Skip to content
Methodology

How Sentari's crypto signals work.

Sentari turns real-time financial news into a clear BUY, SELL, or HOLD verdict, then can auto-execute it on Kraken. This page explains how a verdict is generated, how auto-execution stays under your control, and exactly how we measure the track record, including the real numbers behind Sentari Classic.

Last updated Jun 15, 2026

How a verdict is generated

Sentari reads thousands of financial headlines an hour and scores the sentiment behind them. Rather than trust a single source, it looks for corroboration: when several independent sources point the same way, the signal is stronger and the confidence score is higher. That evidence is distilled into one verdict per asset, BUY, SELL, or HOLD, so a wall of news becomes a single, readable action.

  • Ingest: scan real-time financial news and market sentiment.
  • Corroborate: weigh agreement across multiple independent sources.
  • Render: output a BUY / SELL / HOLD verdict with a confidence score.
  • Gate: only act when confidence clears the conviction threshold you set.

How auto-execution stays under your control

When a verdict clears your conviction threshold, Sentari can place the trade on your Kraken account automatically. It is non-custodial: your funds never leave your account, the API key grants query and trade access only (never withdrawal), and keys are encrypted server-side. You set a capital cap, can pause at any time, and can run the whole thing in simulation mode first.

How we measure performance

Every performance number on this site comes from actual closed trades, computed the same way the app computes your own stats. Realized profit and loss for a trade is size × (close − entry) ÷ entry, and cumulative equity compounds from a $1,000 base. Only closed trades are counted; open positions are excluded until they close.

The headline figure, +231.63%, is the return on that $1,000 base over 760 closed trades since Aug 2, 2025, as of Jun 15, 2026.

Backtest vs live

The Sentari Classic record combines a one-year backtest with the live run since launch, and we always label which is which. Backtests can flatter a strategy because they do not feel real slippage or changing conditions, so live results can trail them. That is exactly why we publish the full breakdown of real closed trades rather than a single marketing number.

The record in full

Sentari Classic, from a $1,000 base, tracked since Aug 2, 2025, as of Jun 15, 2026. Every row is computed from the actual closed trades.

Starting capital$1,000.00
Tracked sinceAug 2, 2025
Closed trades760
Wins / losses / breakeven494 / 262 / 4
Win rate65.00%
Average win$8.93
Average loss-$8.00
Average profit per trade$3.05
Profit factor2.10
Best trade$80.82
Worst trade-$95.00
Total profit$2,316.28
Return on capital231.63%
Max drawdown10.41% ($201.30)
Max consecutive wins15
Max consecutive losses8
Average holding time13h 6m
Median holding time6h 36m

You can also see the equity curve and headline stats on the landing page.

Why the sample size matters

A track record is only as trustworthy as the number of trades behind it. A handful of trades is noise; a few hundred starts to show a real edge. At 760 closed trades with a 65% win rate and a profit factor of 2.10, Sentari Classic rests on a large sample, not a lucky run, while a 10.41% max drawdown shows the roughest stretch along the way.

Frequently asked questions

How does Sentari generate a buy, sell, or hold verdict?
Sentari's AI continuously reads real-time financial news and market sentiment, weighs the evidence across multiple independent sources, and turns it into a single verdict, BUY, SELL, or HOLD, with a confidence score. The more sources corroborate a move, the higher the confidence.
What does the confidence score mean?
Confidence is how strongly the evidence supports a verdict, shown as a percent. A higher score means more, and more consistent, corroboration. You can set a conviction threshold so the bot only acts on verdicts above a confidence level you choose.
How is the total return calculated?
It is the cumulative realized return of Sentari Classic's actual closed trades, from a $1,000 starting balance, tracked since Aug 2, 2025. Realized profit and loss per trade is size times (close minus entry) divided by entry; equity compounds from there. Only closed trades count; open positions are excluded.
Is the track record a backtest or live trading?
Both, and we label it as such. The record combines a one-year backtest with the live run since launch. Live results can differ from a backtest because of slippage and changing market conditions, which is why we publish the real closed trades behind every number.
Is 760 trades a meaningful sample size?
Yes. A common rule of thumb is that 30 to 100 trades is directionally useful and 100 to 300 is genuine evidence of an edge. At 760 closed trades, the win rate and profit factor rest on a large sample rather than a lucky streak.
Can past performance predict future results?
No. No track record guarantees future returns, and any tool that promises a 100% win rate is a red flag. Sentari shows you the real record so you can judge it yourself, and lets you start in simulation mode with zero risk.

Sentari provides software and information, not financial advice. Crypto trading involves risk, including the loss of capital. Past performance is not a guarantee of future results.