Algorithms now execute the majority of trades on every major financial exchange. Algorithmic trading accounts for roughly 60 to 75 percent of total trading volume in U.S. equity markets, European financial markets, and major Asian capital markets. In crypto markets the proportion is similar. The speed, scale, and consistency of automated systems have reshaped market structure in ways that seemed improbable two decades ago.
Yet inside this automated environment, the role of human judgment has not disappeared. It has shifted. Understanding why requires looking at what algorithms actually do, where they consistently fall short, and why the human signal provider model that emerged in the 1990s and expanded through copy trading platforms onto Telegram channels remains structurally relevant in 2026.
What Algorithms Do Well, and What They Do Not
A trading algorithm is a set of rules applied to data. Its strengths are well understood: it executes without hesitation, scales without degradation, monitors multiple instruments simultaneously, and responds to price changes in microseconds. A good algorithm applied to a stable market regime will outperform most discretionary traders on execution quality alone.
The limitation is equally well understood by practitioners, though less frequently stated plainly: algorithms are reactive systems. They respond to patterns in data that were identifiable when the strategy was built. They do not understand context. They cannot read a central bank statement and weigh the political pressures behind it. They cannot assess whether a company founder departure signals strategic drift or routine succession. They process what is measurable, not what is meaningful.
There is growing evidence of investors with bounded rationality in the market whose trading decisions are influenced by sentiment and beliefs, and that investor sentiment is an important factor influencing trading decisions. The mirror image of that observation is equally true: algorithms, which have no sentiment at all, create exploitable conditions for traders who can read what is driving sentiment in real time.
The Crash That Illustrated the Problem
The most consequential demonstration of what happens when automated systems interact without human oversight occurred on May 6, 2010. A large fundamental trader initiated a sell program to sell 75,000 E-Mini S&P contracts valued at approximately $4.1 billion using an automated execution algorithm set to target an execution rate of 9% of the trading volume calculated over the previous minute, but without regard to price or time.
Automated trading systems used by many liquidity providers then temporarily paused in reaction to the sudden price declines, as these built-in pauses are designed to prevent automated systems from trading when prices move beyond pre-defined thresholds, in order to allow traders and risk managers to fully assess market conditions before trading is resumed.
The result was that within twenty minutes, U.S. equity markets lost approximately one trillion dollars in market value before recovering nearly all of it within the same session. The SEC and CFTC joint report published in September 2010 documented the sequence in detail and remains the authoritative account of how cascading automated decisions, none of which were individually wrong by design, collectively produced a market breakdown that no single algorithm anticipated or intended.
The Flash Crash did not kill algorithmic trading. It accelerated the understanding, inside both regulators and trading firms, that automated systems operating without human oversight at key decision points carry systemic risks that individual strategy performance metrics do not capture.
The Value That Human Analysis Provides
The original article this piece replaces made an argument about investor relations: that the market-moving influence of a rational human investor’s buy or sell decision had grown, not shrunk, as algorithmic volume increased. The logic was that algorithms, watching each other constantly, interpret a new fundamental investor’s position as a new statement of intrinsic value and reset their parameters around it. Human conviction, in that framing, becomes the anchor that algorithmic activity orbits.
That argument has held up. Research published in the Journal of Accounting Research in 2025 found that algorithmic trading reduces monitoring from fundamental investors and exacerbates price reactions to bad news, increasing stock price crash risk.
When research-driven investors are crowded out of daily volume by automated activity, the informational quality of prices can deteriorate even as liquidity metrics appear healthy.
This is the structural reason why human signal providers retain value in algorithmic markets. The signal provider who studies an asset, understands its fundamentals, monitors its order flow, and makes a considered judgment about direction is supplying something that no rule-based system can generate internally: interpretation of meaning rather than pattern.
How the Signal Provider Model Evolved
The earliest version of this model was institutional. A fundamental analyst at a bank or asset manager produced research. Portfolio managers read it, formed a view, and traded. The signal traveled through a structured professional process, and the trade that followed carried genuine informational weight.
The retail version emerged in the 1990s when systems like Collective2 allowed independent traders to publish their signals and attract subscribers. The signal provider retained the human judgment component while the execution layer became progressively automated. Subscribers could connect signals directly to their brokerage accounts, and the provider’s track record became visible to anyone evaluating whether to follow them.
Copy trading platforms like eToro’s CopyTrader, introduced in 2010, moved this model onto social infrastructure. Performance became transparent, with drawdown history, win rate, and open positions visible before a subscriber committed capital. The provider was now accountable in real time to a public audience, creating different incentive structures from anonymous system developers selling black-box strategies.
Telegram channels, which became the dominant distribution mechanism for crypto trading signals after 2017, stripped the infrastructure back further. No centralized platform, no standardized performance verification, no regulated accountability mechanism. The signal arrives as a message. The subscriber decides whether to act. The provider’s reputation is the only quality filter, and track records are typically self-reported.
The decentralization increased access and reduced friction. It also removed most of the protections that regulated environments provide, a tradeoff that continues to define the channel-based signal economy.
The AI Challenge to Human Signal Providers
The most serious structural challenge to human signal providers in 2025 and 2026 is not better algorithms. It is AI agents capable of performing tasks previously requiring human analysts: reading earnings transcripts, monitoring news flows, synthesizing sentiment data, and generating trade recommendations with justification attached.
Research published by the Federal Reserve in 2025 found that AI agents make more rational decisions than humans, relying predominantly on private information over market trends, and exhibit less herd behavior than human financial professionals, a finding with significant implications for future financial stability as generative AI gains traction in market decision making.
That is a meaningful capability. An AI agent that reads ten thousand news items overnight and surfaces the three most relevant to a given position is providing genuine analytical leverage. But the same research notes that AI agents are not strictly algorithmically rational and have inherited some elements of human intuition and bias, consistent with findings across the broader literature showing that large language models can replicate human errors.
The practical implication is that AI agents in trading are not replacing human judgment so much as automating certain components of it, while inheriting some of its weaknesses in the process. A human signal provider who understands what an AI agent can and cannot do, and who builds their analysis around the interpretive and contextual work that agents still handle poorly, remains differentiated.
Where Human Judgment Remains Irreplaceable
Three specific domains consistently resist full automation, and they correspond directly to the competencies that distinguish strong signal providers from algorithmic noise generators.
The first is regime identification. Most trend-following and mean-reversion strategies perform well in the market conditions for which they were optimized and fail in others. A human analyst watching macro conditions shift, central bank communication change, or liquidity structures transform can recognize when a regime is changing before the statistical signals that an algorithm would require to confirm it. That early identification is a judgment call rather than a calculable output.
The second is narrative interpretation. Markets price expectations, and expectations are shaped by stories. An earnings beat with lowered forward guidance is not the same as an earnings beat with raised guidance, but both move price in initially similar ways before the narrative distinction emerges. A human reading the call transcript and management tone provides a category of input that quantitative systems treat poorly and language models handle inconsistently.
The third is accountability and conviction. A signal provider with a documented track record, who publishes reasoning alongside trade calls and updates their view when evidence changes, is providing something qualitatively different from an algorithm generating outputs no subscriber can interrogate. In markets where trust is the scarce resource, human accountability is a feature rather than a limitation.
Algorithmic Volume and Human Significance
The paradox at the center of the original argument remains valid: as automated trading came to dominate market volume, the influence of each rational human investor’s decision grew rather than shrank. Algorithms amplify and transmit human conviction rather than replace it.
Research published in Scientific Reports in 2025 confirmed that algorithmic trading can significantly reduce market volatility in normal conditions, but the relationship between automated activity and human investor behavior remains complex, mediated by sentiment, information asymmetry, and institutional structure.
In practical terms: a crypto market where 60 percent of volume is algorithmic is a market where a well-reasoned human signal, published to a subscriber base with the credibility to act on it, moves price in ways that pure automated activity cannot replicate. The algorithms are running. They are waiting for something to react to. The human signal provider who understands that dynamic, and who supplies analysis of genuine quality, is providing exactly what the automated market needs but cannot generate for itself.
References and Further Reading
- SEC and CFTC. “Findings Regarding the Market Events of May 6, 2010.” https://www.sec.gov/news/studies/2010/marketevents-report.pdf
- Hansen, A.L. and Lee, S.J. “Generative AI and Financial Stability: Evidence on AI Agent Herd Behavior.” Federal Reserve Finance and Economics Discussion Series, 2025. https://www.federalreserve.gov/econres/feds/files/2025090pap.pdf
- Ahmed, A.S. et al. “The Rise of Machines: Algorithmic Trading and Stock Price Crash Risk.” Journal of Accounting Research, 2025. https://journals.sagepub.com/doi/10.1177/0148558X251394477
- Wikipedia: Automated trading system. https://en.wikipedia.org/wiki/Automated_trading_system
