▲ Work in progress. This site is built and researched in the open: pages, numbers and findings may change or be wrong, and the record has gaps. Not advice. More
The Lab
Ideas traders use, put to the record · updated 30 Oct 2026 21:56 UTC
Research on the record, in groups: each piece on a page of its own, saying what was tested, what we found, how sure we are and what could still be wrong. No money is traded; these are statistics, not strategies. The work is ongoing, and an answer here can change.
Grids
#12 · Does the market's mood warn of a bot's losses?Not shownThe paper bots' market words (trending, ranging, calm, wild) against their trades a day later: would a gate that blocks trades under a word have helped? A week of record.#13 · First passage: the chance a price touches a levelA grid rung, a take-profit or a liquidation price: the textbook formula against four years of Bitcoin, Ether and Solana, and a repair that held on unseen years.#20 · Grid bots, followed up: range width, the stop, and four years of ETHFour years of ETH, BTC and SOL: the range's width against the stop, the stop's firing rule, and a correction: a wider range is less risk, not shown to earn.#24 · Grids with and without stopsGrid bots on 14 coins over 30 months of prices, with and without a stop.#25 · Grid bots: more rungs or fewer?The same grid with 3 to 120 rungs, on spot and on futures up to 10x, over four years of Bitcoin, Ether and Solana: what the rung count changes, what fees take, and the order sizes an exchange allows.#27 · Grid bots: does an adaptive grid beat a fixed one?Gaps set by volatility, and a range that follows the price out: does either beat a grid that is set once and left alone?#30 · Grid bots: does a small trend bet beside the grid smooth the bad months?Not shownA small position that follows the trend, kept beside the grid to soften its bad months, against the same position long and short on random days.#61 · Two opposing Bitcoin grids: do they beat holding, or the carry?No reliable gainA long grid and a short grid on Bitcoin in one account, against holding Bitcoin and against the carry, 2020 to 2026.
DCA
#17 · Buying the dip in steps: three tests of our DCA ladderOur DCA ladder, three tests: fewer orders with wider gaps and more patience, selling in two halves, and how much leverage a ladder can take once the chance of ruin is counted.#49 · The QFL signal in a TradingView accumulator: does the signal help?No reliable edgeA TradingView accumulator and its QFL signal, rebuilt and tested against random signals on three windows, and its ladder against ours.
Carry
#23 · Funding carry and the hedged basketA coin bought, its perpetual shorted, the funding collected: how much there is, a hedged basket counted correctly, leaning with the trend, and a year that goes nowhere.#29 · The same long, held two ways: plain coins, or coins plus a hedge that earns fundingPlain coins, or more coins with a short that earns funding: the same bet on the price, what the extra money earns, what it costs, and when the short can be liquidated.#41 · The carry trade in one account: does it survive the crashes?Holds, with the coins paid forThe carry trade held in one cross-margin account or in separate ones, through every crash since 2020 minute by minute, 10 October 2025 in focus: leverage on both sides, a cash cushion, and the collateral rate.#48 · The carry trade on a loan: when should it stand down?Not shownThe carry trade on a loan, cut back the moment funding falls under the loan's rate: against waiting, cutting in steps, weighing the fees, never cutting and never borrowing, on funding since 2022.
Strategies tested
#26 · Pairs trading: does the gap between two coins come back?No reliable edgeWhen two coins that move together drift apart, bet the gap closes: six pairs over four years, after fees and funding, against random timing.#31 · "Shannon's demon": does rebalancing 50/50 harvest volatility?No reliable edgeHalf in a coin, half in cash, rebalanced as it drifts: does it harvest the swings after fees, or does the coin's path decide it?#32 · Can a small trader make markets on Bybit?No reliable edgeA buy just under the price and a sell just over it, replayed on a week of every Bitcoin and Ether trade: is it viable at a retail fee, and what fee would it need?#33 · Does a "2% stop, 4% target" bot beat every strategy?No reliable edgeA claim that a bot with a 2% stop and a 4% target beats every strategy: the odds worked out, four years replayed, and 720 versions tuned, then judged on a year they never saw.#38 · Momentum across coins: buy the strongest, sell the weakestNot proven, a leadThe strongest coins bought and the weakest sold, every week since 2021 on every Binance future, the removed ones too: against random books, after costs and funding.#39 · Momentum across coins, tested again on 2018–2020Not confirmed; the lead holdsThe momentum study's one fixed setting, unchanged, on Binance's spot prices from 2018 to 2020, which none of it had seen: every rule written down before those prices were read.#40 · Momentum among blue chips only: Bitcoin and the top 10No reliable edgeDoes momentum still work on Bitcoin and the ten largest coins, which are less likely to vanish? Each year's list as it was known then, on futures since 2021 and spot from 2018.#46 · Ending up with more Bitcoin, without adding moneyNo reliable edgeSelling a slice when Bitcoin runs far above normal and buying it back lower, or hedging to collect funding: counted in Bitcoin against simply holding, 2017 to 2026, and against sales on random weeks.#47 · Buying steadily, or all at once?No reliable edgeDollar-cost averaging against putting it all in on day one, and against keeping half in cash: Bitcoin, Ether and Solana since 2017, and the top 20 coins of 2018, 2020 and 2022, the dead ones too.#53 · Coin-flip traders: can luck beat the fees?No edge after feesTraders who pick the day and the side by coin flips, 90,000 a setting on Bitcoin, Ether and Solana: does the law of large numbers even it out, and can luck beat the fees?#58 · The demo soak, looked back on: 13 to 21 October 2026Our bots on five demo and testnet accounts for eight and a half days, 13 to 21 October 2026: each account's result split into the bots and the coins that sat idle, the worst falls, and where demo's +40% came from.#62 · How long until a strategy comes out ahead, and can the market say which to run?Holding, the carry, two grids and one grid: how often each was ahead after one month to three years, and whether the month you start in says which to run.
Signals tested
Ideas traders use, each turned into a rule and checked against the record; the numbered ones are judged again as the record grows, on record that did not exist when their settings were fixed.
#1 · Order-book pressureA hintWhen more money waits to buy just below the price than to sell just above it, buyers are stronger, so the price should rise next.#2 · Fade the crowdPartlyWhen unusually many traders are long, few are left to buy: the crowd is late, and the price tends to fall.#3 · Follow the top tradersPartlyBinance's biggest traders know more than the crowd; when they lean long, the price tends to rise.#4 · Chasing buyersReal, but smallA burst of market buying shows urgent demand, so the price keeps rising.#5 · Leverage build-upHolds upWhen open interest climbs fast, traders are piling in with borrowed money; the crowded side gets squeezed and the price falls.#6 · Crowded fundingPartlyWhen funding runs high, longs are paying heavily to stay in: the long side is crowded, and the price tends to fall.#7 · Liquidation cascadesA hintWhen a wave of longs is liquidated, the selling is forced, not a view of the price: once it is over, the price tends to bounce back.#8 · The hour's habitA hintEach hour of the day has its habit: an hour that has tended to rise over the last month tends to rise again, because traders in each part of the world wake, trade and sleep on a schedule.#9 · Into the funding paymentPartlyJust before funding is paid, the side that pays closes out to dodge it: when funding is high and longs pay, their selling pushes the price down in the last hour; when it is negative, shorts buy and push it up.#10 · Funding apartNo signWhen traders on one exchange pay much more than elsewhere to hold longs, that exchange's longs are crowded: they are the ones forced out next, and the price tends to fall.#11 · Memes behind BitcoinReal, but smallMeme coins follow Bitcoin, but late: their traders react after the big money has moved, so Bitcoin's last hour shows where the meme coins go in the next.#14 · A TradingView indicator, rebuilt and replayed on two years of BitcoinNo reliable edgeSpot vs Leverage (SvL v4.5), a TradingView indicator, rebuilt from its source: 450 tests on five charts, against chance and after fees.#15 · The SvL on other charts and coinsNo reliable edgeThe SvL away from its design: Bitcoin from the 5-minute chart to the week, then ETH and SOL. 1,035 tests against chance, the faint signs checked on unseen data.#21 · Liquidation levels after a turn: targets, or just distance?No edge after feesAfter a sharp fall turns, are the liquidation levels of traders caught on the wrong side targets the price is drawn to? Bitcoin, Ether and Solana against fake markets, with a risk table worth keeping.#22 · Fibonacci confluence: does it time trades better than chance?No reliable edgeFibonacci levels from several charts, traded where they agree: against plain fixed distances and against random timing, on Bitcoin, Ether and Solana over a falling year and a rising one.#28 · Does the S&P 500's RSI warn of Bitcoin's falls?No reliable edgeA video's claim that the S&P 500's monthly RSI lines up with Bitcoin's tops and bottoms: checked, and asked whether its turn is known in time to warn.#34 · Do coins have a clock? Hours, weekdays, funding times, the US openNo reliable edgeHours of the day, weekdays, funding times and the US open, on ten large coins since 2021: does any habit stand out among 35 tries, and pay the fee on a year it hadn't seen?#35 · Forced selling and crowded trades: open interest, funding, long/short ratiosNo reliable edgeOpen interest, funding and long/short ratios on 54 coins: does the price move in a usual way after a flush or a crowd leaning one way? And a first test that was wrong, said plainly.#36 · Buying a sharp drop: does the market bounce the next day?Failed on fresh dataThe one lead left by the forced-selling study, tested again on 2020 and 2021, prices none of our studies had touched, with every rule written down first.#45 · Elliott waves: do the zones and the wave trades hold up?No reliable edgeThe wave rules, the Fibonacci zones and the wave trades, explained, then tested with the waves counted as they happen, on Bitcoin, Ether and Solana since 2021: against random zones and random entries.#50 · The touch map in calm and wild markets: does a table for each help?No reliable gainThe touch map's odds from one table, or a table for each mood: calm, normal or wild by the volatility rank, or the hidden Markov model's two states, judged on prices none of them saw.#51 · The volume clock: counting time in trading instead of hoursNo reliable gainCounting time in coins traded instead of hours: are the returns closer to a bell curve, and are the touch map's odds better on that clock?#59 · Purple whales: do big orders lead Bitcoin?No reliable edgeBitcoin's order-size lines rebuilt from public trades, and the "purple whales" ($100k to $1M) tested against the price: rebuilt faithfully, no lead found.#63 · Trend over months: does a coin's own past point its way?No reliable edgeTwo published rules, a coin's own last year and Bitcoin's own last week, tested on crypto after the papers' data.#64 · Session momentum: do the first hours predict the last?No reliable edgeDo the first hours of the Asian, London or New York session point the way of its last hours, as a paper found in US stocks?#67 · Real liquidation clusters: magnets, turns, or neither?Not shownPrices where liquidations really clustered, on Bitcoin, Ether and Solana: touched again more than chance, or a turn worth fading? A first look; judged looks in January and April 2027.
Tokenised assets
#54 · Stock perps while the stock market is shutPassed, on a weak barPerpetual futures on shares trade all week while the stock market is shut: does the weekend price tell how Monday opens, what does holding one cost, and how rough is the reopen?#55 · Gold on crypto exchanges: how close, and does a grid suit it?Not shownGold tokens and gold perps against real gold: how closely they follow it, what holding one costs, and whether a grid suits gold.#56 · Tokenised gold and stocks: are they worth it?Holds upWhat a token or a perp costs against the real thing, how gold and the big stock funds compare with Bitcoin and Ether, and whether moving some of a crypto book into them helps.#57 · Tokenised gold and stocks on chain: where they tradeWhere tokenised gold and stocks traded, on chain or on the exchanges, how deep each was near the price, and which one set the price.
Models explained
#16 · Models for reading a market, and what they showed on real pricesMarkov and hidden Markov models, mean reversion, order flow and where market data comes from: each explained, then tested against fake markets.#19 · How far to trust a Sharpe ratioThe Sharpe ratio, the usual score for a strategy: how many days it needs before it means anything, how much of the best of many settings is luck, and the ruin it cannot see.#42 · Survivorship: how the coins were picked, and what it doesWhy a list of coins picked today flatters the past, measured on every Binance future since 2021, the dead ones too, and what it means for each page in the Lab.#52 · Hidden order in crypto prices? The BDS test, before and after a volatility modelAre crypto returns independent from one period to the next, and does a standard volatility model explain what isn't?#60 · How our backtester works, step by stepOur bot's own code replayed on past prices, step by step, for readers who have never seen a backtester: what it does with each candle, where it leans cautious, and what it cannot tell you.#65 · Skill or luck: how many coin-flippers look skilled?90,000 coin-flip traders put through the usual test of skill: how many look skilled, and the better test that catches them.#66 · The best of many tries: our own results through White's reality checkOur own best results put through White's reality check: how much of a best result was only the best of many tries.
Notes
#18 · Bitcoin after the CPI flush: where the liquidations sitWhere traders on leverage get liquidated after the US inflation surprise, and how likely Bitcoin was to reach each level, as of 14:33 UTC on 14 Oct 2026. Dated and frozen: never refreshed.#43 · Bitcoin after the CPI flush: how the levels did, a day laterThe CPI note's liquidation levels scored one day on: which were touched, against the odds it gave, and why one day can't show whether those odds were right. Dated and frozen.#44 · Bitcoin after the PPI flush: where the liquidations sitThe day after CPI, producer prices came in hot too: where traders on leverage get liquidated, which levels the drop and the bounce took out, and the odds for the rest, as of 14:45 UTC on 15 Oct 2026. Dated and frozen.#68 · Bitcoin around the FOMC decision: where the liquidations sitWhere traders on leverage get liquidated around the US rate decision, and how likely Bitcoin was to reach each level, as of 20:00 UTC on 28 Oct 2026. Dated and frozen: never refreshed.#69 · Bitcoin around the FOMC decision: how the levels did, a day laterThe FOMC note's liquidation levels scored one day on: which were touched, against the odds it gave, and why one day can't show whether those odds were right. Dated and frozen.
Theory
The ideas behind the tests, apart from them: where each comes from, its maths step by step, what others found, and what our own tests found.
#T1 · Shannon's demon: making money from a price that goes nowhereRebalancing a fixed split to profit from a price that swings but goes nowhere: where it comes from, the maths, and why a trend starves it.#T2 · The Kelly formula: how much to bet when you have an edgeHow much to bet when you have an edge, and why betting more than that loses even when you are right.#T3 · Steady buying or all at once: the theory of dollar-cost averagingSteady buying or all at once: what the theory says dollar-cost averaging can and can't do.#T4 · Momentum: do recent winners keep winning?Do recent winners keep winning? The idea, the research on stocks and crypto, and the catch.#T5 · Mean reversion: does what drifts away come back?Does what drifts away come back? Where the idea comes from, how to measure the pull, and when it breaks.#T6 · Hard-to-beat markets, and why every test needs a random twinWhy markets are hard to beat, and why every test here is set against a random twin.#T7 · Survivorship bias: the coins you can't seeThe coins you can't see: how a list picked today flatters the past, and how to test fairly.#T8 · Fibonacci and Elliott waves: where the numbers come fromWhere the Fibonacci numbers and Elliott's waves come from, and what tests of them find.#T9 · Funding carry: getting paid to hold a hedged coinGetting paid to hold a hedged coin: why futures pay funding, what the trade earns, and what can break it.#T10 · Quantum finance: what it adds, and what it doesn'tTwo ideas under one name: quantum maths for option prices, and quantum probability for people's choices. What each adds, and what our tests found.
The Lab notebook: what changed, and when →
Historical data: Binance Vision (data.binance.vision), used under CC BY-NC-SA 4.0 (creativecommons.org/licenses/by-nc-sa/4.0); what is drawn from it here is shared under the same licence.
Educational research from public market data; any trading discussed is paper, never real. Nothing here is advice. The full notice.