16 Personality Types Compatibility Chart: How to Read a Pairing

A full chart of the sixteen results is a grid with 256 cells, usually color-coded into ideal matches, good matches, workable pairs and difficult ones. It is the most-shared artifact in the whole four-letter framework, and it is the part with the least behind it. Nobody has demonstrated that the cells predict anything about how a particular couple does.

That does not make the exercise worthless, but it does change what the output is. A pairing cannot tell you whether a relationship will work. It can tell you, fairly reliably, which conversation the two of you will keep having. Someone who prefers closure living with someone who prefers to decide late will have the same argument about plans for years, whatever else is true about them.

So read a chart as a map of predictable friction rather than a verdict. Framed that way, it does something useful: it names the thing early, before the pattern has had five years to become a story about character.

The framework is often searched as MBTI; our test is the 16 Personalities Test, and the pages linked from here describe each result from that test rather than any official instrument.

Energy, the E and I axis, decides what recovery looks like and who runs out first. A mixed pair argues about Saturday: one person is restored by seeing people, the other by not. The workable version is explicit, in advance, with both nights on the calendar. The failure mode is the extravert reading quiet as rejection and the introvert leaving without saying when they will be back.

Information, the S and N axis, is the one most couples underestimate and the one most likely to produce a feeling of not being on the same wavelength. A sensing partner wants the specifics and a plan that survives contact with reality; an intuitive partner wants the possibility and finds detail-first conversations deflating. The repair is translation: say the concrete version and the pattern version of the same point, in that order.

Decisions, the T and F axis, produces a familiar complaint in both directions. The thinking partner is told they are cold when they offer a fix; the feeling partner is told they are illogical when they want the upset acknowledged first. Almost all of it is sequencing. Acknowledgment before solution resolves more of this axis than either person changing.

Structure, the J and P axis, is the practical one: calendars, chores, money, how much notice a change needs. This axis tends to show up every week rather than only in important conversations, which is why it can feel like the biggest of the four. The fix is boring and effective, which is deciding together which commitments are fixed and which are intentions.

Popular sources answer this in two incompatible ways at once. One camp says similarity is comfort: shared letters mean fewer translations and less explaining. The other camp says the ideal pairing is one that matches on the middle letters and differs on energy and structure, on the theory that the two then share the same underlying functions in a different order.

That second theory is where the familiar ideal pairs come from: INFJ with ENFP, INTJ with ENTP, INFP with ENFJ, INTP with ENTJ, and the sensing equivalents. It is an internally consistent story about the framework's own machinery. It has never been tested against relationship outcomes in a published study, and it should be read as folklore about preferences rather than a finding.

The research on similarity itself is clearer than either camp. Montoya, Horton and Kirchner's 2008 meta-analysis of 313 studies found that actual similarity predicted attraction in first-encounter and short-interaction settings but had no significant effect in existing relationships, while perceived similarity predicted attraction in all three. In other words, feeling similar to your partner matters; being measurably similar, once you are already together, does not.

Dyrenforth and colleagues put a number on it in 2010, using three nationally representative samples of couples. After accounting for each person's own traits and their partner's traits, couple similarity explained less than half of one percent of the variance in relationship satisfaction. A person's own traits accounted for about 6 percent of the variance by comparison.

The table below takes the ten pairings that get searched most often, which are mostly the ones popular sources name as ideal, and gives the honest version of each: what tends to work and what tends to grate. Neither column is a prediction. Both are the recognizable shape of that pair's good weeks and bad ones.

Read the row for your own pairing, then read the axis section above for whichever letters differ. If your pairing is not in the table, every result has its own compatibility page with the other fifteen set out in full.

The largest analysis available ran machine learning across 43 longitudinal datasets covering 11,196 couples, published by Joel and colleagues in 2020. The strongest predictors of relationship quality were all about the relationship rather than the people in it: perceived partner commitment, appreciation, sexual satisfaction, perceived partner satisfaction, and conflict. The strongest individual predictors were life satisfaction, negative affect, depression and attachment style.

Two findings from that paper are worth holding onto. First, relationship-specific variables predicted up to 45 percent of the variance in quality at baseline, while individual differences predicted about 21 percent. Second, what a person reported about their own relationship predicted two to four times more than anything their partner reported, and partner reports added nothing once the person's own report was in the model.

Nothing in that list is a type. The things that moved the numbers are all learnable and all available to any pairing of letters: noticing and saying what you appreciate, treating a complaint as information rather than an attack, repairing quickly after a bad exchange, and taking your own low mood seriously instead of routing it through the relationship.

This is why a chart that sorts couples into good and bad cells is worse than useless if it is believed. Two people in an ideal cell who never repair will do badly. Two people in a difficult cell who repair well will be fine, and will have to translate more while doing it.

For the intuition and thinking results, INTJ Compatibility covers the Architect, INTP Compatibility the Logician, ENTJ Compatibility the Commander and ENTP Compatibility the Debater. These four share a habit of engaging a partner's problem as a problem, which is a strength and the origin of most of their relationship complaints.

For the intuition and feeling results, INFJ Compatibility covers the Advocate, INFP Compatibility the Mediator, ENFJ Compatibility the Protagonist and ENFP Compatibility the Campaigner. These four tend to give more than they ask for and to raise a need late, so the recurring theme across all four pages is stating things plainly and earlier.

For the sensing and judging results, ISTJ Compatibility covers the Logistician, ISFJ Compatibility the Defender, ESTJ Compatibility the Executive and ESFJ Compatibility the Consul. These four show care through reliability and taking things on, and the shared risk is a private ledger of what they have carried that the other person never knew was being kept.

For the sensing and perceiving results, ISTP Compatibility covers the Virtuoso, ISFP Compatibility the Adventurer, ESTP Compatibility the Entrepreneur and ESFP Compatibility the Entertainer. These four are strong in a crisis and in the present moment, and the shared challenge is the slow unglamorous maintenance that keeps a relationship steady between events.

That some pairings are doomed. None of the 256 cells describes an unworkable relationship, and treating a cell as a warning is a way of giving up early on something ordinary. The cells that popular charts mark red are usually the pairs who have to translate the most, which is work rather than incompatibility.

That opposites attract, or that they do not. Both claims are stated confidently and neither survives the evidence. Perceived similarity helps and measured similarity stops mattering once you are in the relationship, which means the useful question is not how alike you are but whether you feel understood.

That a chart can tell you whether to leave. It cannot, and that is worth saying plainly because it is a real use people put these grids to. If a relationship is distressing you, the questions that help are about how conflict goes, whether repair happens, whether you are treated with respect, and whether you are able to say what you need. Type is not on that list.

That your result is fixed. A meaningful share of people get different four letters on a retake, usually on the axis where their preference was close to the middle. If you and a partner have each tested twice with different results, that is the measurement being imprecise, not a change in either of you.

The four-letter framework is a popular typology with limited test-retest reliability and no strong predictive validity. McCrae and Costa's 1989 reanalysis showed that the underlying traits behave as continuous dimensions rather than the either-or categories the letters imply, and Pittenger's 2005 review found that the framework does not forecast real-world outcomes with the reliability its popularity suggests.

On four-letter pair compatibility specifically there is no validating literature: research on typology and couples exists, but no study validates a pairing chart like this one. There is no published study showing that any four-letter pairing predicts satisfaction, stability or conflict frequency. The absence is not evidence against the idea, and it does mean that every confident compatibility chart in circulation is an opinion presented as data.

What exists instead is a substantial literature on what does predict how couples do, which is summarized above. That research is the reason this page is organized around friction and repair rather than around matching.

Psychology.com's free 16 Personalities Test takes a few minutes and reports where you fall on each of the four axes rather than only returning four letters, which is the more useful output for this purpose: a near-middle score on an axis tells you that axis is unlikely to be the source of a recurring argument. An optional in-depth report is available for readers who want a longer write-up.

Once you have both results, the practical sequence is short. Find the letters that differ, read the axis section on this page for each one, then read your own compatibility page for the detail. The 16 personality types index covers the four axes and every result if you want the model itself first.