TL;DR
Understanding a task is the beginning. Reliable performance means being able to complete it accurately, notice emerging errors, and adapt when the conditions change.
AFLS starts with a clear standard and a demonstration of the complete task. The learner confirms understanding, prepares, and performs a baseline run. After each attempt, the learner and coach review what actually happened and choose a specific change for the next run.
The task is then repeated under progressively tighter, manageable constraints. These might involve time, complexity, available support, or distraction. A return to the original conditions creates a deliberate contrast: the learner may perceive more room to notice detail and correct mistakes.
If a weak sequence remains, it is isolated, rebuilt, and returned to the complete task. Every claimed improvement is checked for retention after a delay and transfer to a changed context.
The purpose is safe, accurate, repeatable performance with increasing self-correction. Pressure is useful only while it reveals errors the learner can understand and act on.
Where this came from
After experiencing the STA Spirit during intensive training in Switzerland, I wanted to write down the learning logic I had witnessed and experienced. The sequence of demonstration, practice, pressure, decompression, and repair made a strong impression on me.
I called this written framework the Action Feedback Loop System, or AFLS. The training experience and the instructors at STA inspired it; the framework presented here is my interpretation of what I observed and a way to explore its wider application.
Eight stages, one connected loop
The eight stages provide a quick overview. The twelve operating phases below expand the same sequence.
01 The whole learning loop
| Stage | Operating phases | What it does |
|---|---|---|
| 1. Frame | 1 | Define the task, purpose, standard, and boundaries. |
| 2. Model | 2 | Demonstrate the complete task slowly and clearly. |
| 3. Prepare | 3–5 | Confirm understanding, prepare, focus, and commit. |
| 4. Perform | 6 | Complete the whole task at baseline. |
| 5. Score and feedback | 7–8 | Secure, self-assess, review evidence, and choose the next action. |
| 6. Compress | 9 | Repeat under calibrated constraints, with feedback after every run. |
| 7. Decompress | 10 | Return to baseline and use the contrast to refine performance. |
| 8. Repair and validate | 11–12 | Repair when needed, reintegrate, and test retention and transfer. |
Feedback continues throughout the loop. Repair depends on what the evidence reveals. Validation is required whenever improvement is claimed.
The twelve operating phases
1. Frame the whole task
Define the purpose, desired result, and observable performance standard. Explain the complete sequence and its boundaries before concentrating on small details. The learner should know what successful performance looks like and why each part matters.
2. Model at Level 0
Demonstrate the whole task at a slow, controllable pace. Make important decisions, transitions, checks, and likely error points visible. Answer questions and show how the safety conditions are maintained.
3. Confirm understanding and safety
Ask the learner to explain the task, boundaries, and stop conditions in their own words. A readiness signal should reflect demonstrated understanding. Resolve gaps before starting.
4. Prepare
Prepare the tools, information, workspace, protective measures, and contingencies. Mentally rehearse the sequence. The learner gives an honest readiness signal and identifies anything that still needs attention.
5. Focus and start
Use a short focus cue to direct attention to the objective. Remove irrelevant inputs and begin with a clear start signal. Consistent cues make runs easier to compare.
6. Complete the baseline run
Perform the complete task under the baseline conditions. Continue through manageable friction using planned contingencies. The coach observes without interrupting unless a predefined safety or validity threshold is crossed.
7. Secure and reset
Check people, tools, output, and workspace. Restore a safe environment and allow enough recovery to preserve the quality of the next attempt. Resetting is part of the performance sequence.
8. Score and give feedback
The learner self-assesses first. The coach then adds observable evidence. Select one to three priorities and turn them into concrete changes for the next run. “Check the output before the handoff” is actionable; “be more careful” leaves too much open.
9. Compress through Levels 1 and 2
Repeat the whole task with progressively tighter time, complexity, support, or distraction constraints. After every run, secure, score, give feedback, and recover. Increase the constraint only while performance remains safe and the errors remain understandable.
10. Debrief and decompress
Briefly review the stretch run, then return to Level 0 conditions. Perform again with attention to flow, precision, awareness, and previously missed details. Compare what changed when the constraint was removed.
11. Diagnose and repair
If a repeated, consequential gap remains, identify the exact sequence and its start and end points. Rebuild its elements slowly, reconnect them into meaningful chunks, and practise the repaired transition until it becomes smooth.
12. Reintegrate and validate
Return the repaired sequence to the whole task and test the complete performance again. Check retention after a delay and transfer under changed conditions. Continue the loop until the standard is met safely, accurately, and repeatably.
Calibrated pressure and the return to baseline
Pressure should reveal where performance becomes inefficient or unstable. Its intensity must fit the learner, the task, and the cost of error.
02 Challenge, then create contrast
| Condition | Purpose | Typical change |
|---|---|---|
| L0 — Baseline | Establish complete, controlled performance. | Defined time, appropriate support, clear conditions. |
| L1 — Compress | Expose inefficiency. | Moderately reduce time or support. |
| L2 — Stretch | Challenge adaptation within safe limits. | Add a demanding but manageable constraint. |
| D0 — Decompress | Return to L0 and compare performance. | Restore baseline time, support, and conditions. |
| V — Validate | Check whether learning lasts and transfers. | Test after a delay and in a changed context. |
The “more time” effect
One striking part of the experience was the return to baseline after a demanding run. The task could feel more spacious: there seemed to be more time to notice, breathe, and correct detail.
AFLS deliberately includes this contrast. The coach asks what became visible, what felt different, and whether the scorecard shows a useful change. This is a practical observation to test; it is not a guaranteed effect for every learner or task.
If safety, comprehension, technique, or the ability to use feedback deteriorates, reduce the constraint. A harder attempt is worthwhile only while it remains informative.
Repair a weak sequence, then restore the whole
Repeating the complete task can reveal a bottleneck without resolving it. AFLS temporarily isolates that bottleneck so it can be rebuilt with greater precision.
03 Rebuild the part that keeps failing
- Name the failing sequence and define its start and end points.
- Divide it into meaningful elements: A, B, and C.
- Practise A slowly for approximately eight quality repetitions, with breathing and a brief reset. Repeat for B and C.
- Connect A and B, then add C: “AB … C.” Repeat approximately eight times.
- Perform A–B–C as one smooth sequence approximately five times.
- Return it to the complete task.
- Check whether the original error disappeared, persisted, or moved elsewhere. Then test retention and transfer.
The repetition counts are starting parameters. Adjust them to the quality of execution, fatigue, task complexity, and cost of error. A component that works in isolation still needs to work inside the complete performance.
What the coach, learner, and system each own
All three share one reference: observable performance against an agreed standard.
04 Three roles, one shared standard
| Role | Responsibility |
|---|---|
| Coach | Define and demonstrate the standard, collect evidence, calibrate challenge, and protect safety and feedback quality. |
| Learner | Prepare honestly, perform, monitor their state, use contingencies, self-assess, and apply the next change. |
| Training system | Keep cues, constraints, scoring, records, and progression rules consistent enough to compare attempts. |
This structure makes the next attempt a joint decision grounded in evidence. The learner increasingly participates in diagnosing and correcting their own performance.
Measure performance without hiding the trade-offs
Keep speed separate from quality. A faster run can conceal missed steps, weaker decisions, or an unsafe shortcut.
| Dimension | What to observe |
|---|---|
| Safety and integrity | Whether all non-negotiable boundaries were respected. |
| Completeness | Whether required steps were performed. |
| Sequence | Whether the order and transitions were correct. |
| Accuracy and quality | Whether the output met the agreed standard. |
| Time | Whether the run met its defined time constraint. |
| Adaptability | Whether the objective was maintained through manageable disruption. |
| Self-correction | Which errors the learner detected and fixed independently. |
| Retention and transfer | Whether performance survived a delay, reduced prompting, and a changed context. |
Useful debrief questions are: What was the intended standard? What actually happened? What changed under pressure? Which error had the highest cost? What will change next time? What evidence will show that the change worked?
Keep the boundaries explicit
The continuity rule is simple: continue through manageable friction using planned contingencies; stop when a predefined safety or validity threshold is crossed.
Stop conditions include loss of safe control, risk of harm, acute distress, inability to understand instructions, meaningful physical deterioration, and failures that make the exercise invalid. Ethical, legal, privacy, and other integrity boundaries also remain binding.
For demanding sessions, check the learner’s actual state. Ask what they notice physically and mentally, which error they are most likely to make, and whether they need recovery, less complexity, or a different drill. Qualified instructors should define and supervise any high-consequence application.
Fatigue, confusion, and fear can make repetition less useful. The next run should preserve the learner’s ability to understand the task and use the feedback.
How the same logic could be applied elsewhere
These examples illustrate how to adapt AFLS. They are proposed applications, not evidence that the complete protocol produces equal results in every domain.
An executive briefing
Start with a three-minute briefing using notes. Repeat in two minutes without notes, then in ninety seconds with one controlled skeptical interruption. Return to three minutes and review pauses, emphasis, and clarity. Repair the weakest transition, then test the briefing after a delay and with a different audience.
A technical operations workflow
Run a complete process with a checklist in a training environment. Reduce support or introduce a controlled missing input. Return to the original conditions and review verification, sequencing, and error detection. Repair the failing handoff and validate with a different dataset or environment.
A physical procedure
Establish slow, controlled execution with explicit safety checks under qualified instruction. Add a bounded coordination or timing challenge only within the task’s safety limits. Return to baseline, identify a weak transition, repair it, and test the whole procedure again after recovery.
Across domains, the standard, constraints, expertise, and stop conditions change. Whole-task understanding, evidence after each run, a concrete next action, decompression, conditional repair, and validation remain the organising logic.
A small pilot makes the framework testable
Choose one task and define the standard before training begins. Record a baseline run, complete two to four AFLS sessions with the same scorecard, then test retention after several days and transfer in a changed context.
Compare accuracy, time, error patterns, self-correction, safety, and the learner’s state. Confidence is useful to record, but it should be compared with observable performance. To evaluate whether AFLS adds value over current training, use a comparable group or condition receiving that training; a before-and-after result alone cannot establish the cause of improvement.
Evidence and current status
AFLS connects several established approaches into an operating sequence: deliberate practice, challenge-point calibration, mental rehearsal, structured debriefing, pressure training, whole-and-part practice, and tests of retention and transfer.
The references below concern those components. They provide context for the framework; they do not validate the complete AFLS sequence or establish that every component works equally well across tasks. Claims about the combined protocol need prospective field testing with clear standards and documented conditions.
The contribution of this paper is to make the observed training logic explicit, usable, and testable. Its value should be judged by what learners can reliably do after training and carry into their next real task.
References
- Ericsson, K. A. (2008). Deliberate practice and acquisition of expert performance: A general overview. Academic Emergency Medicine, 15(11), 988–994. Read the paper ↗
- Guadagnoli, M. A., & Lee, T. D. (2004). Challenge point: A framework for conceptualizing the effects of various practice conditions in motor learning. Journal of Motor Behavior, 36(2), 212–224. Read the paper ↗
- Keiser, N. L., & Arthur, W., Jr. (2021). A meta-analysis of the effectiveness of the after-action review (or debrief) and factors that influence its effectiveness. Journal of Applied Psychology, 106(7), 1007–1032. Read the paper ↗
- Low, W. R., et al. (2021). Pressure training for performance domains: A meta-analysis. Sport, Exercise, and Performance Psychology, 10(1), 149–163. Read the paper ↗
- Fontana, F. E., et al. (2009). Whole and part practice: A meta-analysis. Perceptual and Motor Skills, 109(2), 517–530. Read the paper ↗
- Toth, A. J., et al. (2020). Does mental practice still enhance performance? A 24 Year follow-up and meta-analytic replication and extension. Psychology of Sport and Exercise, 48, 101672. Read the paper ↗
- Czyż, S. H., Wójcik, A. M., & Solarská, P. (2024). The effect of contextual interference on transfer in motor learning — a systematic review and meta-analysis. Frontiers in Psychology, 15, 1377122. Read the paper ↗
