Suppose today there is a huge institutional investor (a mutual fund, a bank, a pension fund, etc.) who wants to buy a stock but does not want to buy it at the market price, so he hangs a buy order in the market.
Let's say the original market order book is $200. $1.01 x $1.03. 200, and then suddenly this silly institutional investor comes in with a $1.01 payment for 3,000 shares, and the order book becomes $3.200. $1.01 x $1.03. 200. And we usually call this silly institutional investor the "elephant", and the high-frequency trader knows that the $1.01 price is backed by a buy, so they raise their bid price by 1 cent to $1.02, and this strategy is called the Penny Jump.
If the high-frequency trader buys the stock after the price has not gone up because an elephant is supporting it below, he can quickly sell it back to the elephant for $1.01.
For high-frequency traders, the way they make a profit is actually very simple: they use the microstructure in the market to guess the intentions of their counterparts and then build a segment one step ahead of others.
For the elephant, he exposed his trading intentions by hanging a huge ransom note in the market and naturally became a target for high-frequency traders.
In the real world of stock trading, there should be very few such shrewd institutional investors who would openly and brazenly post huge quotes on the market. Instead, it is common for large institutional investors, who want to get out of a stock, to intentionally post huge quotes to create a false impression, to attract high-frequency traders to enter the market to push the stock price, and then another brain dump out, which is what I call the profit trade in the world.
For high-frequency traders, once this strategy is seen and "matched" by gaming, they turn to "counter-matching" and develop strategies to eat the tofu that institutional investors "matched".