgetting of what needs to be done. The mounting difficulty of introducing new elements into behavior chains increases anxiety, the frequency of negative self-appraisals, and the execution of errors. These phenomena further hinder learning. Forward progressions usually produce skills which are executed well in the initial stages but deteriorate and exhibit weaknesses and faults as the sequence progresses.

In spite of shortcomings, forward progressions as a teaching structure have been perpetuated and rarely questioned as to whether or not there is a better alternative.

The more traditional and effective way of instructing non-verbal species should be considered for humans. Discussions about backward or reverse progressions have appeared periodically in exercise-oriented literature (Chelladurai, & Stothart, 1978; Dusault, 1986; Rushall, & Ford, 1982; Sherman, & Rushall, 1993; Spooner & Spooner, 1984; Spooner, Spooner, & Ulicny, 1986). In reverse progressions, chains or skills are learned by teaching the last element of the skill first, the next-to-last second, etc. The completion of the skill or chain is the most practiced element. As the activity is performed, skill does not diminish as it progresses. This results in the execution of very safe landings, good follow-throughs, total skills, etc. None of the problems typically associated with forward progressions emerge in backward sequencing. The following figure contains a schematic of the two progressions and compares their elements.

An example of backward shaping with a golf swing will demonstrate the differences in these two concepts of instructional progressions. The grip is the first element that is taught irrespective of the progression used. The student should be able to form a satisfactory grip and wield the club in a variety of actions. From then on, the forward and backward progressions differ. After the grip, backward shaping dictates that the final follow-through position be taught. The learner is instructed that it is the terminal position that should be attained at the end of every trial. Knowing the criteria for judging that position allows the learner to execute covert positive reinforcement at the end of each trial if the criteria are achieved. Thus, every trial ends with the learner appraising whether the skill progression has or has not been achieved correctly. Successive steps move the club progressively further back in the “ideal” swing. Each step has the new element executed first, followed by the remainder of the skill which has been successfully performed on previous trials. The size of the step progressions should be sufficiently easy to minimize performance errors. When done correctly, this contrasts with the forward progression, which produces errors primarily due to the interference phenomenon and the progressive weakening of the skill strength.

The major advantages of backward shaping/chaining over forward progressions can be summarized as follows:

interference does not occur since each new element precedes all previously “learned” elements,

that is, the learner thinks of and executes a new technique element and follows it with what has been done successfully before,

each step progression does not increase in difficulty since undivided attention can be focused on new content,

attention is focused only on the new step and then established elements are performed to finish in the terminal position,

there is a lack of tension/anxiety in the learner because of the simplicity of the task and its steps, and

emotional problems are highly unlikely because step sizes are small, guarantee a high rate of success, and are typically understood clearly by the learner.

There are some skills that do not readily lend themselves to reverse progressions, for example, diving and jumping. In activities such as these, the total behavior should be executed with a moderate degree of proficiency. From then on, skill refinements should be emphasized in a reverse progression.

When refining established skills, corrections are more effective when they are introduced in reverse order (Sherman & Rushall, 1993). The backward shaping of rowing ergometer technique was shown to produce fewer errors in the learning experience than a forward progression (Rushall, 1984). All the elements of a golf game were performed more efficiently when they were taught in a reverse progression (Simek & O’Brien, 1981).

Comparisons of forward and backward chaining progressions in manual task instruction have shown the reverse procedure to be superior in developing speed, accuracy, fluency, and skill maintenance (Martin, Koop, Tumer, & Hanel, 1981). It was also shown to be superior when teaching response chains (Weiss, 1978) and instructing military tasks (Cox & Boren, 1965).

Reverse progressions are a viable alternative to traditional forward progressions in skill instruction. Teaching tackling would commence with the opponent and tackler lying on the ground in a firm hold position which would constitute “how the tackle should finish.” An equivalent backward progression starting position would occur for various wrestling take-downs and counter moves. A small child learning to throw would practice follow-throughs and release actions prior to first holding a ball. The sequence of steps has to be devised for each sporting action and then the size of the step progressions individualized for the learner’s capabilities.

Below are some personal claims about the efficacy of backward progressions. These statements can be readily assessed through appropriate research projects.

The rate of learning is much faster than for forward progressions. With mentally retarded subjects, the learning rate for a backward progression is approximately equal to the rate for normal subjects using forward progressions.

The fewer errors increases a learner’s self-concept and self-efficacy for physical activity pursuits.

The number of trials to criterion is significantly fewer when compared to forward progressions.

The resulting strength of the latter elements of skill performance leads to a higher level of skill performance than that which is generally achieved through forward progressions.

Backward progressions for shaping or chaining should be tried as an alternative instructional procedure in physical activity pursuits. The experience will be rewarding for both the learner and instructor.[/quote]


#2

D.M. A great piece of information, thanks!


#3

A challenge in applying reverse progression training to pitching is the requirement for momentum to be present to allow certain details of the pitching motion to happen. For example, to get the centre of gravity and torso as far forward as is required or recommended, you must have built up some momentum. Of course, this presupposes that this is a good thing, as I do, which some may debate. Another example is the requirement for the combination of total body movement and shoulder rotation to facilitate the upper arm (humerus) to be able to externally rotate to the amount desired (typically laid back to horizontal).

So, if you start with the learner in whatever you consider a “good” finish position, and have them “learn” that, you then would need to “step back” in the process to add the next element to be “learned”.

What do you do next? I’m actually asking you all this question for your ideas.

We must keep in mind that we want each step to be as “specific” as possible. To be most effective, it must resemble what happens in the game situation as closely as possible. The farther you deviate from that, the less applicable the “learned” skill will be.

Assuming that anyone is interested, I’d like to see the people on this board brainstorm this one. I’ve got some thoughts on it but I know they’re not perfect. Give me your ideas as to what you think the steps might be that would satisfy the intent of the reverse progression thoughts as well as keeping things as close as possible to the game situation, such as momentum and a mound.


#4

In general, any point of energy transfer between links in the kinetic chain could pose a problem for reverse progression. However, it is possible to do drills that put the body in the various positions it goes through during the pitching delivery and then to continue from that point though not necessarily with the same momentum/force/velocity/whatever. The question becomes “is that a sufficient enough emulation of the real delivery to make the drills beneficial?”. I feel it is.

An example of such a drill is the NPA’s “stack & track” version of their towel drill. This drill puts your body in a position where your weight is out over the front foot, back leg is extended back, front leg is bent and firmed up, head and shoulders are upright and online with the target, and glove is up and in front. From this position, the throwing arm goes through its motion and the target is (hopefully) hit with the towel. This drill lets you practice the release and follow-through while in the position they want you to be (according to their model of mechanics).

After this drill, one could then switch to the “rocker” version of the towel drill which starts with a position earlier in the delivery - weight is centered, hips and shoulders are still closed, etc.

Finally, there is the full towel drill that starts at the usual start of the delivery.

Now, although I don’t recall the NPA recommending doing these drills in this order to follow the idea of reverse progression, they obviously can be. Also, this sequence of drills probably doesn’t provide the ideal level of granularity - the steps between them are too big. But I think variations of these drills could be created to provide smaller steps.

And also to achieve maximum hip and shoulder separation.

[quote=“dm59”]So, if you start with the learner in whatever you consider a “good” finish position, and have them “learn” that, you then would need to “step back” in the process to add the next element to be “learned”.

What do you do next? I’m actually asking you all this question for your ideas.[/quote]
You do as I described above - you practice each step in a more mechanical manner than occurs during a normal pitching delivery. You just have to accept that there is value in this. Tom House claims there is value in his drills simply from getting pitchers use to what it feels like for their bodies to be in each of these positions.

I think there has to be compromises in the case of pitching.

One concern I have with this reverse progression stuff is the assumption that things already learned are no longer a distraction. I’m not sure I agree with that. But I speak naively about this so I’m curious to hear what others have to say.


#5

I can’t say if it really applies to pitching or not but we used the principles in teaching tennis strokes 25 years ago.


#6

[quote=“CADad”]I can’t say if it really applies to pitching or not but we used the principles in teaching tennis strokes 25 years ago.[/quote]Yeah. He tends to say that a lot, not necessarily tennis, but other sports. He claims that other sports have been there for a while and that baseball hasn’t really caught up.


#7

[quote]One concern I have with this reverse progression stuff is the assumption that things already learned are no longer a distraction. I’m not sure I agree with that. But I speak naively about this so I’m curious to hear what others have to say.
[/quote]

Digging up some older posts. Do you think reverse progression is only useful to someone just learning a skill or can it be useful for someone trying to improve a skill. I know the article says anyone can, and I agree that it’s tough to believe that things already learned are no longer a distraction.

As far as a progression, how about starting in with the legs in the stride position (or whatever we want to call it), and doing the rock back and throw as Roger suggests. Then the next step could be starting with feet shoulder width apart and simply striding into and through the rest of the delivery. The next step could simply be from the stretch, leg lift and go. Then the full wind up could finish it up.

Does that sound like too simple of a progression?


#8

Based on what the article said, then yes that is too simple of a progression. The article indicated that the steps of the progression need to be small such that the difficulty of advancing from one step to another is insignificant and can be overcome by focusing entirely on that next step. I don’t feel this is possible for pitching for the reasons that DM stated.


#9

[quote=“palo20”]Do you think reverse progression is only useful to someone just learning a skill or can it be useful for someone trying to improve a skill.[/quote]It’s no doubt great for beginners but I do believe it can be tremendously effective for improving things. It’s the idea of putting it in context that I find intriguing.

[quote=“palo20”]As far as a progression, how about starting in with the legs in the stride position (or whatever we want to call it), and doing the rock back and throw as Roger suggests. Then the next step could be starting with feet shoulder width apart and simply striding into and through the rest of the delivery. The next step could simply be from the stretch, leg lift and go. Then the full wind up could finish it up.

Does that sound like too simple of a progression?[/quote]I kind of like this thought. The rocker will allow a bit of momentum to be generated, which is what I’ve always had difficulty with in my attempts to stay specific. I’ve tried getting into that “stride out” position but I got the pitcher to get the front foot up onto the toes with the heel up and turned over so that the hips are closed. Then they would rotate to land and throw with everything the already “learned”. This isn’t so bad but the rocker just might be a better approach.

Hey, Roger. Did you suggest that to me before and I just didn’t get it through my thick skull?


#10

Yeah, sort of. (I suggested the progression - not that your skull is thick :lol: ) I described doing the “stack and track” version of the towel drill followed by the “rocker” version of the towel drill followed by the total towel drill. Of course the towel could be replaced with a baseball if you don’t like the towel. But what I described was just 3 steps - a rather coarse progression.


#11

Well, I like the rocker step idea but would prefer to actually throw a regulation ball due to the specificity issue.

Sounds good guys.


#12

The challenge of reverse progression is in the applicability to specific sport. Obviously skydiving and the like are out. Applying this to pitching might require the use of an additional piece of hardware to support the pitcher at each desired phase of the delivery. Kind of like the isometric drill with a towel held during each phase of the delivery only backwards…

Its interesting to me because the reverse progression is used in the interview process with crime victims as a technique to fill in previously forgotten or latent information not developed by the investigator in the classic forward progression interview.


#13

In the progression, is there any “learned” movement backwards from one progression to the next? Trying to find the best way to word this. Say we start with the follow through. Assuming that we move back to the position right at foot plant, before the throwing begins at foot plant, is there a segway where the pitcher moves from the follow through back to the stride position backwards? IOW, does he move backward at all (much like Wolforth has pitchers pull bands backwards to build up back muscles)?

So far the progression we have is:
Follow Through
Stride Position, rock and throw
Feet shoulder width, stride and throw
Stretch position
Full Wind up

Any other thoughts or bridges in between these steps?