DEV Community

Mateosoul
Mateosoul

Posted on

Modeling Temporary vs Permanent Price Impact

#polymarket #tutorial #automation #architecture

How quantitative traders model market impact to build smarter execution algorithms for prediction markets and automated trading systems.

Market impact is one of the most overlooked factors in algorithmic trading.

Polymarket trading bot

Many developers spend months improving prediction models, experimenting with machine learning, or optimizing signal generation, only to discover that their live trading performance falls well below backtest expectations. In many cases, the culprit isn't the strategy—it's execution.

Every order interacts with the market. Large trades consume liquidity, move prices, and influence the behavior of other participants. Understanding this phenomenon is essential when building a Polymarket Trading Bot or any automated trading system.

In this article, we'll explore the difference between temporary and permanent price impact, explain why they matter, model them mathematically, and demonstrate how they can be incorporated into Python-based execution algorithms.

Why Price Impact Matters

Imagine a prediction market with the following order book.

SELL

$0.64   120
$0.63   250
$0.62   500

--------------------

$0.60   600
$0.59   900

BUY
Enter fullscreen mode Exit fullscreen mode

A trader wants to purchase 1,500 YES shares.

Since only 500 shares exist at $0.62, the remaining order must consume liquidity at progressively higher prices.

Execution might look like this:

Price Shares Filled
0.62 500
0.63 250
0.64 120
0.65 300
0.66 330

Although the initial market price was $0.62, the average execution price becomes much higher.

This difference is known as market impact.

Unlike trading fees, market impact is often invisible in backtests—but it directly affects profitability.

The Two Types of Price Impact

Quantitative finance generally separates price impact into two categories:

1. Temporary Price Impact

Temporary impact is caused by liquidity consumption.

When a large order removes resting limit orders from the book, the market price moves upward. However, after liquidity providers replenish the order book, prices often return toward their original equilibrium.

Price

0.66 ──────▲
           │
0.64 ──────│──────
           │
0.62 ──────┼──────────────
           │
0.60 ──────┘

      Buy Order

Time →
Enter fullscreen mode Exit fullscreen mode

Characteristics:

  • Exists only while liquidity is being consumed.
  • Often reverses within seconds or minutes.
  • Highly dependent on order book depth.
  • Can be reduced using smarter execution algorithms.

2. Permanent Price Impact

Permanent impact reflects new information entering the market.

Suppose breaking news significantly changes the probability of an event occurring.

Example:

  • A central bank unexpectedly cuts interest rates.
  • Election results begin to arrive.
  • A major sports player suffers an injury.
  • An ETF receives regulatory approval.

Traders rapidly update their expectations, and the market establishes a new equilibrium.

Price

0.75 ─────────────────────────

0.65 ────────────────

0.55 ───────

Time →
Enter fullscreen mode Exit fullscreen mode

Unlike temporary impact, this movement generally does not reverse because the market has incorporated genuinely new information.

Temporary vs. Permanent Impact

Temporary Permanent
Caused by liquidity Caused by information
Usually reverts Usually persists
Execution-related Fundamental repricing
Depends on market depth Depends on new probabilities
Can often be minimized Cannot usually be avoided

Understanding the difference is critical.

Many inexperienced traders mistake temporary price movement for genuine market information, leading them to chase prices that soon revert.

Conversely, ignoring permanent impact can result in trading against a market that has already incorporated new information.

Why This Matters in Prediction Markets

Prediction markets differ from traditional equities in several important ways:

  • Liquidity is often fragmented.
  • Retail participation is high.
  • Prices represent probabilities rather than intrinsic value.
  • News can instantly alter expected outcomes.
  • Order books may be relatively thin.

These characteristics make prediction markets especially sensitive to both temporary and permanent price impact.

For example, a political market may trade quietly for hours before a televised debate begins. During the debate, thousands of participants react simultaneously, producing a mixture of genuine information-driven repricing and temporary liquidity shortages.

A sophisticated execution engine should distinguish between these two effects whenever possible.

A Simple Mathematical Model

One of the simplest ways to express observed execution prices is:

[

P_{execution}

P_{market}
+
Impact_{temporary}
+
Impact_{permanent}
]

Where:

  • Market Price is the quoted mid-price.
  • Temporary Impact reflects liquidity consumption.
  • Permanent Impact reflects newly incorporated information.

Although real-world models are more sophisticated, this decomposition provides an intuitive framework for analyzing execution quality.

Python Example

Let's build a simple simulation.

market_price = 0.62

temporary_impact = 0.015
permanent_impact = 0.010

execution_price = (
    market_price
    + temporary_impact
    + permanent_impact
)

print(f"Execution Price: ${execution_price:.3f}")
Enter fullscreen mode Exit fullscreen mode

Output:

Execution Price: $0.645
Enter fullscreen mode Exit fullscreen mode

Even a few cents of impact can significantly reduce expected returns over hundreds or thousands of trades.

Why Professional Trading Bots Care

Professional execution systems don't simply decide what to trade—they also optimize how to trade.

Instead of submitting one large market order, advanced systems may:

  • Split orders into smaller pieces.
  • Monitor liquidity in real time.
  • Adjust execution speed based on volatility.
  • Use passive limit orders whenever appropriate.
  • Pause execution during periods of excessive market impact.

Execution quality is often the difference between a profitable quantitative strategy and one that fails in production.

Coming Up Next

In Part 2, we'll dive deeper into:

  • Kyle's Lambda model
  • Almgren–Chriss intuition
  • Python simulations of temporary price decay
  • Order book imbalance analysis
  • Execution cost estimation
  • Building a market-impact module for a Polymarket trading bot

References

Top comments (0)