SBTI Hidden Types: How to Unlock DRUNK and HHHH

To understand DRUNK, you need to know a piece of backstory that most people never hear.

SBTI's existence is directly tied to a "please stop drinking so much" story. A content creator had a friend who drank too much — not casual social drinking, the kind that makes people worry. Talking didn't help. Sharing health articles didn't help. Forwarding liver disease case studies definitely didn't help. So the creator tried a different approach: build a personality test website, but bury a trap inside it. If the friend revealed during the test that they were a drinker, the test would skip everything else and stamp them as DRUNK (The Drunkard), then deliver the result in a half-joking, half-serious tone: "Your core personality trait is: alcoholic."

DRUNK was the very first SBTI type ever designed. Before CTRL, before BOSS, before DEAD — there was only DRUNK. The other 26 types were added later to make the whole thing look like a legitimate product. The entire SBTI personality test originated as a friendship-driven drinking intervention prank. That might be one of the most absurd product origin stories on the internet.

Did the intervention actually work? According to the creator, the friend did find it funny — but whether they actually drank less is a separate question. Either way, DRUNK survived as an "origin Easter egg," becoming the most story-rich type in the entire SBTI system.

DRUNK isn't triggered during the 30 main questions. It's triggered through bonus questions that appear after the main test is complete. Here's the exact flow:

After you finish all 30 questions, the system presents a "gate question" about drinking habits. This question has 4 options, covering a range from "don't drink at all" to "drink regularly." If your answer trips the "drunkard path," the system immediately presents a second confirmation question with only 2 options. This is a "are you sure?" design — giving you one chance to back out.

If you confirm on both questions, the system stamps you as DRUNK and voids every answer from the previous 30 questions. No pattern matching. No radar chart comparison. No similarity calculation. Just a hard cut to the DRUNK result. The bluntness of this logic is itself a joke: you thought you were taking a serious personality test? Sorry, you drink, and nothing else matters.

From a technical standpoint, DRUNK is the only type that bypasses the entire scoring pipeline. All other 26 types go through the full "answers → dimension scores → pattern matching" flow. DRUNK takes a shortcut. In the code, a dedicated drink_trigger flag fires, and the normal scoring function's entire execution path gets skipped.

In the share URL, DRUNK shows up as the 16th digit in the result encoding string. That digit is either 0 (not DRUNK) or 1 (DRUNK). When someone opens a DRUNK share link, the system reads that flag and renders the DRUNK result page directly — no need to reconstruct dimension scores or run matching calculations. The other 15 digits in the string still contain your raw dimension scores from the 30 questions, even though they were never used for type assignment. So technically, if you triggered DRUNK but are curious what your "real" type would have been, the data exists in your share URL — someone who knows the encoding format could decode it.

If DRUNK is a hidden type you can choose to trigger, HHHH is an unexpected verdict the system hands you.

HHHH (The Silly Laugher) has exactly one trigger condition: your 15-dimension pattern, when compared against all 25 regular type templates using Manhattan distance, produces a maximum similarity below 60%. In plain language: you don't really look like any of the 25 types. You're too "non-standard" for any existing template to claim you.

That sounds cool, but HHHH is extremely rare in practice. The reason: the 25 regular type patterns already cover a very wide range of dimension combinations. Think about the math — 15 dimensions, each with 3 possible levels (L/M/H), gives a theoretical space of 3^15 = 14,348,907 combinations. But because Manhattan distance works by summing per-dimension differences, you only need to be "roughly similar" to a type across most dimensions to clear the 60% threshold. 25 carefully selected standard types blanket the vast majority of realistic personality patterns.

So what answer pattern actually triggers HHHH? Usually something extreme AND contradictory across multiple models. Example: Self Model all H (extremely confident), but Emotional Model all L (extremely avoidant); Action Drive all H (hyper-driven), but Attitude Model all L (deeply pessimistic). These combinations are rare in actual humans because psychological dimensions tend to correlate — most people don't max out confidence while simultaneously bottoming out on emotional engagement.

By the numbers, HHHH has the lowest user count of all 27 types. Its existence is primarily an engineering safeguard — ensuring that no matter how unusual your answers are, the system always produces a result instead of a "type not found" error.

This is one of the most-asked questions, especially from completionists who've collected a few regular types and want the full set.

DRUNK: 100% controllable. All you need to do is select the drinking-related options in the bonus questions after the main test. The 30 main questions don't affect DRUNK triggering at all. You could answer all 30 questions by picking the first option every time, then select the drinking options in the bonus round — DRUNK, done. Zero randomness. Completely deterministic.

HHHH: Theoretically possible, practically very difficult. You need to make your best match across all 25 regular types fall below 60% similarity. The problem: while you're answering, you can't see the real-time matching calculations. You don't know your current distance from each type. Randomly "mashing buttons" probably won't work — random answers tend to produce middling scores that end up matching something like OJBK (The Whatever), which has a lot of M values in its pattern.

One approach: go extreme and contradictory. Max out all Self-dimension scores, bottom out all Emotional-dimension scores, then alternate high and low across Attitude and Action dimensions. But even this doesn't guarantee you'll dodge every single type's 60% threshold — some types have inherently extreme patterns (like DEAD, which is almost all L), and your extreme answers might accidentally match them.

The honest truth: the most "legitimate" way to get HHHH is to genuinely have a psychological pattern that falls outside all 25 templates. Not performed extremes, but a real combination that happens to land in the gaps between types. That's why HHHH carries a certain badge of honor in the community — not because it's better, but because it represents a genuine "unclassifiable."

DRUNK and HHHH give the SBTI test system more depth than most people realize.

DRUNK is a reminder that this test didn't start as a serious psychological project — it started as one person's concern for a friend, wrapped in a joke. When you trigger DRUNK, you're retracing the exact path the original creator designed when the whole thing was just a prank webpage for one specific person. In a way, DRUNK is SBTI's "root." The other 25 types are branches that grew from it.

HHHH represents a form of design humility: admitting the model has limits and some people won't fit any label. Most personality tests force-assign you a result no matter what — even if the match is only 30% similar, they'll shove you into a category and call it a day. SBTI set a 60% floor. Below that, it openly says "our templates can't capture you." This kind of honesty is almost unheard of in entertainment tests, but it's precisely this transparency that makes HHHH the most curiosity-inducing type in the system.

Together, the two hidden types also illustrate the range of the SBTI design. DRUNK is deterministic — two specific answers, guaranteed outcome, no math involved. HHHH is probabilistic — it emerges from the entire 15-dimension scoring pipeline only when no standard template clears a specific numerical threshold. One is a hard-coded Easter egg from the project's origin. The other is an emergent property of the matching algorithm. Both exist in the same system, and that combination of intentional humor and genuine structural depth is what makes SBTI more interesting to pick apart than your average internet personality quiz.

Next time someone asks "isn't SBTI just some random internet quiz?" — tell them about the drinking intervention origin story and the 60% HHHH threshold. Those two details are enough to show that this test is more thoughtfully designed than it might appear on the surface.