What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Yes, but only under specific conditions. Automated strategies can react to price, volume, and liquidity changes that other automated strategies have partly caused. When those reactions add pressure in the same direction, they form a feedback loop, and a participant on the wrong side of it can end up with worse execution and real losses. Losses are not a universal outcome of bots interacting, and most of the detailed evidence comes from studies of foreign-exchange execution algorithms. It should not be transplanted to equities, futures, crypto, or retail trading without separate evidence.
How the feedback loop forms
The loop has five links. None of them is guaranteed, and each can break.
- A market event changes price or turnover. The trigger might be news, a large sell order, or a sudden drop in trading volume.
- Reactive algorithms adjust. Execution programs change their pace, their direction, or their quotes in response to the new state of the market.
- Their orders change the order book. Those adjustments move the book and the prices at which trades execute.
- Other systems read the change as a signal. Strategies that watch price or volume respond to the new state, which adds their own orders to the pressure.
- Liquidity thins. Liquidity providers may widen, reduce, or withdraw their quotes under stress, leaving less depth to absorb the next wave of orders.
The “play against themselves” phrase is a metaphor. The sequence describes independent strategies whose responses happen to reinforce one another. Nothing in the evidence shows that one bot deliberately trades against its own positions, or that the strategies share an objective or coordinate with one another.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →When the loop is most likely
The Bank for International Settlements (BIS) report on FX execution algorithms, published 30 October 2020, points to conditions rather than a general law. Feedback risk is associated with the following three conditions, and it is not an inevitable property of automated trading.
#1 Best Overall
Outsized orders
When one order or program is large relative to the depth available, its own footprint moves the price enough to trigger other systems. The same order in a deep, stable book may leave prices largely unchanged.
Correlated and crowded strategies
Algorithms designed individually to reduce their own market impact can collectively reinforce it when their logic and responses are correlated. Several similar strategies reacting to the same external event can have the market impact of a much larger order, even though each one looked small on its own.
Thin liquidity
With less depth in the book, the same amount of selling or buying moves prices further. This is the step where a loop becomes harder to absorb, because each round of orders meets fewer resting quotes.
Recommended Free Tools
A worked example: selling into a flash crash
The BIS report’s central example is a participation-of-volume (POV) algorithm. A POV program scales its execution pace to a share of market turnover. During a flash crash, turnover rises sharply as prices fall. A selling POV algorithm then trades faster into the decline, which can add to the selling pressure. The report also describes the opposite case: a buying execution algorithm can support prices during a drop and help start a rebound. Direction and context therefore decide the outcome.
| Execution logic during a sharp move | Reaction as turnover rises | Possible effect on prices (BIS example) |
|---|---|---|
| Selling POV algorithm | Sells faster, because its pace is tied to turnover | Can increase selling pressure during a flash crash |
| Buying execution algorithm during a drop | Buys into the decline | Can support prices and help start a rebound |
The same logic that intensifies selling can intensify buying. Whether it helps or hurts depends on which side of the market the algorithm is on, how large it is relative to depth, and what other strategies are doing at the same time.
When the same activity steadies prices
The BIS report does not treat execution algorithms as purely destabilising. It says they can improve matching efficiency and liquidity provision, and that they can introduce new execution and market-structure risks at the same time. Its discussion of the Covid-19 pandemic adds a balancing point: initial observations suggested the self-reinforcing risks might not have been as acute as previously believed. Those were early observations from 2020, not a final verdict, and the report does not put a figure on how often the loop occurs.
What “at a loss” can mean in practice
A strategy that trades into a feedback loop can lose money through several distinct channels. Only some of them involve interaction with other bots.
- Adverse execution: fills at prices worse than the price the strategy expected when it decided to trade.
- Slippage: the gap between expected and actual execution prices, which tends to widen when liquidity is thin.
- Crowded trades: many strategies on the same side try to exit at the same time, so the exit price falls as the exit crowd grows.
- Trading-system risk: a system that sends erroneous orders, or lacks limits to stop it, can cause losses without any market interaction at all.
The last item matters because it is a separate failure. Interaction between independent strategies is a market-level effect. A single firm’s malfunction is an operational risk, and the regulatory controls discussed below target it directly.
The 2010 Flash Crash as a carefully attributed case
On 6 May 2010, US equity markets fell sharply and recovered within minutes. The US Securities and Exchange Commission (SEC) staff report on those events describes a complex episode involving interacting liquidity and trading dynamics. The staff review found evidence that liquidity withdrawal and algorithmic trading could contribute to feedback effects. It also summarised studies that are generally consistent with the view that high-frequency traders did not cause the crash, although their withdrawal may have exacerbated the declines.
The event therefore shows how feedback can deepen a move. It is not evidence that bots caused the crash.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Institutional controls for algorithmic trading
The Financial Conduct Authority (FCA) Handbook, section MAR 7A.3, says firms engaging in algorithmic trading must have effective systems and controls. The rule text lists the control areas shown below. The FCA rule page gives 1 January 2021 as its last update, so check the live Handbook wording before quoting it as current law.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problems| Control area (as listed in MAR 7A.3) | Question to ask of your own strategy |
|---|---|
| System resilience and capacity | How does the system behave when message volume spikes well above normal? |
| Appropriate trading thresholds and limits | What are the maximum order size, position, and loss before the strategy stops? |
| Prevention of erroneous orders or disorderly-market contribution | What checks stop a bad price or quantity from reaching the market? |
| Business continuity | If the connection or data feed fails, what happens to open orders? |
| Testing | Has the strategy been tested on stressed price histories, not only calm ones? |
| Monitoring | Can you see order behaviour and execution quality while the strategy runs? |
What the FCA’s 2025 review added
The FCA’s multi-firm review of algorithmic trading controls, published 21 August 2025, says algorithmic trading firms can materially affect price formation and liquidity. The reasons are their trading footprint, their strategies, and their role linking fragmented markets. The FCA says controls and oversight need to keep pace with complexity, speed, and technological change. The review created no new requirements. It was intended to help firms comply with existing ones.
Managing the risk in your own strategy
The FCA rules apply to regulated firms, not to individual traders. The steps below translate the same control categories into practical checks for a personal or small-team strategy. They are a sensible framework, not a regulatory requirement.
Quick Recap
- Size orders against visible depth. Compare your order size with the depth you see at the prices you trade. If you would make up a large share of the visible book, split the order or reduce it.
- Set hard limits before you need them. Define the maximum order size, position, and daily loss in advance, and make the strategy switch off automatically when a limit is breached.
- Test under stress. Include periods when volume and price spiked, not only calm sessions. A strategy that behaves well in quiet markets may respond very differently when turnover jumps.
- Watch execution, not just profit and loss. Track slippage against the price at the moment of the decision, and fill rates. Flag cases where slippage rises alongside falling volume.
- Plan for failure. Decide what happens to open orders if the connection drops, and keep a manual kill switch that works without relying on the code itself.
What the evidence does and does not establish
- No cited source measures how often feedback loops occur, so claims about frequency are not supported.
- Controls reduce the chance of a self-inflicted failure and of disorderly-market contribution. The cited material does not show that they guarantee protection from losses caused by market interaction.
- The BIS findings on feedback loops rest on FX execution algorithms, and the Covid-era observations were preliminary when published in 2020.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
