Guide 8 of 10 · Psychology · By Match Academy Editorial Team · Updated 2026-09-17
Match Academy is an independent training platform. We have no partnership with Tinder, Hinge, Bumble or any other dating app; brand names are used only to describe how those apps work, and trademarks belong to their owners. Descriptions of app features, tiers and prices reflect our understanding at the time of writing and can change at any time — check the app itself. Reply rates and other figures are coaching experience, not statistics published by the apps.
Seven findings from attraction research — similarity, reciprocity, self-disclosure, the pratfall effect, peak-end, anticipation and shared arousal — each translated into what to actually do in a chat, and what not to do with it.
Attraction research has a public-relations problem. The findings are solid — decades of them, replicated, boring in the way good science is boring — and almost every popular summary turns them into tricks. "Use scarcity." "Push-pull." "Never text first." The research doesn't say that. It says something more useful and less cynical: attraction follows a handful of patterns, most of which reward being honest, specific and a little brave.
This guide takes seven of those patterns and does two things with each: says what the research actually found, and translates it into what to do in a dating-app conversation. Each one also has a don't — the way the principle gets abused — because that's where the pickup version lives, and because the abused version reliably backfires. The Match Academy coach was built on this material; its flirt balance score is essentially a measure of whether you're using these principles or performing them.
The oldest finding in the field: we like people who are like us. Donn Byrne's attitude-similarity experiments in the 1960s showed that liking rises almost linearly with the proportion of attitudes two strangers share. Later work sharpened it. What predicts attraction isn't shared hobbies — it's shared values, shared humour, and the sense that you see the world the same way. Two people who both climb but find different things funny will bore each other; two people with nothing in common on paper who laugh at the same absurdity will not.
Do: look for the "me too" moment that's about how you think, not what you do. The way they describe their sister, what they find ridiculous, how they treat the waiter in a story — those are the similarities that compound. Pick up their language a little: their phrasing, their rhythm, their emoji habits. Mirroring is mostly unconscious and reads as we're in sync.
Don't: fake shared interests. Claiming to love a band you've never heard has a shelf life of one date. And don't confuse agreement with similarity — cheerfully disagreeing about pizza toppings while agreeing about how to treat people is far more attractive than nodding along to everything.
Knowing someone likes you makes you like them more. This sounds obvious and is routinely ignored by everyone playing it cool. The refinement matters, though: Aronson and Linder's gain-loss work found that liking which becomes apparent — warmth that builds — is more powerful than liking that was total from the start. Not because indifference is attractive, but because warmth that's earned feels like it's about you, and warmth that's automatic feels like it's about anyone.
Do: show interest clearly and specifically. "You ask better questions than anyone I've matched with this year" is reciprocity done right — it's about them in particular. Let your warmth be visibly a response to who they turn out to be, which is what naturally happens if you're actually paying attention.
Don't: perform indifference. Waiting three hours to reply to a message you read in three seconds, pretending you're not interested, withholding compliments as a strategy — these mimic the gain pattern by faking a low start, and people can feel the fake. The research reward goes to warmth that grows, not to coldness that thaws on schedule.
Arthur Aron's 1997 study — the one behind the famous "36 questions" — had strangers ask each other progressively more personal questions for 45 minutes. Many left feeling closer than they did to some long-term friends. The active ingredients were two: the disclosure escalated gradually, and it was reciprocal. One person confessing to a silent partner doesn't produce closeness; two people going one step deeper at a time does.
Do: go one rung deeper than the conversation currently is, and then wait for them to match. Answer "what do you do?" with the job and one true thing about how you feel about it. Ask why, not just what. The depth ladder is the practical version: facts, preferences, stories, feelings, one step at a time.
Don't: skip rungs. A childhood wound at message five is not intimacy; it's a weight handed to a stranger. And don't interrogate — asking deep questions while revealing nothing yourself breaks the reciprocity that makes the whole mechanism work.
One rung deeper, then wait
Coach's read: Nothing here is deep. It's one step past small talk on both sides, which is exactly the pace that builds closeness without anyone feeling exposed.
In Elliot Aronson's 1966 experiment, listeners rated a highly competent person more likeable after hearing them spill coffee on themselves; the same clumsiness made an average person less likeable. The finding has held up and grown: recent work on what researchers call the "beautiful mess effect" shows that we judge our own moments of vulnerability harshly and other people's as courageous and endearing. The nerves you're embarrassed about read, from the outside, as honesty.
Do: own a small flaw lightly. "I'm going to be honest, I rewrote this message four times" or "I'm terrible at first dates and excellent at second ones" — said once, without fishing — makes everything else you say more credible. It's also the psychological reason self-deprecating openers work.
Don't: make it a pattern. One pratfall is charming; a running commentary of self-criticism asks the other person to keep reassuring you, and the effect flips. And the flaw has to sit on top of competence — the study's whole point. Coffee on the shirt of someone who's clearly good at their life is endearing. Coffee on the shirt of someone who's told you three times they're a mess is just the mess.
Kahneman's peak-end rule: people don't remember experiences as averages. They remember the most intense moment and how it ended, and they barely weigh the duration at all. A long conversation that slowly thinned out is remembered as thin. A shorter one with one genuinely funny exchange that ended while it was still good is remembered as great — even if the total amount of good in it was less.
Do: engineer one peak per conversation — a real laugh, a moment where someone says something true — and then end on it. "Okay, I have to go be an adult. Tell me the rest of the barista story tomorrow" leaves them holding the peak. The tomorrow is the anticipation principle, below, doing free work for you.
Don't: keep a conversation alive until it dies. The instinct when it's going well is to hold on; the research says let go while it's good. And don't end on logistics or a dead-end "haha yeah" — if the last message is flat, the memory is flat.
This is the principle the pickup world built its entire business on, so it's worth getting exactly right. A 2011 study by Whitchurch, Wilson and Gilbert found that women were more attracted to men whose level of interest in them was uncertain than to men they knew liked them a lot — the not-knowing kept the men on their minds. Combine that with a robust general finding that anticipation of a reward is often more pleasurable than the reward itself, and you have the scientific basis for "play hard to get".
Except that isn't what it shows. In the study, the uncertainty was real — participants genuinely didn't know. Manufactured uncertainty — deliberately delayed replies, withheld warmth, calculated aloofness — is detected as a tactic far more often than people think, and once detected it destroys the one thing attraction can't survive without, which is trust. What the research actually supports is much simpler: a person with a real life, real plans and real limits on their availability is more attractive than a person who is always instantly there. Not because they're playing a game. Because they have somewhere to be.
Do: have the life. Make plans you'd keep whether or not the chat goes well. Reply when you have time and attention, not the instant your phone buzzes out of anxiety. Leave a thread for tomorrow instead of exhausting it tonight. Let the date be the place you find out more, rather than telling everything in text.
Don't: schedule your coldness. Setting a timer before replying, deliberately ignoring a good message, pulling back after a warm exchange to "keep them guessing" — these are the tactics we specifically flag in coach reports as flirt-balance failures. They occasionally work on someone anxious enough to chase, which is a bad filter for who you end up with.
Real anticipation versus manufactured scarcity
Manufactured: (reads her message at 7:02, sets an alarm, replies at 10:15 with something short and cool so she wonders) — The behaviour is fake, and fake behaviour is fragile. If she notices — and people notice — every warm thing you've said gets re-read as strategy.
Real: "Just seen this — was at five-a-side and I'm now eating chips in a car park like a champion. The barista story deserves better than car-park typing. Tomorrow?" — You were unavailable for a real reason, you said so, and you set up tomorrow. That's anticipation without a single manufactured second.
Dutton and Aron's 1974 "shaky bridge" study: men who'd just crossed a high, swaying suspension bridge were more likely to later contact an attractive interviewer than men on a low, stable one. The standard interpretation is misattribution of arousal — a racing heart from the bridge gets read as a racing heart about the person. The broader, better-supported version is that shared activation of any kind — laughter, mild challenge, novelty, movement — bonds people faster than calm conversation.
Do: in text, create mild activation through play — a challenge ("defend that opinion in one sentence"), a game, a bet, banter that makes both of you slightly quicker. For dates, pick something with movement or novelty over a static dinner: a walk, a market, an arcade, a climbing wall if you both climb. It's one of the reasons the date-idea matrix is weighted towards activities.
Don't: engineer stress. The principle is mild activation. A first date at an extreme-sports venue, a chat that turns into a debate, deliberately winding someone up — that's arousal with the wrong label, and the label it gets is "this person is exhausting".
Push-pull — a compliment followed by a small rejection, or the reverse — is the most famous "technique" in dating advice, and it's a real pattern in flirting between people who already like each other. Our issue isn't that it exists; it's that as a deliberate tactic it's a manufactured version of principles 2 and 6, and manufactured versions of those principles are precisely what erodes trust. Chapter 5 of the journey includes a scenario called The Push-Pull, but you play the person on the receiving end, and the drill is to recognise it without matching it. The flirting guide covers what actually works instead: teasing about choices, warmth that's specific, a future you casually put them in.
| Moment | Principle at work | What it looks like |
|---|---|---|
| Opener | Reciprocity, pratfall | A specific observation that shows interest; one light admission of effort or nerves |
| Messages 2–6 | Similarity, self-disclosure | Finding how you both think about something; going one rung deeper and waiting for the match |
| Mid-conversation | Shared arousal | A playful challenge, a bet, a small argument about films |
| Ending each chat | Peak-end, anticipation | Leaving on a laugh with a thread saved for tomorrow |
| Between chats | Anticipation (real) | Living your life; replying when you have attention, saying so when you don't |
| The date | Shared arousal, self-disclosure | Something with movement; conversation that goes a rung deeper than the chat did |
Beyond initial physical attraction, the most replicated findings are: similarity in values and humour, clearly reciprocated interest, gradual mutual self-disclosure, competence paired with a visible small flaw (the pratfall effect), and shared moments of mild excitement or laughter. Conversations are also remembered by their peak and their ending, not their length. All of these reward being specific, honest and a little brave rather than strategic.
Not in the way it's usually meant. Research shows genuine uncertainty about someone's interest can keep them on your mind — but manufactured uncertainty (deliberate delays, withheld warmth) is usually detected and destroys trust. What the evidence supports is having a real life with real limits on your availability. That produces anticipation without deception.
Aronson's 1966 finding that a competent person becomes more likeable after a small, visible blunder. In dating it means one light admission — nerves, a rewritten message, being bad at first dates — makes everything else you say more credible. It only works once or twice and only on top of otherwise coming across well; constant self-criticism has the opposite effect.
The peak-end rule: memory weights the most intense moment and the final moment far more than duration. A chat that ends on a laugh is remembered as great; one that ran until it thinned out is remembered as thin. Ending while it's still good — with a thread saved for tomorrow — also creates anticipation, which is its own attraction principle.
As a natural rhythm between two people who already like each other, no — it's just flirting. As a deliberate tactic applied to someone to make them chase, yes, and it tends to select for anxious partners while repelling secure ones. We teach recognising it and staying steady, not deploying it.
The applied version of principles 2, 4 and 6 is the flirting over text guide; the version for your own head — why confidence is a practice rather than a prerequisite — is dating confidence: 7 mindset shifts. And when the principles have done their work, the next guide is the one that turns a good chat into a date: when and how to ask.