Crypto Charting Resources — API Fields, Formulas, and Checklists

CoinDock candle data

Endpoint — no account required:

GET /api/v1/markets/{symbol}/candles?interval=1h

Intervals: 1m, 5m, 15m, 30m, 1h, 4h, 1d, 1w

Fields per candle:

Field	Meaning
`open_time`, `close_time`	Interval boundaries
`open`, `high`, `low`, `close`	The four prices
`volume_base`	Quantity of the base asset traded
`volume_quote`	Value traded in the quote asset
`trades_count`	How many trades built this candle
`is_closed`	Whether the interval has finished

The last two are what let you distinguish a real market from a handful of transactions. Full documentation: coindock.online/docs/api.

Candle formulas

How a candle is built

open   = price of first trade in interval
high   = max(price) across interval
low    = min(price) across interval
close  = price of last trade in interval
volume = sum(quantity) across interval

Body ratio — how decisive an interval was

body ratio = |close − open| / (high − low)

Near 1: decisive movement. Near 0: wide range, ended where it started — a lot was rejected.

Percentage change

change % = (close − open) / open × 100

Percentage changes are not symmetric (a 50% fall needs a 100% rise to recover) and do not add across periods — they compound.

Range

range   = high − low
range % = (high − low) / low × 100

Data-quality checklist

Run before analysing any candle or chart:

is_closed — exclude forming candles, or handle them explicitly
trades_count — flag or filter single-digit-count candles
Volume denomination — base for one pair over time, quote across pairs
Empty intervals — omitted, or carried forward as flat candles?
Timestamp convention — labelled by open time or close time?
Interval boundaries — which time zone?
Venue — candles aggregate only that venue's trades

Chart-reading routine

Four prices identified from body edges and wick tips
Direction read from the platform legend, not assumed
Body-to-range ratio noted
Preceding candles reviewed for context
Trade count checked
Closed status confirmed
Order-book depth checked if I intend to act

Timeframe selection

Holding for	Primary	Context
Minutes to hours	`1m` – `15m`	`1h`
Hours to days	`1h` – `4h`	`1d`
Days to weeks	`4h` – `1d`	`1w`
Weeks and longer	`1d` – `1w`	`1w`

Use two — one for decisions, one for context. Choose before looking. Searching eight timeframes will always produce a pattern in noise.

Volume sanity checks

Check	Inconsistency looks like
Volume vs resting depth	Large volume, empty book
Volume vs `trades_count`	Large volume, very few trades
Volume over time	Suspiciously regular rather than uneven
Volume vs spread	Heavy trading, persistently wide spread

Backtest pitfalls

Unclosed candles excluded (look-ahead bias)
Low-trade-count candles filtered — the largest single problem in crypto
Slippage modelled — assuming fills at the close ignores your own price impact
Fees modelled — maker/taker, and gas on DEXs
Survivorship addressed — delisted and abandoned tokens are missing from any current listing
Out-of-sample data reserved before starting
Parameter sensitivity checked — if it works at 14 and fails at 13 and 15, it is fitted to noise
Multiple regimes covered — more rows of one regime is not more evidence

What CoinDock does not publish

No signals, no forecasts, no pattern or indicator success rates. We have not run that research, and repeating uncited figures would be worse than saying nothing.

Where any source quotes a specific win rate, ask: what market, what timeframe, what definition, what period, and out of sample?

Guides

Concepts

How-to

Reference

Charting Resources — Formulas, Fields, and Checklists