Friendship Calculator

Ten questions, eight categories

It asks you ten questions about a friendship, each answered on a scale from 1 to 10, and turns your answers into a profile across eight categories: trust, communication, loyalty, fun, shared interests, reliability, emotional support and conflict resolution.

The output is a percentage, a radar chart showing the shape of the friendship, a named friendship personality, and a full breakdown showing which of your answers produced which category score.

The important thing about this calculator, and the reason it exists alongside the name-based one, is that its output reflects something real. Not because the scoring is scientific — it is not — but because the input is your own observation of an actual friendship rather than the spelling of two names. Garbage in, garbage out applies; thoughtful answers in, useful reflection out.

Why this is not a name calculator

The friendship calculator by name compares letters. It is entertaining and it tells you nothing.

This one never looks at the names at all. They appear on the result as labels and are otherwise ignored, which you can verify by running the same ratings under different names and getting an identical score.

That difference is the whole point of having both. One is a game; the other is a structured way of thinking about a friendship. Neither is a measurement in any rigorous sense, but only one of them is looking at the friendship.

Who it is for

People who want a prompt rather than an answer. Sitting down and rating ten aspects of a friendship takes two minutes and tends to surface something you had not articulated — usually that one area is much weaker or stronger than the general impression suggested.

Pairs of friends doing it together, which is where it becomes genuinely interesting. Rating separately and comparing produces a conversation that most friendships never have, and the disagreements are more useful than the agreements.

People reflecting on a friendship that has changed. Rating it as it is now, then as it was three years ago, gives two shapes to compare. That comparison is often clarifying in a way that thinking about it in the abstract is not.

How to use it

1. Add names if you want them on the result. Entirely optional and with no effect on the score.
2. Choose who is answering. Leave it on "Just me" for a quick profile, or switch to separate rating if you are doing this with the other person.
3. Move each slider to the number that matches your honest view. Every slider starts at 5, which is deliberately neutral.
4. Press the button and read the radar chart before the percentage. The shape carries more information than the number.

The ten questions

Each is worth a moment's thought, because several of them are easy to answer reflexively and get wrong.

How long have you known each other?

Not a duration in years but a sense of it — 1 for someone you met last month, 10 for someone who has been there most of your life. It feeds loyalty, because shared history genuinely does accumulate.

How frequently do you communicate?

Messages, calls, anything. This is about contact rather than depth, and it is separate from how often you meet on purpose — plenty of strong friendships involve constant messaging and almost no meeting.

Trust level

Whether you can say something unguarded and know it will stay there. This carries the heaviest weight of any single question, appearing in four of the eight categories.

Communication level

How easy conversation is, not how much of it there is. A friendship with infrequent but completely open conversation scores high here and low on frequency, and that is a real and common pattern.

Shared interests

How much overlap there is in what you actually enjoy. Worth rating honestly rather than generously — plenty of excellent friendships score low here, and the calculator weights it least of the eight for exactly that reason.

Reliability

Whether they do what they said they would. Distinct from trust: someone can be entirely trustworthy with a secret and hopeless at turning up.

Emotional support

Whether this is a friendship you can bring a problem to. Along with trust, this is the most heavily weighted question, because it is what separates a friend from a pleasant acquaintance.

How well you handle disagreements

The question people most often mis-rate. A high score means disagreements get aired and resolved — not that you never disagree. Friendships where nothing is ever said should rate low here, however peaceful they feel.

How often you meet in person

Deliberately separate from communication frequency, so long-distance friendships are not penalised twice.

Comfort level

Whether you can be quiet together, arrive unannounced, or be seen at your worst. It is a good proxy for how established a friendship really is.

The eight categories and their weights

Each category blends three questions. The full mapping is below — every one of these numbers is also shown on the result.

The weights encode a straightforward view: trust and emotional support matter most, shared interests matter least. That is an opinion, and it is published rather than hidden so you can weigh the result accordingly if you disagree.

The reasoning is that common interests are excellent at starting friendships and poor at sustaining them. People change hobbies, jobs and cities; friendships that were built only on a shared activity usually do not survive the activity ending. Trust and support are the things that carry a friendship through periods when there is nothing obvious to do together.

How the 1–10 scale is converted

A rating of 1 becomes 0, and a rating of 10 becomes 100:

This matters more than it might appear. The obvious alternative — rating divided by 10 — would make the worst possible answer worth 10%, so a friendship rated 1 on every single question would still score 10% overall. That is not a meaningful floor.

With this conversion, the bottom of the scale is genuinely the bottom. All ones gives 0%, all tens gives 100%, and the default of 5 across the board gives 44%, which is a reasonable representation of "neutral on everything".

Reading the strength profile

The radar chart plots all eight categories at once, each on its own spoke, with the rings marking 25%, 50%, 75% and 100%.

An even octagon means a balanced friendship. Everything is at a similar level, whether that level is high or low.

A spiky shape means a few categories are carrying it. This is extremely common and not a problem — plenty of good friendships are enormous fun and rarely serious, or deeply supportive and rarely social.

A lopsided shape, high on one side and low on the other, usually indicates a particular kind of friendship. High trust and support with low fun and interests describes a friendship built through difficulty. The reverse describes a social friendship that has never been tested.

The shape is more useful than the percentage because it does not collapse eight different things into one. Two friendships scoring 68% can have completely different profiles.

Rating separately

Switching to separate rating gives each person their own slider for every question and unlocks three extra outputs.

Each person's top three areas shows what each of you rates most highly, which is frequently not the same thing.

Shared strengths lists the categories appearing in both top threes. Where there is no overlap, the result says so — that is common and simply means you value different aspects of the same friendship.

The biggest difference of opinion is the single most useful output on the page. It names the category where your two ratings diverge most and gives both numbers. A gap of fifteen points or more is also flagged directly on the category bar.

A large gap is not a warning. It usually means one person has thought about that area more, or has a higher standard for it, or has had an experience the other has not. It is a conversation, not a diagnosis.

Worked example: one person rating

Worked example: two people disagreeing

Rating honestly

The output is only as good as the input, and there are predictable ways people distort their own answers.

Generosity inflation. Rating a friend highly because rating them low feels disloyal. The ratings are private and nobody is being judged; an inaccurate 9 just produces an uninformative result.

Recency bias. A recent argument or a recent good evening can swing several answers. Try to rate the friendship as it has been over the past year rather than the past week.

Confusing frequency with quality. Seeing someone often is not the same as communicating well with them. The two questions are separate because the two things are separate.

Rating the person rather than the friendship. A wonderful person can be an unreliable friend. The questions are about the friendship as it operates, not about their character.

Avoiding the extremes. Clustering everything between 4 and 7 produces a flat, uninformative profile. If something is genuinely excellent, say 9. If something genuinely does not happen, say 1.

What to do with the result

The weakest category is the useful output, not the percentage.

Most of these categories respond to attention. Communication and meeting frequency are almost entirely a matter of someone taking the initiative. Conflict resolution improves the first time something is said out loud rather than absorbed. Shared interests can simply be created by doing something together on purpose.

Two do not respond quickly. Time known cannot be changed at all, and trust moves slowly and mostly through accumulated small evidence. If those are the low scores, patience is the only available strategy.

If you rated separately and found a large gap, that is the place to start. "You rated this a lot lower than I did — what am I missing?" is a genuinely good question, and the calculator has done its job if it prompts it.

Rating a friendship that has changed

One of the more revealing ways to use this is twice: once for the friendship as it is now, and once for how it was at some earlier point. Comparing the two profiles is usually more informative than either on its own.

Friendships change shape rather than simply getting better or worse. A university friendship might have scored very high on fun, shared interests and meeting frequency, and moderate on trust and support simply because nothing had tested them yet. Ten years later the same friendship might score low on meeting and interests and very high on trust and support. The overall percentage could be almost identical while describing two completely different relationships.

This is worth knowing because people often interpret a change in one area as decline. Seeing each other less is not the friendship weakening; it is the friendship occupying a different place in two busier lives. The radar chart makes that visible as a change of shape rather than a drop.

What is worth attention is a drop in trust, support or conflict resolution, since those are the load-bearing categories. A friendship that has become less social but more trusted is usually fine. One that has stayed social while trust has quietly fallen is the pattern worth thinking about.

If you do run it twice, write down the first set of ratings before starting the second. Nothing is stored between calculations, and it is surprisingly hard to remember what you put once you have started thinking about a different period.

Long-distance friendships

The questionnaire deliberately separates how often you communicate from how often you meet, and that separation matters most here.

A friendship conducted entirely by message can legitimately score 9 on communication and 1 on meeting. Those two answers feed different categories at different weights: communication frequency contributes 30% to the communication category, while meeting frequency contributes 20% to it and 30% to fun. So a long-distance friendship takes a hit on fun and shared interests without being penalised across the board.

That is intentional. Distance genuinely does reduce some things — you cannot spontaneously do something together — while leaving trust, support and conflict resolution entirely intact. A close friendship across two continents should not score the same as an acquaintance you see weekly, and with these weightings it does not.

The characteristic long-distance profile is high on trust, support, loyalty and communication, low on fun, shared interests and meeting. On the radar chart it appears as a shape weighted heavily to one side. If that is what yours looks like, it is describing the situation accurately rather than marking you down.

One thing worth rating carefully in a long-distance friendship is conflict resolution. Distance makes it easier to let something go unmentioned, because you are not going to see the person and be reminded of it. Friendships that operate at a distance often score lower here than the people in them expect.

When the profile looks worse than you expected

It happens, and it is usually informative rather than alarming.

The most common cause is that a friendship is carried by two or three strong categories and the other five have never been tested. That is not a weak friendship; it is an untested one. A profile can only reflect what has actually happened.

The second most common cause is a single low answer with wide influence. Trust appears in four categories, so rating it low pulls down trust, loyalty, reliability and conflict resolution simultaneously. If the overall figure surprised you, open the working and look for which single question is doing the damage.

It is also worth checking whether you rated the friendship or the moment. A disagreement in the past fortnight makes conflict handling feel much worse than it is over a longer view.

And sometimes the number is right, and the honest answer is that a friendship has thinned out. That happens to almost every friendship at some point, usually because circumstances changed rather than because anyone did anything wrong. A calculator noticing it is not a catastrophe. If the friendship matters, the low categories are a reasonable list of where to start.

Rating a group rather than a pair

The questionnaire assumes two people, but it works for a small group if you rate the group as a unit. Answer each question about the group as a whole rather than about any individual in it, and the profile describes the dynamic rather than a relationship.

Group profiles tend to look different from pair profiles in a consistent way: fun and shared interests run high, because a group usually forms around an activity, while conflict resolution runs low, because groups are markedly worse at addressing friction than pairs are. Nobody wants to raise something in front of an audience.

If you rate a group and the conflict resolution score comes out low, that is the expected shape for a group rather than a warning about your particular friends. Groups and pairs are different things, and the questionnaire was designed around pairs, so read a group profile as a rough sketch rather than as a considered reading.

What it cannot tell you

Whether a friendship is good. It reflects your ratings back at you in an organised form; it has no independent view.

How your friend sees it, unless they rate it themselves. A single-person profile is one perspective on a two-person relationship.

Anything validated. The categories are ordinary words, the weights are a stated opinion, and no part of this has been tested against research on friendship. It is a structured prompt, not an instrument.

What to do about a low score. That depends on circumstances the calculator knows nothing about.

Common mistakes

Leaving every slider at 5

The default produces 44% for every friendship, which tells you nothing. Move them.

Rating conflict handling high because you never argue

Never arguing is not the same as handling disagreement well. If things go unsaid, that is a low score.

Reading the percentage and ignoring the shape

Two friendships on 68% can be completely different. The radar chart is where the information is.

Treating a disagreement gap as an accusation

It means two people noticed different things. That is normal.

Comparing scores between friendships

Your ratings are relative to your own standards, and those standards shift depending on who you happen to be thinking about at the time. Comparing two of your own profiles against each other is reasonable and often revealing; comparing your figure against someone else's tells you only that two different people used the same slider differently.

Limitations

Everything rests on self-report, with all the bias that carries. The calculator cannot detect a generous rating or a resentful one.

Ten questions cannot capture a friendship. Humour, shared values, practical usefulness, how you handle each other's other relationships — none of these is asked about.

The weights are fixed and cannot be adjusted. If shared interests matter more to you than the 0.9 weighting suggests, the overall figure will not reflect that, though the individual category score still will.

Combining two people's ratings by averaging treats both views as equally accurate, which is the only defensible default but not always true.

The friendship personality comes from a finite table, so uncommon combinations fall back to a description based on the single strongest category.

Frequently asked questions