ta.go — Technical Analysis for Go

Technical analysis for Go (Golang): 150+ indicators, oscillators, bands, and statistics — RSI, MACD, Bollinger Bands, ATR, ADX, Stochastic, Ichimoku, SuperTrend, Parabolic SAR, VWAP, OBV, and 19 moving averages (SMA, EMA, WMA, Hull, KAMA, T3, VIDYA, ZLEMA). Includes a goroutine-parallel Monte Carlo simulator for backtesting and strategy research.

Pure Go, zero runtime dependencies, no cgo. Ties or beats go-talib and cinar/indicator at 100k bars on every benched indicator — see BENCHMARKS.

Built for algorithmic trading, quantitative finance, crypto and stock-market analysis, and backtesting systems that need indicators without a TA-Lib C dependency.

go get github.com/Bitvested/ta.go@latest

Contents

• Quickstart
• Packages
• Bar conventions
• Upgrading
• Performance
• Versioning
• Contributing
• License

Quickstart

ta.go is split into category subpackages. Import the ones you need:

package main

import (
	"fmt"

	indicators "github.com/Bitvested/ta.go/src/indicators"
	ma         "github.com/Bitvested/ta.go/src/moving_averages"
)

func main() {
	prices := []float64{
		44.34, 44.09, 44.15, 43.61, 44.33, 44.83, 45.10, 45.42,
		45.84, 46.08, 45.89, 46.03, 45.61, 46.28, 46.28,
	}

	fmt.Println("SMA(14):", ma.Sma(prices, 14))
	fmt.Println("RSI(14):", indicators.Rsi(prices, 14)) // Wilder smoothing
}

Packages and Indicators

Import path	What's inside
…/src/moving_averages (movingaverages)	Sma, Ema, Wma, Smma, Hull, Kama, Lsma, Cwma, Vwma, Vwwma, Wsma, Pwma, Hwma, Dema, Tema, T3, Trima, Vidya, Zlema
…/src/indicators	Rsi, Wrsi, Macd, MacdSignal, MacdBars, AroonUp, AroonDown, Mfi, Roc, Cop, Kst, Obv, Vwap, Mom, Tsi, Bop, Fi, Asi, Alligator, Pr, Stoch, Ichimoku, Atr, Fib, Fractals, FractalsPrice, HalfTrend, ZigZag, Psar, Supertrend, Elderray, Hv, Rvi, Rvi_signal, Rsi_divergence, Divergence, Adl, Apo, Cci, Cmf, Cross, Dpo, Pdm, Mdm, Pdi, Mdi, Dx, Adx, Adxr, Emv, Kdj, Mass, Natr, Nvi, Pvi, Ppo, StochRsi, Trix, Ulcer, Vortex
…/src/oscillators	Gator, MomOsc, ChaikinOsc, AroonOsc, Ao, Ac, Fisher, Kvo, Ult
…/src/bands	Bands, Bandwidth, Keltner, Don, Fibbands, Envelope
…/src/statistics	Median, Mad, Aad, Ssd, Std, StdSeries, Variance, Cor, Covariance, Dif, Drawdown, Normsinv, Ncdf, Sim, Percentile, Kmeans, Er, Ar, Winratio, Avgwin, Avgloss, Se, Kelly, Zscore, Normalize, NormalizePair, NormalizeFrom, Denormalize, Standardize, Log, Exp, Martingale, Antimartingale, Sum, Permutations, Mse, Cum, RecentHigh, RecentLow, Pvalue, ExpectedTrails, LrSlope, LrIntercept, LrAngle, Tsf, ReturnPositive, ReturnNegative
…/src/chart_types (charttypes)	Ha (Heikin-Ashi), Ren (Renko)
…/src/experimental	Support, Resistance, Line, LineCalc, DivergenceState
…/src/random	Prng, Pick, Range, Float, Order
…/src/misc	Fibnumbers, TimesUp, TimesDown
…/multi	Sim, SimSeeded — goroutine Monte Carlo

All paths are prefixed with github.com/Bitvested/ta.go. The root package only exports the Version constant — there is no monolithic re-export.

Bar conventions

Indicators that take OHLCV-style bars use a fixed column ordering:

Indicator family	Bar layout
Atr, Cci, Pdi/Mdi/Dx/Adx/Adxr, Pdm/Mdm, Vortex, Ult, Stoch, Kdj, Ichimoku, Halftrend, Psar, Supertrend, ZigZag, Don, Keltner	[H, C, L]
Adl, Cmf, ChaikinOsc, Kvo	[H, C, L, V]
Mfi	[H, L, C, V]
Emv	[H, L, V]
Nvi, Pvi, Vwma, Vwwma, Vwap	[close, volume]
Rvi	[O, H, L, C]

Restructure your input slices to match before calling.

Upgrading

If you're moving from a pre-v2.0.0 release, the most likely breakage points:

Area	What changed
Imports	The root ta package no longer exports indicators — switch to the category subpackages above.
indicators.Rsi	Now uses Wilder smoothing (TA-Lib convention). The legacy simple-average behaviour is preserved in Wrsi.
indicators.Roc	Returns rate of change in percent (×100). Cop and Kst cascade.
indicators.Mfi	Input is [][]float64 of [H, L, C, V] bars.
indicators.Atr, oscillators.Fisher, movingaverages.Smma	Output length is N − l + 1; TR uses absolute values. Keltner, Supertrend, Halftrend cascade.
statistics.Median	Even-length windows return the average of the two middle values. Mad cascades.
indicators.Macd	The base Macd returns just the MACD line. Signal and histogram split into MacdSignal and MacdBars.
indicators.Cross	Returns []CrossSignal{Index int; Cross bool}. Inputs are no longer mutated.
NaN policy	Rsi, Roc, Dif, Mom, MomOsc, Mfi, Fisher, Pr, Stoch return math.NaN() on flat / divide-by-zero edges instead of silent coercion.
No mutation	Std, Cross, Divergence, Percentile, and experimental.Support/Resistance no longer mutate caller inputs.

Full Phase-by-Phase breakdown in CHANGELOG.md.

Performance

ta.go ties or beats markcheno/go-talib and cinar/indicator at 100 k bars on every benched indicator. Headline numbers (lower is better, ns/op):

Indicator	ta.go	go-talib	cinar
SMA	141 µs	183 µs	221 µs
EMA	301 µs	299 µs	502 µs
RSI	544 µs	531 µs	1912 µs
MACD	936 µs	1108 µs	1602 µs
Bollinger	501 µs	625 µs	669 µs
ATR	513 µs	974 µs	762 µs

Full table and reproduction instructions in bench/BENCHMARKS.md.

Versioning

The user-facing version is 2.1.0, exported from the root package as ta.Version. The Git tag is v0.21.0 — the v0.x prefix lets the existing github.com/Bitvested/ta.go import path keep working without Go's /v2 module-path requirement. To pin the 2.1.0 line specifically:

go get github.com/Bitvested/ta.go@v0.21.0

import ta "github.com/Bitvested/ta.go"
fmt.Println(ta.Version) // "2.1.0"

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change. Please make sure to update tests as appropriate.

License

MIT

👨‍💻 Maintained By

ta.go is an open-source project created and actively maintained by Nino Kroesen.

This library was built with performance in mind for high-frequency data and algorithmic trading systems. If you are interested in seeing these indicators in action within a live quantitative trading environment, check out Bitvested.

---