Curriculum·G905 Backtesting and Validation·about 33 min

Chasing a signal that means nothing

By the end of this lesson you can

  • Explain that a signal with no real information is noise, and acting on it is acting on nothing
  • Describe how Long Island Iced Tea's stock spiked on a blockchain name change with no business behind it
  • Reason that a correlation without an economic reason is likely spurious
  • Require a real, causal reason before trusting a signal a backtest happens to reward

Graduate · enrolled learners

This lesson opens with Long Island Iced Tea to Long Blockchain, December 2017.

What happened
Long Island Iced Tea Corp was a small beverage company whose shares traded quietly. In December 2017, at the height of the crypto mania, it announced that it would rename itself Long Blockchain Corp and pivot toward blockchain technology, and on the name change alone its stock spiked several hundred percent in a day, briefly making the renaming one of the most profitable events imaginable to have traded. There was no blockchain business behind the name, no product, no revenue from any such pivot, only the word in the company's title, and the stock later collapsed, the company was delisted, and the SEC pursued related insider trading. Nothing about the price move was a signal in the meaningful sense: the market rewarded a word, a pattern with no information about the company's actual value, so a trader or model that traded the pattern was trading noise dressed as opportunity. The move was large, real, and completely empty of the thing that makes a signal worth acting on, a reason.
The decision point
A signal is only worth acting on if it carries real information, a reason connected to value, and a pattern that moves prices without such a reason, a name change, a coincidence, a correlation with no mechanism, is noise, so acting on it is acting on nothing, however large and real the move looks. Long Island Iced Tea is the case: a stock spiked several hundred percent because the company put the word blockchain in its name, with no business, product, or revenue behind it, a pattern that was large and real and utterly empty of information about the company's value, and it later collapsed. This is the counterpart to overfitting from the other side: overfitting finds patterns in the past that are noise; this lesson is about acting on patterns in the present that are noise, and both fail because a pattern without a reason does not predict, it merely happened. The error is to treat any pattern that has moved prices, or that correlates in the data, as a tradeable signal, when a correlation without an economic mechanism is likely spurious and a price move without a value reason is likely to reverse. So the discipline is to require, before trusting a signal, a real and causal reason why it should connect to value, to ask what mechanism links the pattern to the thing being priced, and to reject a signal that rests only on the fact that prices moved or numbers correlated, because Long Island Iced Tea showed that the market itself will chase a word with real money, and a system built to follow the market's patterns will follow it straight into the noise unless it is required to find the reason first.

What you will be able to answer

  • Why did Long Island Iced Tea's stock spike (December 2017)?
  • What makes a signal worth acting on?
  • What is a correlation without an economic mechanism?
  • What must a trader require before trusting a signal?

Orientation and Year One are open: anyone can read them without an account. From Year Two onward the lessons are for enrolled learners, because progress through the later years only means anything if it is tracked against a record.

It is free. We do not sell the list and there is nothing to buy at the end of it.

Sources and review

Confidence high·Volatility low·Reviewed 2026-09-18·Owner unassigned

Contested

The loss is recorded as 0 because the harm fell on those who bought the spike and on the later delisting, not as a single quantified figure; the SEC pursued related insider trading. The lesson uses the meaningless-signal mechanism rather than a dollar loss.

The Long Blockchain episode involved later enforcement and other details; this lesson uses only the core fact, a price spike on a name change with no business behind it, which is the transferable validation lesson.