Trading idea check for BTC and ETH
Tell us your idea in plain words. We turn it into an exact rule and lock it before we count anything. Then we check it on past BTC and ETH data against what happens normally, without your idea, and show you every time it failed.
No password. The pre-check opens from a link in that email, or in our Telegram bot. We use your email only for this pre-check and, if you order, for updates on that order.
The verdict is paid, from $29. It answers two things: did your idea come true more often than chance in the past, and would it have made money after fees.
We tell you by email or Telegram as soon as your card is ready. Nothing you write here goes public unless you say so.
Questions before you start? Ask us in Telegram or write to us here.
Draft: nothing has been tested yet. This is only the rule as we read your idea.
You pay when you order, whatever the verdict. Full refund if you change your mind before the verdict card is delivered.
We keep this draft for 30 days. The link to come back is in your email or Telegram. After you order, we send every update and the finished card by email, and by Telegram if you connect it.
They don’t have our archive. You can write a backtest there in minutes, but the data isn’t there: no hourly snapshots of the Deribit option chain, no Hyperliquid whale positions, no Bybit retail long share. We keep them in one place, together with DVOL, funding, ETF flows and daily candles since 2017.
And a rule there is easy to bend. In half an hour a backtest gets tweaked until it looks good. Here the rule is locked before the count, every version is counted, the result is compared with chance, and the misses go into a public ledger.
A real card from our own test, not a client’s. The index is the classic Crypto Fear & Greed Index by alternative.me; BTC and ETH prices from Binance.
“Buy when the market is in extreme fear: the next month beats an ordinary one.”
Hypothetical count on past data, not a forecast.
Bought at the daily close on the first day of a fear streak, sold 30 days later, after fees.
The fingerprint is a short code made from the exact rule: change one word and the code changes. You get it, with the time, before any result exists, so the rule can’t be quietly changed later. If we read your idea wrong, version 2 is free, and both versions stay on the card.
Every idea we test goes into a public ledger, and yours will too if you choose to share it. Most ideas marked “hold up” were checked only on the years where the idea was found. We’re now checking them on new data and will record those results the same way, even if they’re worse.
Ledger snapshot 2026-11-06 08:45 UTC.
A sharp jump in BTC implied volatility within a few hours is followed by a bigger move over the next 24 hours.
2024 to May 2026: 66 jumps. Over the next 24 hours price moved 3.4% on average, against 1.6% on a normal day. May to October 2026: 13 jumps, 1.4% against 1.3%, no real difference. It worked in one period, not the next, and it never said which way price would go.
Options, DVOL
Price gets pulled toward max pain before options settle.
482 BTC and ETH expiries since March 2026. Price at expiry ended near max pain only 37% of the time, and yesterday’s price was the better guess. The famous magnet wasn’t there.
Price vs options, 3 November 2026
When perps get extremely expensive or cheap against spot, price snaps back over the next week.
BTC: in the most extreme 5% of readings, the next 7 days beat an average week by +0.7% in the snap-back direction, too big to be luck. ETH pointed the same way, but so weakly it could be luck, so the idea only stands for BTC.
Futures and perps
Seven are chart tools: the 200-day average, the golden and death cross, RSI below 30 and above 70, MACD, Fibonacci, round numbers. Three are market lore: funding, a crowd of retail longs, and “Uptober”. We wrote each one as an exact rule and time-stamped it before counting anything. To pass, an idea had to show up on both BTC and ETH, and random days had to rarely give the same result. Six weren’t confirmed by the data; four had too few cases to judge.
“Price below the 200-day moving average means a bear market.”
Over the next 90 days BTC gained 2.5% on average after days below the line and 10.9% after days above, so the direction matches. But random days give a gap that big 14% of the time on BTC and 10.3% on ETH, too often to trust. And since 2021 the gap points the other way.
“A golden cross, the 50-day average crossing above the 200-day, starts a rally, and a death cross starts a fall.”
Seven to nine crosses of each kind per coin, at least 90 days apart. That is too few to tell from chance, and after death crosses price actually rose more than in an ordinary 90 days, on both coins.
“RSI below 30 means the market is oversold, and a bounce follows.”
BTC rose over the next 14 days in 42% of cases, while an ordinary 14 days rose in 54%. ETH fell 6.2% on average after such days, so neither coin shows the bounce.
“RSI above 70 means the market is overbought, and a pullback follows.”
Over the next 14 days BTC rose 5.2% on average against 1.2% normally, ETH 3.2% against 0.9%. So there was no pullback. We hadn’t planned to test the opposite direction, so we only note it.
“A weekly MACD cross upward marks the start of a rally.”
There were 14 crosses on BTC and 13 on ETH, at least 12 weeks apart. Too few to tell from chance; after them ETH rose less than in an ordinary 12 weeks.
“A pullback tends to stop at the 61.8% Fibonacci level.”
Five days after touching 61.8% price was back on the trend side in 53% of cases on BTC and 65% on ETH. Neighbouring levels between 55% and 70% did about the same, 49% and 64%.
“Round numbers such as $100,000 act as support and resistance.”
Three days after a touch price was still on its side of a round level in 54.4% of cases on BTC and 51.5% on ETH. For shifted, non-round levels it was 54.4% and 55.4%, so round numbers held no better.
“Very high funding means longs are overheated, and a drop follows.”
Over the next 7 days BTC rose 1.9% on average against 0.6% in an ordinary week, ETH 2.7% against 0.5%. So there was no drop. We hadn’t planned to test the opposite direction, so we only note it.
“When most retail traders are long, price moves against the crowd.”
Our Bybit archive starts in March 2025; the first months only set the usual level, so the count starts in August 2025. That leaves 10 such days on BTC and 5 on ETH, a week apart. Price fell a little more than usual after them, but that is too few cases to tell from chance.
“Uptober: October is usually the best month for BTC.”
BTC rose in 7 of 10 Octobers, 14% on average against 1.5% in other months, but the last two Octobers fell and ETH averaged only 2.5%. Eight full years in the test are too few to tell a month from chance.
Not client results: our own check. Rules time-stamped on 5 November 2026, before the count. Data: Binance daily candles for BTCUSDT and ETHUSDT from August 2017, Binance funding from September 2019, Bybit retail long share from March 2025. This describes the past and is not a forecast.
We lock the rule before anyone sees a result, so it can’t be bent to look good. A changed rule is a new version, and both verdicts stay.
One message in plain words. No code, no jargon needed.
SecondsOpen the pre-check from your email or Telegram. Check the exact conditions, the numbers and five past dates when it would have triggered. If it’s right, you say “yes, that’s my idea”.
Free, as many edits as you likeAfter you pay, we lock the rule and time-stamp it. You get its fingerprint right away: a code that shows the rule wasn’t changed later.
Seconds after paymentWe count how the rule did on past data and compare it with what happens normally without your idea, for example a level just as far away on the other side of the price. Fees and spread come off. The card has the verdict, five checks and every losing case with its date.
Ready within 48 hours, within 24 on Pro. We message you by email or Telegram as soon as the card is readyTick what your idea needs and see how far back the test can go. Add one options-chain condition and the whole test can only start in March 2026. You see this before you pay.
test starts
— months of data
For every Deribit expiry we publish a price range. Over 482 expiries, price ended inside it 82% of the time. A range of the same width, simply centred on yesterday’s price, did better: 90%. So where we placed our range added nothing, and the docket page says so. Your idea gets the same treatment.
Source: INDICIA docket, 482 expiries on 3 November 2026.
Every card runs the same five checks and ends with a grade from A to F. You see which part of your idea works and which part is dead weight.
“Undetermined” is also an answer: on this data the idea can’t be told apart from luck. The pre-check tells you beforehand whether there are enough cases, so this verdict is paid like any other.
If the pre-check accepted the data and it then fails on our side, you get your money back.
Packs bring the price of one check down. Subscriptions add alerts on your rules and a monthly re-test on new data.
Our own archive, collected from public exchange APIs: hourly snapshots of the whole Deribit BTC and ETH option chain since 10 March 2026, hourly DVOL and funding since May 2024, daily BTC and ETH candles since August 2017, spot ETF flows since January 2024, and the positions of every account in our Hyperliquid whale registry every 30 minutes since July 2026.
Before you pay, the pre-check shows which of these sources your rule uses and from which date we have the data.
No. Describe the idea the way you would explain it to a friend. We write the rule and you confirm it before anything runs.
You can, and it takes minutes. Two things are missing there. The data: our archive keeps the hourly Deribit option chain since March 2026, DVOL and funding since May 2024, spot ETF flows since January 2024, Bybit retail long share since March 2025, Hyperliquid whale positions every 30 minutes since July 2026 and daily BTC and ETH candles since August 2017, all in one place. And someone to hold the rule: here it is locked before the count, every version is counted, and the result is compared with chance.
Before we count anything, we save the exact conditions with a time stamp and a digital fingerprint: a code that shows the rule wasn’t changed later. You get the fingerprint right away. If the rule changes later, it becomes version 2 and both verdicts stay on the card.
Some data is new. Options-chain history starts in March 2026, so a rule that uses it has only seen the market since then. Before you pay, the pre-check shows from which date we have the data.
The pre-check counts them first. Below 30 we do not sell the check in that form and suggest a looser condition instead.
That’s where we send your pre-check, and it lets us keep your draft for 30 days. No password and no newsletter: we use the address only for this check. There is a daily limit of free pre-checks per person; you see it only if you reach it.
The pre-check only counts past cases. The verdict is the real test: did your locked idea come true more often than chance, and would it have made money after fees.
Only if you agree. Then it appears in the ledger without your name, 30 days after the verdict. We delete it if you ask.
By email, and by Telegram if you connect it on the order page. You get a message when the rule is locked, if we have a question, and the moment the card is ready. The order page shows the status at any time.
No. You get the verdict card, kept for you for good. The data stays with us.
Did not find your answer? Ask us in Telegram, or write here.
Describe it, confirm your email or Telegram, and see your rule and how much past data it has, for free. Pay only for the verdict.