Trading Journal vs P&L: ~3% of Polymarket's 1.72M Accounts Move Prices — the Gap Between Plan and Execution as the Only Edge Author: Дністер Published: 2026-09-06T03:01:30.000Z Language: en URL: https://neurodrift.org/en/blog/tradzhurnal-yak-psykhiatr/ Original (Ukrainian): https://neurodrift.org/blog/tradzhurnal-yak-psykhiatr/ Tags: markets, trading, psychology, discipline, journal Original source: https://neurodrift.org/blog/tradzhurnal-yak-psykhiatr/ An LBS and Yale study of 1.72M Polymarket accounts found that about 3% move the price. The other 97% are the liquidity those 3% live on. This is a text about why most traders are in the bottom 97% — and the one tool that moves you from there to the top. Not a model. Not speed. A journal that behaves like a psychiatrist you never hired, whose bill is paid in losing trades. ----- "3% of accounts move the price. The other 97% lose to them." — from a working paper by London Business School and Yale on 1.72 million Polymarket accounts. I. At 2 a.m. a psychiatrist you never hired quietly sits at your desk Two a.m. The third monitor glows red. In the cup — cold coffee with a ring from the previous session. In your head — "one more trade and I'm even." Right at this moment a psychiatrist you never hired quietly sits at your desk and starts keeping your chart — with every next stop. Imagine an invoice arriving for a therapy session no one ordered. Under "service" — "deep diagnosis of recurring behavioral trauma." Under "payment" — not euros. Losing trades. That's the trading account: the only psychiatrist in town who takes you with no appointment, works around the clock, and bills in kind — in the trader's money, one stop-loss per session. First a cold, foreign, peer-reviewed number — so no thought arises that this is a motivational text about "the discipline of winners." Four economists from London Business School and Yale — Gómez-Cram, Guo, Jensen, and Kung — took every Polymarket trade for 2023–2025: 1.72 million accounts, nearly $13.8 billion in volume. And tested the industry's sacred mantra — that the market is wise because it gathers the collective wisdom of the crowd. The result: about 3% of accounts move the price toward the correct outcome. The other 97% lose in aggregate to that minority — their combined deficit equals the 3%'s surplus. Not "the crowd is sometimes wrong." The crowd systematically finances the minority that measured itself. Now without euphemisms. "The collective wisdom of the crowd" is a marketing euphemism for "97 people out of 100 chip in for the salary of three." The retail trader isn't a "market participant." He's a revenue line for those who counted themselves earlier. A casino at least is honest: it doesn't tell you you're taking part in a democratic discovery of a fair price. There's no poster at the casino entrance saying "Your losses are our lessons." In trading there is — it's just in the trader's head. The market did one better: it takes and hands you a diploma in amateur psychology — a paragraph of self-knowledge for every losing trade. The only industry where tuition costs a minus on the account, and the teacher doesn't even show up to class. This structure isn't new — it existed among 1990s Taiwanese day traders, and in the 2000s forex rooms. But in 2020–2026 it became infrastructure. Commission-free mobile brokers (Robinhood, eToro, Bybit) lowered the entry barrier to zero: opening an account is now easier than ordering pizza, and gamification with confetti and social feeds of "what others are trading" turned speculation into everyday leisure. Polymarket and Kalshi together went from zero to $24 billion monthly volume in two years. The difference isn't an individual trader's behavioral weakness — it's a system now actively growing liquidity from the bottom, because its business model rests on that liquidity. This text's question is simple and unpleasant. Not "how to become one of the 3%." But: how do you even know which percentage you're in? The P&L won't say — it's too noisy and too flattering. Only one thing will say it: a journal that behaves not like accounting but like a psychiatrist. The same one you never hired, whose bill is already growing. And until the trader reads its notes — the bill keeps accruing. !An annotated breakdown of the scene with numbered pointers and an explanatory legend. The same desk, broken into a session. ① the empty psychiatrist's chair — the trade journal is that therapist; ② three monitors stroke the ego with green days, but only the journal makes the diagnosis; ③ the open journal — the only honest record: what you DID, not what you felt. II. The P&L lies to you in compliments — the journal makes the diagnosis The first thing to accept without anesthesia: your P&L is a flatterer. It's not a work report. It's a horoscope. In a green month it whispers "you're a genius, you have an edge," in a red one — "you just got unlucky, the market's irrational right now." And both are the same sentence said in a different tone. The P&L shows only what the trader wants to think about himself — and just enough not to close the account. Because the broker is also grateful for the company. The market rewards those who measured their own gap between plan and execution — and punishes those who replace measurement with the feeling of an edge. Until that gap is recorded, the trader is voluntary liquidity for those who can count their own failures. It's literally a place in the food chain. Imagine a doctor who diagnoses purely by how much money the patient has in his wallet on the way out. Got richer in a week — "healthy." Got poorer — "dying." No psychiatrist works this way, because it's absurd: the wallet is a consequence, not a symptom. And the typical trader reads himself exactly like this. The P&L is the wallet on the way out. The journal is the case history: "blew two stops, entered three times bigger because 'I have to get even' — another −3R." No pretty excuses, only diagnoses. Here's where the real edge hides — and it's not where they look for it. The edge isn't a model or speed but the gap between you-who-made-the-plan and you-who-pressed-the-button. Until you measure that gap, you're not a trader with an edge — you're liquidity for those who measured theirs. The P&L says how much you lost. The journal says who you were the moment you were losing. The first is a horoscope. The second is a case history. Before moving on — the honest counter-position this argument deserves. Not every trade journal is a diagnostic tool. Most are theater of self-discipline. The market is full of people who keep perfectly formatted spreadsheets with color triggers and pretty R-counts — and still stay in the 97%. That's not a lack of journal. It's "survivorship bias theater": the journal becomes a way to feel disciplined, not to measure the real pattern. We're told about traders who kept a journal and went profitable. About those who kept a journal and still blew the account — silence. So the argument "keep a journal → you'll reach the 3%" is weak, and that must be admitted plainly. But here's what this critique doesn't explain: without a journal, the behavioral pattern never becomes visible at scale. One revenge trade is an exception. Fourteen revenge trades a month is a disease with a price list. The difference between "journal as theater" and "journal as psychiatrist" is one: whether the trader counts a Deviation column with an R-price, or just describes his state. The first is a mirror. The second is a self-report for Instagram. The critique of journals in general is fair. The critique of the measured behavioral gap is not. One more qualification is missing: the outcome of a single trade is a decision plus market noise, and over a short distance there's so much noise the signal isn't heard at all. A green day proves nothing. A red week proves nothing. Only the pattern in the journal, measured over a month, tells the truth. The rest is a PR department in your head, rewriting the script in hindsight so the trader was always right. III. Meet the Behavioral Edge Ledger — three columns that expose you Here's a tool, not a sermon. The ledger is a black box. A plane crashed — everyone looks not at the pilot's vacation photos but at the flight record. The P&L is a selfie against the wreckage. The ledger is the FDR, the flight data recorder. I'll name it so it sticks: the Behavioral Edge Ledger. It's not the table where you write "bought / sold / profit." It's a map of three columns, and each one cuts: ColumnQuestionWhy this one TriggerWhat happened the moment before your hand reached for the button?Not the chart. The trader's state: lost the previous one, bored, everyone in the chat is already in the position, just closed in the green and felt invincible DeviationHow did the action differ from your own plan?Not "from ideal trading." From what the trader wrote down to himself before the market opened. Moved the stop. Entered with no setup. Increased size on emotion Cost in RHow much did it cost in units of risk, not dollars?R is the planned risk per trade. The dollar pities and masks: a $50 loss out of boredom seems trivial until you see it's a full day's risk. R is the lens that won't let you pity yourself The second column is the main one: here lives the difference between "I planned it this way" and "I did it this way." The rest is support staff. If you count only dollars, not R, you're not managing risk, just watching how expensive the illusion of control is. Through this lens the trader looks not like a victim of the market but like a person who, for the third time this week, confuses boredom with an investment strategy. And the formula it was all for. The real edge isn't the quality of decisions or the quality of execution separately: Edge = (Decision − Execution) × Repeatability Translation without euphemisms: you can have a brilliant strategy (decision = 10) and blow the account because execution = 3, and that same gap reproduces forty times a month with grim regularity. The hole isn't in the model. The hole is in the gap — multiplied by repeatability, it eats the edge faster than any bad setup. Until you can name these three numbers, the edge is absent. There's only the feeling of an edge, and feelings aren't traded. !Three columns. The middle one is the densest, because that's where the confession is. An analyst duck reads it like an FDR after a crash: no pity, no excuses, only the pattern. Three columns. The middle one is the densest, because that's where the confession is. An analyst duck reads it like an FDR after a crash: no pity, no excuses, only the pattern. IV. A catalog of the walls you bang into — by name and with a price list Behavioral finance long ago gave precise names to your favorite ways of harming yourself — and even peer-reviewed proof that these errors aren't unique. Because the journal's most galling discovery: the "unrepeatable mistake" has an academic name and sits in a textbook. No trader is a trailblazer. Each is an illustration to a chapter. Here's the catalog of walls, each with a face and a price list in R: WallTriggerWhat they tell themselvesWhat it really is Revenge trade10:03. Just caught a stop, the hand reaches to double down in the same direction"I'll get even right now"Thaler–Johnson's "break-even" effect: after a loss, any chance to recover seems indecently attractive. A revenge trade is an emotional credit card: instant relief now, the bill with interest at month's end in R House-money effectJust won"This isn't my money, it's profit, I can risk it"The same Thaler–Johnson, mirrored: a win is mentally separated from "real" money and risked the way you'd never risk a paycheck. Not for nothing did the term come from the casino — they saw it every evening Moved stopPrice approaches the stop"I'll give it a bit more room, the market will shake out the small fry"Disposition effect: in the loss zone the psyche turns risk-loving and refuses to lock the loss. The trader isn't "giving room." He's turning −1R into −4R, haggling with reality Boredom entryNothing is happening"Something must be there, I'm a trader after all"Confusing activity with work. The market isn't obligated to entertain — and a fee for entertainment is charged every time. Month's total: −4R out of boredom, paid like therapy FOMOEveryone in the chat is already in the position"The train's leaving without me"Entering on the very impulse the 3% sell the top into. The trader calls it "intuition." They call it "exit liquidity" A recurring trigger isn't a "minor slip." It's metastasis. The patient is interested not in the first tumor but in how many times it already fired into the account's other organs. Notice the pattern? Not a single cell in the "trigger" column has a chart. Everywhere — the trader's state. The market just presses a button that's already wired in. And here's the black joke of this section: years are spent searching for a better indicator, a better timeframe, a better model — while the wall you actually bang into is the same the whole time, and it's not on the chart. It's in the mirror. In Uncut Gems the protagonist parlays several bets, and every time he "gets even" he only raises the stakes further. He sees neither revenge trade nor house-money — because he doesn't look. The journal is the only mirror that doesn't flatter: it just shows that the "moved stop" line appears for the fourteenth time this month, and all fourteen times it was called "patience." You're searching for a better indicator. The wall you bang into has been the same the whole time — and it's not on the chart. It's in the mirror. And it doesn't need a better indicator — it needs one record. V. Market noise is the best ally of your inner con artist Now understand why, without a journal, this is incurable — structurally, not for lack of willpower. Because in trading the feedback is poisoned: a right decision regularly gives a loss, a dumb one regularly gives a profit. You enter perfectly by plan — and catch a stop from a random price spike. You enter on pure revenge, against all your rules — and accidentally hit the jackpot. The market just taught you: "ignore your plan, do what you want — that gets paid." And it really paid. Market noise is friendly fire: your own side hits you when you did everything right, and gives you a medal when you aim at your own foot. Imagine a trainer who gives the dog a treat for disobedience and hits it for an executed command — randomly, every other time. A month later — a neurotic animal that understands no connection between action and consequence. The only difference is that the dog at least isn't given leverage. The market is exactly this trainer. Without a journal, learning happens not in reality but in a casino pretending to be a teacher and grading at random. Day trading pretends to be a school: you pay like for an MBA, learn from Twitter memes, and one in a hundred gets the diploma — and only because he came to learn, not to play. The journal breaks this circle the only way possible — it separates the quality of the decision from the result. A trade is judged not by how much it brought but by whether you followed your own plan. "Lost, but executed everything flawlessly" is a winning record, even if the P&L is red. "Won by breaking all the rules" is a losing record, even if the account is green: a premium was just paid for behavior that will kill the account over the distance. The P&L applauds this luck. The journal is the only one in the room that doesn't clap, but quietly records: "got lucky to survive, protocol violated." It's exactly these "winning mistakes" that are most dangerous — the casino-teacher just handed out a scholarship for what will eventually ruin the account. !The market trains you like a neurotic dog: a treat for disobedience, a blow for discipline — every other time, at random. The one difference from the dog: it isn't given leverage. The duck records: "feedback poisoned." The market trains you like a neurotic dog: a treat for disobedience, a blow for discipline — every other time, at random. The one difference from the dog: it isn't given leverage. The duck records: "feedback poisoned." VI. Why most traders are in the bottom 97% Back to those 1.72 million accounts — because the number is now painfully concrete. The study's conclusion, no ceremony: about 3% of experienced traders move the price, and everyone else finances their profits. This isn't a metaphor or hyperbole — it's the market's structure, measured trade by trade over two years. And before the thought "well, I'm in the 3%" surfaces, remember: that's exactly what all 100% think. It's statistically impossible, and that's precisely why it's funny. If the thought "I'm not a sucker" arises — that's the ideal client for any broker, and the profile already has a "voluntary liquidity" tag. As the old poker rule goes: if in twenty minutes you haven't figured out who the sucker at the table is, you're the sucker. Or, as in Rounders: Mike understands it not when he loses, but when he sees he's being read like an open book from the start. The study simply counted how many times 1.72 million people sat at the table, and each sincerely scanned the room for someone else. And this isn't a narrow quirk of crypto prediction markets. Barber and Odean of Berkeley studied real Taiwanese day traders for years — the full sample over five years. The verdict: in a typical half-year more than 8 in 10 day traders are in the red; after fees only about 1% are predictably profitable; most quit within the first few years. Their conclusion, verbatim: "trading to learn is no more rational or profitable for the individual investor than playing roulette to learn." Day trading is roulette pretending to be a business school. Roulette at least doesn't pretend there's a skill issue in it — it spins honestly. What they think about themselvesWhat the measured structure says "I have an edge, just a losing streak"~3% move the price; the other 97% are in the red against them (LBS+Yale, 1.72M accounts) "I learn from mistakes, I'll go profitable over time"Experienced speculators are also in the red; "learning by trading" ≈ learning by roulette (Barber/Odean) "Most blow up, but I'm not most"8 in 10 in the red every half-year; ~1% profitable after fees; most leave after a few years "The sucker is definitely not me" + "I have a special mindset"That's what 100% at the table think. Add "special mindset" — and that's the ideal client for any broker The hard moment, no cushion: probabilistically, most traders are in the bottom 97%. Not because they lack mental ability — among that 97% there are plenty of smart people with degrees and three monitors. But because belonging to the 3% isn't a question of intellect or information. It's a question of whether you've measured your own gap between plan and finger. And here we must name those for whom this structure is profitable forever — because it isn't going anywhere. Brokers and platforms earn on every trade, regardless of outcome: commission, spread, funding rate on leverage. Robinhood paid Citadel for order flow — that's not conspiracy, it's public SEC record. eToro takes a spread on every trade, even if the trader is in profit. Polymarket and Kalshi live on volume fees — they profit from maximum flow, not from a minimum number of losses. So gamification, push notifications "your asset is moving," and social feeds of "what others are trading" aren't UX decisions. They're product design for growing liquidity from the bottom. The losing 97% aren't a side effect of the system. They're its fuel. If you're really in the 3% — this text annoys you. If you enjoy it like a motivational pep talk — you're in the 97%. But at least one side will surely read the last sentence aloud and say "well, that's not about me." There's your statistics, live. !At the table everyone scans the room for the sucker. If in twenty minutes you haven't found him — you're the sucker. The dealer is a duck: the only one who counted and knows the answer in advance. At the table everyone scans the room for the sucker. If in twenty minutes you haven't found him — you're the sucker. The dealer is a duck: the only one who counted and knows the answer in advance. VII. How to keep a Behavioral Edge Ledger without turning life into a tribunal Here I must restrain myself, or I'll hand you a new inner prosecutor instead of an edge. Because the journal has two deaths. The first — it isn't kept at all (your case right now, let's be honest). The second, sneakier — it's turned into daily self-flagellation: a tribunal after every trade, hatred for every "−1R," and within a month you drop this torture chamber along with trading. If your journal sounds like an inquisition transcript, not a medical chart — that's not treatment, it's burning. The journal isn't a court, it's a psychiatrist's appointment. A psychiatrist doesn't shout "you're at it again!" — he coldly records the pattern and shows the frequency. Here's the protocol that holds between the two deaths: Write BEFORE, not just AFTER. The psychiatric journal is filled on entry into the position, while the blood is cold: setup, entry trigger, planned stop, size in R. Because after the trade, the inner PR department has already rewritten the story so the trader was always right. The pre-trade record is the only testimony that can't be bribed in hindsight. The Deviation column is the main one. Never skip it. "Executed per plan" or "deviated"? If deviated — in one word: revenge, boredom, FOMO, fear. It's a diagnostic code. Without it the journal is just accounting. Count in R, never in dollars at the table. "Boredom cost −2R today" hits more soberly than "−$47." The dollar pities. R doesn't. Once a week — group by trigger, not by date. This is the moment of diagnosis. You'll see not "I had a bad Tuesday" but "revenge trade cost −11R this month." The wall becomes visible the moment you stop looking at trades one by one. One trigger at a time — one experiment. Don't fix everything. Take the most expensive wall and set one barrier (e.g., after two losses in a row — a one-hour stop). Next week measure whether that trigger's price dropped in R. Trying to "fix everything" at once isn't treatment, it's renovating an apartment that's on fire. Pick one room. The moment the ledger is kept to look disciplined in your own eyes — the number becomes theater, not diagnostics. The metric devours the meaning exactly when it stops being a mirror and becomes a costume. The difference is subtle but saving. The tribunal asks "what a worthless person am I?" The journal asks "which pattern cost the most R and which one barrier reduces it?" The only KPI here isn't how much you earned. It's how many times you honestly recorded "entered on revenge" — before the market slapped on another stop. The ledger doesn't make you a genius. It just takes away your right to honestly call yourself a sucker. A pre-trade record can't be bribed in hindsight. The truth about the gap costs less than another month in the bottom 97%. VIII. Insiders, $630 thousand, and the last honesty about the edge Finally — the harshest calibration, so you don't leave with the fantasy that a journal makes you invincible. The same LBS and Yale study showed another side of those 3%: in the days before Maduro's removal in Venezuela, three freshly created Polymarket accounts poured large bets into the "will Maduro be removed" contract when the market priced it at ~10%. After the raid they jointly made over $600 thousand. Part of the "informed minority" is informed not because it keeps a better journal. It's informed because it knew in advance. This game has insiders. You're not one of them. But you can at least stop being their cashback. The journal doesn't guarantee a place among the 3% — because part of them plays a game we structurally don't play, and no Deviation column overcomes that. But the journal guarantees something else, and it's the only thing in the trader's hands: it pulls you out of the category of voluntary liquidity — those who lose not for lack of information but from an unmeasured gap between their own plan and their own finger. We already counted it: Edge = (decision − execution) × repeatability. For most this number has a minus sign. Until both numbers are measured, the edge is absent. There's the feeling of an edge. And feelings, as established above, have no status. Only the recorded has status. This text changes nothing if tomorrow you again open the platform without an empty "Trigger / Deviation / Cost in R" table beside it. The psychiatrist you never hired already sits at your desk and keeps your case history — with every losing trade. The only question is whether you'll finally start reading his notes, or keep paying the fee without entering the office. The psychiatrist isn't offended. He gets paid whether or not you read his notes. You can't become an insider. But you can stop being the fuel that lights its own match. The most expensive wall isn't the one you bang into. It's the one whose frequency of impact you refuse to count. Sources & context "~3% move the price, 97% finance them" — Gómez-Cram, R., Guo, Y., Jensen, T. I. & Kung, H. (April 2026). London Business School & Yale working paper on every Polymarket trade 2023–2025: 1.72M accounts, ≈$13.8bn volume; ≈3% of accounts provide most price discovery, the other 97% on average lose to that minority — Coindesk; The Block; Cointelegraph. Maduro insider case / $600K+ — same study: three freshly created accounts entered large bets in the Maduro-removal contract (market ~10%) before the operation; after the raid — jointly over $600 thousand — Coindesk; Crowdfund Insider. Prediction-market volume (April 2026) — Kalshi $14.81bn + Polymarket $9.01bn; the sector nearing $150bn cumulative volume — AInvest; The Block (data); Pew Research. "8 in 10 day traders in the red; ~1% profitable" — Barber, B. M., Lee, Y.-T., Liu, Y.-J. & Odean, T. "Do Day Traders Rationally Learn About Their Ability?" Full sample of Taiwanese day traders 1992–2006: in a typical half-year >80% in the red; after fees only ≈1% predictably profitable; "trading to learn ≈ playing roulette to learn" — Barber & Odean (Berkeley, PDF). House-money effect + "break-even" (revenge) — Thaler, R. H. & Johnson, E. J. (1990). "Gambling with the House Money and Trying to Break Even: The Effects of Prior Outcomes on Risky Choice," Management Science 36(6): 643–660 — Management Science. Disposition effect (moved stop / holding losers) — the tendency to lock gains early and hold losses long; prospect theory, the S-shaped value function — overview; ScienceDirect. R-multiple / risk-unit as a unit of account — the concept of judging trades in multiples of risk (R), not absolute money; Van K. Tharp, "Trade Your Way to Financial Freedom." Used as a normalization tool — overview. "Behavioral Edge Ledger" and the formula Edge = (Decision − Execution) × Repeatability — this text's authorial frame (NeuroDrift), not an external term; a synthesis of the behavioral data above into a practical three-column journal. Disclaimer: the text is about behavioral self-analysis, not investment advice. Specific numbers and cases are as of April–May 2026 sources; market volumes are volatile. The "sucker at the table" is a retelling of the folk poker rule, not a quote from a specific person.