Red flags began as a reasonable metaphor. In relationship contexts it described genuinely concerning behaviours — patterns of control, contempt, or dishonesty that research reliably links to relationship dissatisfaction and harm. The metaphor was useful because it named a real category: behaviours that are worth taking seriously rather than explaining away, that appear early in relationships and predict later difficulties, and that people in the early stages of attraction have a tendency to minimise.
The metaphor has since been industrialised into a content format, a game, and a personality system. Red flags now include preferring different music genres, not texting back within an acceptable window, having complicated feelings about family, talking too much about an ex, and a long list of behaviours that are at best personal preferences and at worst normal features of complex human beings. Green flags have been added as the inverse, producing a binary scoring system for human beings that treats compatibility as a checklist and relationships as product evaluation. What began as useful shorthand for genuinely harmful patterns has become a mechanism for eliminating people before they become fully visible.
Myth 1: Compatibility Can Be Assessed Through a Flag System
The implicit model behind red and green flag content is that human compatibility is a function of observable surface behaviours that can be catalogued in advance and used to make reliable partner selection decisions. The research on what actually predicts relationship satisfaction tells a different story. The variables most reliably predictive of long-term relationship quality — how two people navigate conflict together, whether they bring out specific qualities in each other, the particular chemistry of their communication styles — are interaction-dependent and largely unobservable from the behaviours that flag content describes.
Research by Eli Finkel and colleagues on speed dating and partner selection found that people’s stated preferences about what they want in a partner predict their actual choices poorly. What people find attractive in interaction differs substantially from what they say they find attractive in advance. The flag system — which is essentially a set of pre-stated preferences applied to early-stage observations — is therefore operating at exactly the stage and level of analysis where predictive validity is lowest.
Myth 2: Red Flags Are Objective Properties of People
The red flag framework treats flagged behaviours as properties of the person exhibiting them rather than as features of a specific interaction between two specific people in a specific context. Someone who texts infrequently is labelled an avoidant red flag, regardless of whether the person evaluating them has a history of anxious attachment that makes normal texting frequency feel insufficient. Someone who talks about a past relationship is labelled emotionally unavailable, regardless of whether the relationship in question ended recently and the conversation is entirely appropriate processing.
What the content industry presents as objective red flags are very often the interaction of one person’s behaviour with another person’s specific sensitivities, preferences, and histories. The same behaviour — making a joke at an inappropriate moment, cancelling plans once, taking time to respond to messages — will be a red flag for some people and entirely unremarkable for others. The objectification of these interactions into person-level properties produces confident judgements about people based on how they interact with one specific person’s specific nervous system, which is a very narrow sample.
Myth 3: Early Elimination Is Protective
The practical application of red flag content is early elimination: the identification of concerning behaviours before significant investment, allowing exit before the cost of attachment makes exit difficult. The protective logic is real — early exit from genuinely harmful patterns is considerably less costly than late exit. But the framework produces a systematic tendency toward over-elimination that has costs of its own.
Human beings are complex, inconsistent, and context-dependent. Most people will produce something a red flag list could flag, especially in the early, high-stakes, anxiety-producing context of early dating. Someone who is nervous will seem distant. Someone who has been hurt will be guarded. Someone who is curious will ask questions that could be read as interrogative. Applying a binary flag system to these early-stage presentations eliminates people on the basis of their worst, most anxious, most performative moments rather than on the basis of who they actually are across time and contexts.
Red flag content systematically trains attention toward reasons to exclude rather than reasons to continue. This produces an asymmetric evaluation process in which negative signals receive more weight than positive ones — a pattern that, applied consistently, can eliminate genuinely compatible people on the basis of their early-stage imperfections while providing no protection against people who are skilled at presenting well in early encounters.
Myth 4: Green Flags Indicate Good Partners
Green flag content is the inverse of the red flag list: a catalogue of behaviours that indicate a healthy, securely attached, emotionally intelligent partner. The list typically includes things like active listening, taking accountability, remembering small details, having close friendships, and speaking well of their exes. These are genuinely positive traits, but the green flag content around them produces a specific problem: the traits that appear on green flag lists are also the traits that are easiest to perform in early dating when the incentive to impress is highest.
Active listening, accountability, remembered details, and warmth toward an ex are all behaviours that someone who wants to seem like a good partner can produce on demand regardless of whether they represent stable character features or situational performance. The green flag system is therefore most predictive of someone’s ability to present as a good partner in early dating — which is not quite the same as being a good partner in an established relationship, when the performance incentive has diminished and the actual character features become more visible.
Myth 5: The Content Is Primarily Protective
Red and green flag content presents itself as protective education — helping people avoid harm and identify healthy relationships. The content ecosystem that has built up around the framework suggests an additional function: entertainment and community formation around shared judgements. Content that presents escalating catalogues of red flags generates engagement through recognition, agreement, and the social pleasure of shared evaluation. The flags become increasingly fine-grained over time not because the research on relationship harm supports finer granularity but because the content format rewards escalation.
The result is a genre that has drifted from its protective origin toward something more like a competitive exclusion sport, in which the person with the most refined and extensive flag list signals the most sophisticated relationship awareness. The sophistication being signalled is not the same as the wisdom being implied, and the relationship outcomes produced by someone who has memorised an extensive red flag taxonomy are not reliably better than those produced by someone who pays attention, moves slowly, and responds to actual behaviour over time rather than to early-stage pattern-matching.
The Framework vs What Actually Predicts Compatibility
| The Flag Framework Assumes | What Relationship Research Shows |
|---|---|
| Compatibility is assessable from surface behaviours | Key compatibility factors are interaction-dependent and emerge over time |
| Flags are objective properties of people | Most flags reflect the interaction of behaviour with a specific observer’s sensitivities |
| Early elimination is reliably protective | Over-elimination has its own costs; early presentation is high-variance and unrepresentative |
| Green flags indicate stable good character | Early-stage positive behaviours are partly performance and highest-incentive-moment outputs |
| More refined flags mean better judgement | Attention to actual behaviour over time outperforms pre-specified pattern-matching |
What Actually Helps in Early-Stage Assessment
- Move slowly enough that early-stage anxiety and performance diminish before making significant assessments
- Pay attention to how someone behaves when things go slightly wrong — small inconveniences or disagreements reveal more than smooth early dates
- Notice whether their stated values and their observable behaviour are consistent over time rather than at a single moment
- Distinguish between behaviour that concerns you because it genuinely conflicts with your values and behaviour that activates your specific anxieties
- Weight how someone treats you across repeated ordinary interactions over how they present in curated early-dating moments
The Genuine Version of the Concern
The original concern that produced red flag content is real. People do have systematic tendencies to explain away concerning behaviour when they are attracted to someone. Early patterns do predict later patterns more reliably than the optimism of early attraction accounts for. Contempt, consistent dishonesty, and controlling behaviour are genuinely worth taking seriously rather than minimising, and naming them makes them easier to identify and act on. None of this requires the content industry’s expanded taxonomy of increasingly granular flags or the gamified scoring system it has produced.
The useful version of this framework would say: pay attention to how people behave when they are uncomfortable, frustrated, or inconvenienced; notice whether they take accountability or deflect; watch for consistent patterns rather than isolated incidents; and distinguish between behaviour that reflects character and behaviour that reflects the high-anxiety context of early dating. That version doesn’t generate much content, because it doesn’t produce a list.
Frequently Asked Questions
Are red flags a useless concept?
No. Behaviours like contempt, consistent dishonesty, controlling patterns, and cruelty in conflict are reliably linked to relationship harm and worth taking seriously early. The concern is with the expansion of the concept to include personal preferences and context-dependent behaviours that the research doesn’t support as predictors of relationship failure.
How do I distinguish a genuine red flag from my own anxiety being activated?
A useful question: would this behaviour concern me if I observed it in a friend’s relationship, or does it specifically activate something in my own history? Genuine red flags tend to be consistent across observers with different sensitivities. Anxiety-activated flags tend to be specific to the interaction between the behaviour and the observer’s particular nervous system.
Isn’t it good to have standards in dating?
Yes. The distinction is between standards grounded in your actual values and the kind of relationship you want to build, and a taxonomic flag system that converts early-dating variance into confident character assessments. Standards are useful when they reflect what you genuinely need; flag lists are less useful when they reflect what the content ecosystem has convinced you to notice.
Why has red flag content become so popular?
It provides a vocabulary for experiences that feel hard to name, creates community through shared recognition, offers a sense of control in an uncertain domain, and generates very high engagement. The format rewards escalation and granularity, which is why the lists get longer over time regardless of whether the additions have research support.
What’s a better way to evaluate early-stage compatibility?
Move slowly enough that early anxiety and performance diminish. Pay attention to how someone behaves when things go slightly wrong. Notice consistency between stated values and observable behaviour over time. Weight repeated ordinary interactions more than curated early-dating moments. And distinguish behaviour that reflects character from behaviour that reflects the context of early dating.
Related Reading
How a clinical framework became a compatibility scoring system.
Why more options produce worse decisions in partner selection.
The relationship advice that conveniently has no finish line.
Why the one is a statistically incoherent concept.
The through-line connecting all these content formats.
On the gap between popular advice formats and evidence.
Where the optimise-everything mythology begins.
Another checklist-based approach to a complex human problem.
Books on Relationships That Don’t Use Flag Systems
“The Seven Principles for Making Marriage Work” by John Gottman
Research on what actually predicts relationship success — based on behaviour over time, not early-stage pattern-matching.
“Mating in Captivity” by Esther Perel
A nuanced examination of what sustains desire and connection — not reducible to a checklist.
“Why Won’t You Apologize?” by Harriet Lerner
On accountability and repair in relationships — what genuine red flags around honesty and responsibility actually look like.
“How to Not Die Alone” by Logan Ury
A behavioural scientist’s approach to dating decisions — evidence-grounded and sceptical of checklist thinking.
As an Amazon Associate, this site earns from qualifying purchases made through the links above.