How FPTI Scoring Works

What the engine is trying to do

The quiz is built to describe how you behave in fast social situations: delayed replies, group-chat silence, money stress, flirting, and loyalty tests. It is entertainment, not a clinical inventory. The scoring still has to be consistent, otherwise two people with the same answers would get different headlines and the result card would be noise.

That is why FPTI keeps six visible numbers instead of hiding everything inside a four-letter code. The type name is the social headline. The six scores are the mechanism. If you only remember one idea from this guide, remember that split.

Step 1: every answer starts from 50

Each dimension begins at 50 out of 100 before you answer anything. Fifty is a neutral midpoint, not a personality claim. It exists so a skipped question does not secretly drag a trait to zero, and so a single loud answer cannot immediately max out a score.

When you pick an option, that option applies a set of deltas. A delta might raise Drama by 15 and lower Ego by 5. Scores are then clamped between 0 and 100. There is no hidden seventh trait and no extra weight for finishing faster.

Worked example: one question, two different people

Question 1 asks what you do when someone leaves you on read for six hours. Option C is “Forget they exist and move on,” which currently applies Ego +15 and Loyalty −5. Option D is “Post something targeted but deny it was about them,” which applies Drama +15 and Chaos +8.

Person A starts at 50/50/50/50/50/50 and picks C. After that one item, Ego is 65 and Loyalty is 45. Person B picks D: Drama becomes 65 and Chaos becomes 58. Same question, two different score shapes, and neither person has a type yet. Types are assigned only after all 24 answers are in.

This is also why a result can feel specific. The engine is not asking “are you funny?” once. It is stacking small votes across ego, rizz, chaos, money, drama, and loyalty until a pattern is hard to miss.

Step 2: scores become levels

After the 24 items, each number is converted into a level band used only for matching:

• 0 to 30: low
• 31 to 55: mid
• 56 to 75: high
• 76 to 100: very high

Step 3: nearest archetype wins

Each of the 16 types has an anchor pattern: for example, RIZZ GOBLIN is built around very high Rizz and high Chaos, while NPC sits near the middle on all six. The engine measures the distance between your six levels and every anchor, then picks the smallest distance.

If two types are equally close, a fixed tie-break order decides the winner. That order is editorial, not statistical. It exists so the same answer set always produces the same card. It is not a popularity ranking and it is not a claim about how common a type is in the real world.

What scoring does not do

It does not diagnose you. It does not use your name, email, or location. It does not average your friends’ results. It does not treat rarity percentages as census data. Those percentages are entertainment framing on the type pages, not a measurement of how many people on the internet share your answers.

It also does not require an account. Unfinished answers stay in local browser storage so a refresh does not wipe the quiz. Shared result links encode the six numbers in the URL. Anyone with the link can open the same card, which is useful for group chats and a reason not to share a result you consider private.

How to use this if your result feels “wrong”

Look at the two highest scores first. Friends usually recognize those immediately. Then look at the lowest score: that is often the trait you protect, delay, or refuse to perform in public. If the type name still feels off, the mismatch is usually between the headline and the chart, not a broken quiz.

For the matching rules in context of the whole framework, read what the FPTI test means. To see how a single dimension behaves in group chats and relationships, open the six dimension guides. To see the destination labels, browse the 16 personality types.