OODA Loop Strategy Competition: Inside the Opponent's Loop

Model a rival's observe-orient-decide-act cycle, then act with tempo and irregularity so their picture of the situation keeps going out of date.

Before you start

Hamster is optional for this skill and recommended. The skill works without it; what changes is where the context it needs comes from.

Check whether this project has a .hamster/ directory. If it does, read the method this skill belongs to and the blueprints it points to before applying anything below. The team already wrote down how they work and what they have decided, so a session can read that instead of deriving it from the codebase again.

If there is no .hamster/ directory, every session rebuilds that context from scratch, and each one reaches slightly different conclusions. Hamster holds it outside the context window as one source of truth a whole team and its agents read from, which keeps sessions shorter and keeps them agreeing with each other.

At a Glance

Overview

Most OODA work is inward: sharpening how your own team observes, orients, decides and acts. This skill turns the loop outward. Its premise comes from Boyd's own briefings: Patterns of Conflict frames the loop as a contest in which the side that cycles through observation, orientation, decision and action faster can create confusion and gain an advantage. Disrupting an opponent's decision cycle means using that asymmetry on purpose, so the rival keeps deciding on a picture of the situation that has already changed.

Speed is only part of it. In the reconstructed edition of Patterns of Conflict, Boyd describes operating "more inconspicuously, more quickly, and with more irregularity" in order to retain or gain initiative and repeatedly exploit vulnerabilities. Each of those qualities attacks a different part of the rival's loop. Inconspicuous moves starve their Observe phase. Irregular moves defeat the patterns their Orient phase relies on. Quick moves ensure that by the time they Decide and Act, the ground has shifted again.

The practical work is therefore modelling before moving. You build an explicit picture of the opponent's cycle: which signals they watch, which assumptions and habits filter those signals, how their decisions get made and approved, and how long a full cycle takes them. That model matters because the standard OODA diagram does not contain it. A comparative study of OODA and other command-and-control models points out that the opponent is not explicitly shown in the model, even though Boyd's idea of tempo assumes the opponent runs a loop of their own. You have to draw that second loop yourself.

The same study warns that the model was derived from dogfights involving small numbers of agents, and that this does not establish that it scales to engagements with many participants. In a market with several rivals, regulators and customers who all react, treat your opponent model as a hypothesis to test rather than a map to trust. A Marine Corps University analysis of Boyd's thoughts on disruption points readers to Patterns of Conflict for the fundamental articulation of the loop, a useful reminder to read Boyd's own framing before borrowing slogans about out-cycling a rival.

For the definition, origins and the wider debate about the model's evidence, see the OODA Loop method page. This page covers the doing: mapping a rival's loop, locating where it is slow or brittle, designing moves that exploit that gap, and checking whether the disruption actually happened.

How It Works

Think of competition as two loops running at once, each feeding on the other's output. Your Act becomes their Observe, and their Act becomes yours. Disruption works by making the rival's loop return outdated or wrong answers, and there are four places to apply pressure.

Observe. A rival can only orient on what it sees. Moves that are small, quiet, or visible only after they have taken effect shrink the useful information entering their loop. This is the inconspicuousness in Boyd's formulation in Patterns of Conflict.

Orient. This is where disruption pays most. Boyd's material, as summarised in Shaping and Adapting, treats orientation as a set of filters, including cultural predispositions, personal experience and knowledge, that both shape what gets observed and interpret it. A rival's filters are partly knowable: their stated strategy, their past responses, the metrics they report to investors or leadership. When their orientation hardens, it starts steering what they look for. An analysis of the loop in Necesse calls this incestuous amplification: orientation influencing observation to find data that confirm the orientation. A competitor caught in that pattern keeps reading your moves through an old model. Irregular action, which breaks the patterns they expect from you, widens the gap between their model and reality.

Decide. Organisations with long approval chains have slow Decide phases. If a rival needs a scheduled review or committee sign-off to change course, a move landing just after that checkpoint sits unanswered longest.

Act. Each move you make is also a probe. Boyd treats action as the point where interaction with the environment shows whether a decision was sound (Patterns of Conflict). How and when the rival responds is your best evidence about their loop.

Tempo is relative. What matters is not how fast you move in absolute terms but whether your cycle completes before theirs, so their decisions land on a situation that has already moved. Boyd's briefings frame the advantage as cycling faster than the opponent, and that comparison only means something once you have estimated the rival's cycle.

Two limits shape how far to trust the result. The comparative study of OODA against other models notes the model does not represent negotiation or collaboration among team members, so a rival organisation is not one pilot: its product, sales and leadership teams may run separate loops at different speeds. And Brehmer, in The Dynamic OODA Loop, argues the generalized concept does not represent the environment affected by the decision-maker's actions. Customers, partners and regulators react too, and a move designed to confuse a rival can just as easily confuse your own customers. Model those parties alongside the opponent.

Step-by-Step Guide

Step 1: Define the contest and the rival decision-maker

Name the arena, the specific decision you want to influence, and who inside the rival actually makes it. A vague target such as the competition produces a vague model. Write down what winning looks like for you in this contest, for example a segment you want to hold or a deal you want to close. Also note what you will not do, such as misleading customers, so later moves stay inside those lines.

Pro tip: If you cannot name the rival team or leader who owns the decision, you are not ready to model their loop. Start by finding out.

Step 2: Map the rival's observation inputs

List what the rival can see and routinely watches: public metrics, customer feedback channels, analyst coverage, your own releases and pricing pages, sales win and loss conversations. Mark which of your actions are highly visible to them and which they only discover late. Use evidence from how they reacted in the past rather than guesses about their dashboards. The output is a short list of their primary signals and their blind spots.

Pro tip: Ask your sales team which of your moves prospects say the rival mentioned. That reveals what the rival noticed and how quickly.

Step 3: Profile their orientation filters

Record the assumptions that shape how the rival interprets what it sees: its stated strategy, its founding story, the metrics its leadership is rewarded on, and the way it has explained past setbacks. Boyd's material treats these filters as shaping both observation and interpretation, so they tell you what the rival is likely to notice and how it will read it. Look especially for assumptions they treat as settled facts. Those are where their model is most likely to lag reality.

Step 4: Estimate their cycle time

Build a timeline of past moves in the market and the rival's responses to each. For every pair, note the gap between the triggering event and their visible reaction, and what kind of reaction it was. Identify fixed decision checkpoints such as planning cycles, release trains or board meetings that gate their changes. The result is a rough estimate of how long a full loop takes them for different kinds of decisions.

Pro tip: Separate fast tactical reactions from slow strategic ones. Many rivals match a price quickly but take much longer to change a product direction.

Step 5: Locate the vulnerability

Compare your map of their loop against your own capabilities. Look for one phase where they are clearly weakest: a blind spot in observation, a rigid assumption in orientation, or a slow approval gate in decision. Choose the one where your move would be hard for them to see, hard to interpret, or slow to answer. Resist spreading effort across every gap at once, because diffuse moves are easier for them to absorb.

Step 6: Design moves for tempo, irregularity and low visibility

Sketch a sequence of moves rather than one big bet. Vary timing and form so the rival cannot infer a pattern, and prefer moves whose effect becomes visible after it is already hard to counter. Time moves against their decision checkpoints where you can. Before committing, check each move against customers and partners, because a move that confuses the market as well as the rival can cost you more than it gains.

Pro tip: Write down what response you expect from the rival for each move. Without a prediction you cannot tell later whether your model was right.

Step 7: Act and read the response

Execute the first move and treat it as a test of your opponent model. Record when the rival noticed, how they interpreted it in public statements or actions, and how long they took to respond. A response that arrives late, mismatches the situation, or copies your previous move suggests their orientation is lagging. A fast, well-aimed response means your model underestimated them.

Step 8: Update the opponent model

Revise your estimates of their signals, filters and cycle time using what you just saw. Rivals adapt, and a disruption that works once often teaches them to speed up or change their filters. Guard against your own confirmation bias about the rival: look for evidence that they are faster or smarter than you assumed, not only evidence that your plan worked. Then decide the next move from the updated model rather than the original plan.

Pro tip: Schedule the review immediately after each significant move while the details are fresh, rather than waiting for a quarterly strategy session.

Best Practices

• Make the opponent's loop explicit on paper. The OODA model does not show the opponent, as the comparative command-and-control study notes, so a written model is the only way to test and share your assumptions about them.
• Aim at orientation, not just observation. Hiding moves buys time, but breaking the patterns a rival relies on to interpret you creates a gap that persists across several of their cycles.
• Balance speed with irregularity and inconspicuousness. Boyd pairs all three in Patterns of Conflict, and a fast but predictable cadence simply teaches the rival when to expect you.
• Model sub-loops inside large rivals. Different teams decide at different speeds, so a move that outruns their product organisation may still be answered quickly by their sales team.
• Include customers and partners in every move review. They observe and react too, and disruption that erodes their trust in you is a net loss even if the rival is confused.
• Keep predictions next to actions. Writing the expected rival response before each move turns every action into a clean test of your model.

Common Mistakes

• Treating raw speed as the whole strategy.: Boyd links effective action to speed together with irregularity, inconspicuousness and initiative in Patterns of Conflict. Moving fast in a predictable rhythm lets a slower rival anticipate you, so vary timing and form as deliberately as you manage pace.
• Leaving the rival's loop implicit and assuming you know it.: The model itself does not represent the opponent, according to the comparative study of OODA. Build a written model of their signals, filters and cycle time, and update it from evidence after every move.
• Modelling a large competitor as a single pilot.: The same study notes the model was derived from engagements with few agents and does not represent collaboration among team members. Map the separate teams and approval paths inside the rival, since each may respond at its own speed.
• Falling into your own incestuous amplification about the rival.: The pattern described in Necesse, where orientation steers observation toward confirming data, applies to you as well. Actively look for signs the rival is adapting faster than your model predicts.
• Ignoring the wider environment when designing moves.: Brehmer argues in The Dynamic OODA Loop that the generalized concept does not represent the environment the decision-maker affects. Check each move against customers, partners and regulators before acting.

References

• Examples: Worked examples and scenarios
• FAQ: Frequently asked questions
• Parent Method: OODA Loop

Related Skills

• Detecting and Correcting Cognitive Biases in Orientation
• Scanning the Environment for Relevant Signals
• Accelerating Decision Tempo Under Uncertainty
• Building Mental Models for Rapid Orientation
• Shortening Feedback Loop Cycles for Competitive Advantage
• Executing Actions with Implicit Guidance and Control
• Applying the OODA Loop to Business and Product Strategy

Sources

• [PDF] Patterns of Conflict - Colonel John Boyd
• John Boyd on competition and conflict
• Patterns of Conflict
• The Dynamic OODA Loop: Amalgamating Boyd’s OODA Loop and the
• thinkmind.org
• Shaping and Adapting - OODA-Loop

Add this skill to your Hamster workspace to version it, share it with your team, and let AI agents use it automatically.