Manual · 1 of 5
ask TPSL manual
What TPSL searches
TPSL holds every 15-minute candle of Bitcoin and Ethereum since 1 January 2020 from six exchange feeds, and for each candle 110 technical indicators computed the same way every time. A search returns the bars that satisfy a condition — and what happened after them.
- Coins
- BTC · ETH
- Bar
- 15 min, UTC
- History
- since 1 Jan 2020
- Bars per coin
- ≈ 234,000 / feed
- Feeds
- Binance · OKX · Coinbase · Upbit
- Indicators
- 110 per feed
- Freshness
- ≈ 3 s after each quarter hour
- Precision
- 4 decimals
Two ways to ask
1. Ask in plain language
Describe a market condition in English or Korean. TPSL translates it into a condition of the query language, runs it over the full history, and shows every matching bar with what happened afterwards. The four cards below are the home-page examples: two questions and two conditions written out (the ones starting with #).
Show me BTC when it dropped more than 3% within an hour
Try this →BTC volume spike with a big green candle — did the move continue?
#bn_spot_volume_rank_long>=0.95 and bn_spot_body>=1 and entity=BTCTry this →이더리움이 하루 만에 5% 넘게 빠졌을 때
The three moments of a row
One row is one coin on one 15-minute bar. Every number on the result page belongs to one of three moments, and the page labels them so you always know what was knowable when.
| Moment | Bar | Which bar | Columns | Known… |
|---|---|---|---|---|
| Setup | t − 1 | the previous completed bar | every indicator column, p:close | known before this bar opened |
| Signal bar | t | the row's own bar | open · high · low · close · volume, Result, X/Prev | known when this bar closed |
| Outcome | t + 1 … t + 96 | the bars after it | +15 m / +1 h / +4 h / +1 d, run-up, drawdown | known only later |
Why the indicators are one bar behind
Reading a result
- Executed query — the exact condition that ran. Copy it, edit it, share the URL.
- Outcome summary — for +15 m, +1 h, +4 h and +1 d after each matched bar: the share that closed higher, the average and median move, the best and the worst. Run-up and drawdown show the typical highest high and lowest low reached within 1 h and 4 h, with their 10th percentiles as take-profit / stop-loss reference lines.
- Sample — how many bars matched. Fewer than about 30 is an anecdote, not a pattern.
- Table — one row per matched bar, newest first; expand a row for its Setup / Bar / Outcome details. The Data tab shows every column.
- Signal bar and Setup snapshot — averages of the matched bars and of the indicators just before them.
- Coverage notice — appears when some matched bars carry no indicator values (warm-up or exchange gaps); those cells show “—” and are left out of the averages.
Everything on the page is descriptive history. A condition that averaged +0.4 % the next day is a tendency measured after the fact, never a forecast — see the risk disclosure.
Five-minute quick start
- On ask TPSL, type a question or a condition starting with
#and press Enter. Leave the dropdowns on Binance spot for now. - Read the Executed query line. Adjust one number and rerun until it says what you mean.
- Look at the sample count and the +1 h / +1 d rows of the Outcome summary. Check the drawdown line before you get excited about the run-up line.
- Narrow or widen: add
and date>=today-30, change the coin withentity=ETH, or compare thresholds with a comma (bn_spot_rsi<=30,25,20). - Open the Data tab for every column, or send the condition to Deep Analysis to let it search for stronger variants.
#bn_spot_rsi<=30 and bn_spot_volume_rank_long>=0.9 and entity=BTC
Manual sections
Credits & pricing
- 1 credit = $0.10 — top up on the Credits page.
- Daily Future challenges cost 15 c (Non-Guaranteed) or 30 c (Guaranteed).
- Data Market preview unlocks cost demo credits (e.g. 3,000 c); live updates and delivery are not connected.
- No money is charged in this preview. Questions about future billing can be sent through the Support Center.
- The full price list lives on the Pricing page.
Challenges & the refund rule
Guaranteed and Non-Guaranteed challenges contain identical predictions. The Guaranteed Challenge is 100% refunded if the prediction is not profitable; the Non-Guaranteed Challenge is non-refundable but costs half. Results and settlement land in My challenges.