HAWK - Correlation

Correlation

The Correlation has been divided into five parts -

PnL Correlation - Calculates the correlation between the Equity PnL and the Settle Price of the contract.
PnL Diff Correlation - Shows the correlation between Day PnL and the daily price change of a contract, calculated as the difference between today's settlement price and the previous day's settlement price.
Portfolio Correlation - Calculates the correlation between the selected portfolio price and the Settle Price of the contract.
Portfolio Diff Correlation - Shows the correlation between the daily change in portfolio price and the daily price change of the contract, calculated as the difference between today's settlement price and the previous day's settlement price.

For Portfolio Correlation and Portfolio Diff Correlation, the portfolio price is calculated using the portfolio as of the date selected in the date selection header on the top right, and is assumed to be kept the same for the duration selected

PnL Correlation

PnL Correlation displays the correlation between the Equity PnL and the Settle Price of the contracts listed in the table. Along with it we also have a filter which can be used to see absolute values above and below the set value

The PnL Correlation section helps analyze the relationship between Equity PnL and the Settle Price of selected contracts.

Key Features:

Use the filter button to display only correlations above a specified threshold.
The table organizes data by Product and Delivery, with options to select different structures like Outright, 1-month spreads, etc.
The table can display up to 5 products at a time, and users can select products from a dropdown menu. This dropdown only filters the table display and does not impact the correlation calculation.
Each table cell represents the correlation value between Equity PnL Exc R and the settle price of the corresponding contract and structure.
The highest correlation value is highlighted to easily identify the structure most closely linked to Equity PnL.

Clicking on any cell in the table opens a sidebar displaying detailed graphs that visually explain the correlation values, helping users better understand the relationship between Equity PnL and Settle Price for the selected contract and structure.

The Portfolio Graphs display the trend of the settle price for the selected contract, along with the Day PnL and Equity PnL Exc R in the bottom graphs. These visualizations provide a graphical representation of the data used to calculate the correlation, making it easier to interpret the relationship between PnL and settle price movements.

The bottom graph in the side panel has two tabs: Close Price and Difference Close Price, each providing insights into the relationship between PnL and contract price movements.

Close Price

This scatter plot shows the relationship between Equity PnL and the Settle Price of the selected structure.
A line of best fit is drawn, and hovering over it displays alpha and beta values based on the equation: Y = alpha * X + beta

Difference Close Price

This scatter plot compares Day PnL with the daily price change of a contract.
It includes a line of best fit along with alpha, beta, and R-squared (R²) values:
1. Alpha: How much Day PnL changes for every 1-unit change in price.
2. Beta: The base PnL when there’s no price movement, often indicating scalping behavior.
3. R² (R-Squared): Indicates how well the price change explains the variation in Day PnL (a higher value means a stronger correlation).

PnL Diff Correlation

PnL Diff Correlation displays the correlation between the Day PnL Exc R and the change in Price of the contracts listed in the table. Along with it we also have a filter which can be used to see absolute values above and below the set value

The PnL Correlation section helps analyze the relationship between Day PnL and the change in Settle Price of selected contracts.

Rest of the functionalities and widgets remain same as the PnL Correlation tab

The PnL Diff Correlation section analyzes the relationship between Day PnL Exc R and the daily price change of the selected contracts.

A filter option allows users to display only correlation values that fall above or below a specified threshold, helping to focus on significant relationships.
This section functions similarly to the PnL Correlation tab, maintaining the same interactive widgets and features.

By examining how Day PnL moves with price fluctuations, traders can identify key patterns and refine their strategies effectively.

Portfolio Correlation

This section analyzes the relationship between the selected portfolio price and contract prices using correlation metrics.

The portfolio date to be analyzed can be selected from the date pane in the top right corner of the table.
A filter option allows users to display only the absolute correlation values above or below a specified threshold.

Understanding the SOD Price Calculation

One key aspect of this analysis is how the Start-of-Day (SOD) Price is calculated, as it plays a crucial role in determining the correlation values.

Example:

Imagine you select February 20 as the portfolio date, and the application's date range is from February 1 to February 19.

To compute the SOD Price series, we assume:

The SOD of February 20 remains unchanged throughout the period from February 1 to February 19.
The trader does not execute any trades during this period.

Using this assumption, we generate a PnL trend based on how the portfolio would have performed if no trades were made, using the same SOD price for the entire historical period. This PnL trend is then used to compute the correlation with the contract prices.

This method helps in assessing how much the selected portfolio’s value is influenced by market movements over time, independent of trading activity.

Clicking on any cell in the table opens a sidebar displaying detailed graphs that visually explain the correlation values, helping users better understand the relationship between SOD Price and Settle Price for the selected contract and structure.

The Portfolio Graphs display the trend of the settle price for the selected contract, along with the portfolio price of the said date in the bottom graphs. These visualizations provide a graphical representation of the data used to calculate the correlation, making it easier to interpret the relationship between SOD Price and settle price movements.

The bottom graph in the side panel has three tabs: Close Price and Difference Close Price, each providing insights into the relationship between PnL and contract price movements.

Portfolio Scatter:

This scatter plot shows the relationship between SOD Price and the Settle Price of the selected structure.
A line of best fit is drawn, and hovering over it displays alpha and beta values based on the equation: Y = alpha * X + beta

Port Diff Price:

This scatter plot compares Day SOD PnL calculated from the portfolio selected with the daily price change of a contract.
It includes a line of best fit along with alpha, beta, and R-squared (R²) values:
- Alpha: How much Day SOD PnL changes for every 1-unit change in price.
- Beta: The base PnL when there’s no price movement, often indicating scalping behavior in the same contract
- R² (R-Squared): Indicates how well the price change explains the variation in DaySOD PnL (a higher value means a stronger correlation).

Rolling Portfolio Correlation

Rolling Portfolio Correlation extends the standard portfolio correlation analysis by calculating the correlation of each day’s Start-of-Day (SOD) price with the settle price of the selected contract over a rolling period.

Instead of computing correlation for just one selected portfolio date, this approach calculates correlations for all dates in the selected date range.
Users can specify a rolling window, which determines how many past business days are considered when calculating the correlation.
The results are plotted against each date in the selected timeframe, showing how the correlation evolves over time.

Example:

Let’s say the selected date range is February 1 to February 20, and the chosen rolling window is 30 business days.

For each date in this range, the process works as follows:

Take the SOD price of that specific date (e.g., February 3).
Assume this SOD remains unchanged for the previous 15 business days (going back to January 3).
Compute the PnL trend that would have been generated if no trades were made.
Find the correlation and beta values between this simulated PnL trend and the settle price of the selected contract.
Plot the correlation result against February 3 on the graph.

This process is repeated for each date in the selected range, giving a rolling view of how the portfolio’s correlation with the market evolves over time.

Portfolio Diff Correlation

This section analyzes the relationship between the daily change in portfolio price and the daily change in contract prices using correlation metrics.

The portfolio date to be analyzed can be selected from the date pane in the top right corner of the table.
A filter option allows users to display only the absolute correlation values above or below a specified threshold.

Understanding the SOD Price Calculation

One key aspect of this analysis is how the Start-of-Day (SOD) Price is calculated, as it plays a crucial role in determining the correlation values.

Example:

Imagine you select February 20 as the portfolio date, and the application's date range is from February 1 to February 19.

To compute the SOD Price series, we assume:

The SOD of February 20 remains unchanged throughout the period from February 1 to February 19.
The trader does not execute any trades during this period.

This method helps in assessing how much the selected portfolio’s value is influenced by market movements over time, independent of trading activity.

Then similar to Portfolio Correlation we have the table with the correlation values which opens the Portfolio graphs for further analysis

Custom Correlation

This section helps further deeply analyse the desired correlation in one place which means that any type of correlation with the custom portfolio analysis can be performed here. We can reach the tab by selecting the custom correlation from the tabs at the top of page

With the help of the view dropdown we can select the desired type of correlation of which we want to do the deeper analysis

After selecting the correlation type we need to navigate to the respective correlation tab to select the product contract structure combination which we want to further analyse and return to the same tab which will have the updated graphs found on the right side of the page

After having the graph we can then proceed to open the actual place to enter custom portfolio by clicking on the "Try it here" button. To select a certain date SOD Position for a quick start click on the portfolio graph for teh respective date as it will populate the SOD data in the section opened.

If you want to start fresh or have all the values removed then we can use the "Reset" button for the same and the table gets reset and allows users to enter values manually. Also for better user experience users can traverse the table with the help of the arrow keys wherever the values are allowed to be changed.

On clicking the calculate button we get the desired correlation plots for the portfolio entered.

Based on the type of view selected at the top everything on the page changes so user doesnt need to navigate back to select the value for populating the graphs

If the selected view is PnL Correlation or PnL Diff Correlation, then PnL Correlation, PnL Diff Correlation and Rolling PnL Correlation types of charts of populated and the meaning of the same have been explained below
PnL Correlation (Equity PnL vs. Settle Price)

This graph shows the trend of Equity PnL alongside the Settle Price of the customized structure.
Since different structures of products have varying tick sizes, we adjust the price values by their dollar values for consistency. This may result in larger numbers on the axis, but the trend remains accurately represented.

PnL Diff Correlation (Day PnL vs. Daily Price Change)

This graph compares Day PnL with the daily price change of the customized structure.
It also displays alpha, beta, and R-squared values, helping to understand the relationship between price movements and PnL generation.

Rolling PnL Correlation

This graph calculates the correlation between Equity PnL and Settle Price over a rolling period.
For example, if the selected timeframe is Feb 20, 2025 – Feb 27, 2025, and the rolling window is 10 days, the correlation for each date is computed by looking back 10 business days.
For Feb 20, the calculation would use data from Feb 7 to Feb 20 to determine the correlation coefficient.

If the selected view is Portfolio Correlation or Portfolio Diff Correlation, then Portfolio Correlation, Portfolio Diff Correlation, and Rolling Portfolio Correlation populate and the meaning of the same are :

Portfolio Correlation (SOD Price vs. Custom Portfolio Price)

This graph shows the trend of SOD Price alongside the Settle Price of the customized structure.
Since different structures of products have varying tick sizes, we adjust the price values by their dollar values for consistency. This may result in larger numbers on the axis, but the trend remains accurately represented.

Portfolio Diff Correlation (Day SOD PnL vs. Daily Price Change)

This graph compares Day SOD PnL with the daily price change of the customized structure.
It also displays alpha, beta, and R-squared values, helping to understand the relationship between price movements and PnL generation.

Rolling Port Correlation

This graph calculates the correlation between SOD Price and Custom Price over a rolling period.
For example, if the selected timeframe is Feb 20, 2025 – Feb 27, 2025, and the rolling window is 10 days, the correlation for each date is computed by looking back 10 business days.
For Feb 20, the calculation would use data from Feb 7 to Feb 20 to determine the correlation coefficient.

For the Portfolio and Portfolio Diff Correlation, we have an option to have another table to populate different portfolio for comparison. This is used to analyzing a single custom structure, users can compare two custom strategies and analyze their positions side by side.

Click on "Try Another Strategy" to add a second table for comparison.
Users can either manually enter positions or populate the table by clicking on a date in the side graphs to load the SOD positions from that day.
After entering both strategies, clicking "Calculate" generates two new graphs to visualize and compare their performance.

Portfolio Correlation (Custom Portfolio1 vs. Custom Portfolio2 Price)

This graph shows the trend of both the customize structure Price.

Rolling Port Correlation

This graph calculates the correlation between both the custom portfolio over a rolling period.
For example, if the selected timeframe is Feb 20, 2025 – Feb 27, 2025, and the rolling window is 10 days, the correlation for each date is computed by looking back 10 business days.
For Feb 20, the calculation would use data from Feb 7 to Feb 20 to determine the correlation coefficient.

Page updated

Google Sites

Report abuse