Curriculum·S210 Automation Literacy: What Bots Actually Do·about 31 min
The unifying insight
By the end of this lesson you can
- →State the single property every automated strategy shares, and derive its consequences
- →Identify which market state a given bot is betting on, from its rules alone
- →Explain why a bot cannot detect the regime it is wrong in
- →Predict a strategy's failure mode before deploying it, rather than discovering it
Sophomore · enrolled learners
This lesson opens with The grid that did exactly what it was told.
- What happened
- During the deleveraging of 10 October 2025, more than $19B of leveraged positions were forcibly closed across more than 1.6 million trading accounts, and CoinGlass put long positions at roughly 85 to 90 percent of the liquidations. We found no published account of how grid bots specifically fared that day, so what follows is derived from the rules rather than reported. A futures grid carries leverage, and its operators document that a sustained move against it can end in liquidation, so a long-biased futures grid faced the same liquidation engine as every other leveraged long. A spot grid cannot be liquidated, because no borrowed capital is involved, and in a decline like that one its rules do what they were designed to do: keep placing buy orders into the fall and finish holding the full intended position at an average cost far above the prevailing market. Neither outcome requires any component to fail.
- The decision point
- A grid is a rule that buys declines and sells rallies. A sustained trend is therefore not an edge case for it; it is the condition under which the rule is exactly wrong, in every increment, with increasing size. And the rule contains nothing capable of noticing this, because a rule that could tell which regime it was in would not need to be a grid.
What you will be able to answer
- →What property does every automated strategy share?
- →Which state does a grid bet on, and how does it fail?
- →Why can a bot not detect the regime it is wrong in?
- →How do you predict a strategy's failure mode?
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
- https://www.coinglass.com/learn/2025-annual-report-en
- https://www.coingecko.com/learn/october-10-crypto-crash-explained
- https://www.coindesk.com/research/market-spotlight-the-19-billion-liquidation-that-shook-crypto
- https://intercom.help/pionex/en/articles/14616311-futures-grid-bot
- https://onlinelibrary.wiley.com/doi/abs/10.1111/0022-1082.00163
- https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2308659
Confidence medium·Volatility high·Reviewed 2026-08-06·Owner unassigned
Contested
S203-04 owns the three-state taxonomy and the regime classification argument. This lesson applies it to automation. Keep the split.
The claim that spot grid strategies accumulate the asset in a decline is a property of the mechanism rather than an empirical finding about any particular deployment. Present it as derivable from the rules, which it is. We found no published account, from a venue, a bot operator or a researcher, of how grid bots specifically fared on 10 October 2025, so the lesson does not say they were a failure point that day; it says what their rules imply in a decline of that kind.
The $19B and the 85 to 90 percent long share are from CoinGlass's 2025 annual report, which adds that the true nominal figure was likely $30B to $40B once disclosure delays are counted. The 1.6 million figure is trading accounts, from CoinGecko, not unique traders.