Calculating Profits and Losses for Algorithmic Trading Strategies: A Short Guide

James B. Glattfelder,Thomas Houweling

2024-11-21

Abstract:We present a series of equations that track the total realized and unrealized profits and losses at any time, incorporating the spread. The resulting formalism is ideally suited to evaluate the performance of trading model algorithms.

Trading and Market Microstructure

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: how to accurately calculate the profit and loss (PnL) in algorithmic trading strategies, including realized and unrealized profit and loss, and take the spread into account. Specifically, the author proposes a set of formulas to track the total profit and loss at any point in time, thus providing an ideal framework for evaluating the performance of trading model algorithms.

### Main Problem Decomposition

1. **Calculating Profit and Loss**:
- The paper details how to calculate profit and loss in different situations, including handling fees, compound interest calculations, and distinguishing between realized and unrealized profit and loss.
- For example, when buying or selling a currency pair, the difference between the buying and selling prices (i.e., ask price and bid price) needs to be considered.

2. **Formulaic Description**:
- The author defines how to calculate profit and loss through a series of formulas. For example, for a currency pair, the holding amounts of the base currency and the quote currency are represented by \( b_i \) and \( q_i \) respectively:
\[
b_i=\sum_{j = 1}^i u_j,\quad q_i = -\sum_{j = 1}^i x_j u_j
\]
where \( u_j \) is the quantity of the \( j \) -th transaction and \( x_j \) is the transaction price.

3. **Average Price and Profit and Loss**:
- The unrealized average price \( \bar{x}_i \) is defined as:
\[
\bar{x}_i = -\frac{q_i}{b_i}
\]
- The formula for calculating the profit and loss \( p_b^i \) is:
\[
p_b^i=\begin{cases}
b_i\left(1-\frac{\bar{x}_i}{x'_i}\right),&\text{if }b_i\neq0,\\
-u_i\left(1 - \bar{x}_{i - 1}+\hat{x}_i\right),&\text{otherwise}
\end{cases}
\]
where \( x'_i \) is the opposite price corresponding to \( x_i \) (i.e., bid/ask pair), and \( \hat{x}_i \) and \( \mu \) are determined according to the trading direction.

4. **Wealth Change**:
- The initial wealth \( W \) includes the base currency balance \( B \) and the quote currency balance \( Q \), and can be represented as:
\[
W_b^i = B+\frac{Q}{x'_i},\quad W_q^i = Q+x'_i B
\]
- As trading activities change, the wealth \( W_b^i \) and \( W_q^i \) will also change accordingly.

5. **Performance Evaluation**:
- By representing the trading unit as a percentage of the initial balance, the performance value \( \tilde{p}_i \) of each transaction can be calculated, and the total profit and loss can be further calculated:
\[
p_b^i=\tilde{p}_i B,\quad p_q^i=\tilde{p}_i B x'_i
\]

### Summary

The core objective of this paper is to provide a set of accurate and easy - to - understand methods for calculating the profit and loss in algorithmic trading strategies, ensuring a comprehensive evaluation of the performance of trading models. By introducing the spread and detailed formula derivations, the author hopes to provide a reliable tool for traders and researchers to help them better understand and optimize trading strategies.