Reconcile entries in a session history
This Vox casino guide explains a practical way to read and document what a game or account screen shows. A history can show one event at different levels. Read the identifier and the record type before totaling amounts. A summary may already include detail rows. Similar amounts and nearby timestamps are clues to compare, but do not establish that entries are duplicates.

Worked example
A fictional summary for event J40 contains 21 credits, with related detail rows of 11 and 10. If the summary already includes both details, use 21 once. A separate event J41 worth 21 credits belongs on its own line even though the amount is identical.
What to record
Use a row for each unique event and an indented note for its details. Compare timezones and units before reading the total. If a relationship is unclear, mark it as unconfirmed and ask the account service to verify its own record.
Practice cases
Each case changes the view or the information available. Use the worked example as a starting point, then give the case its own record. Start with the stated condition. The review prompts are there to check your notes before you use them to explain a real issue.
1. A summary with a parent identifier
In this hypothetical case, look at “A summary with a parent identifier” and suppose that the record omits the relationship label. Your job is to write a note that can be checked later. Decide what the screen supplies directly and what you need from a second source before answering.
Compare two records. Create two rows with the same column headings. Add the time or version to each row before you enter any values. Put the records side by side and circle the field that changes the interpretation. A name or number that matches can help you locate a record, but confirm what each row actually represents.
Use a row for each unique event and an indented note for its details. Use that check here, then name the exact part of the example that supports your reading. Record whether the condition affected the result, the display, or the strength of the evidence.
As a final check, ask whether you can explain why these rows belong together without using the artwork or a similar-looking label as proof. If you cannot, describe the two observations separately and identify the field you still need.
2. A summary with a parent identifier
In this hypothetical case, look at “A summary with a parent identifier” and suppose that one screen shows a wider date range. Your job is to write a note that can be checked later. Decide what the screen supplies directly and what you need from a second source before answering.
Identify the missing evidence. Split the observation into what the screen confirms, what you can reasonably infer from the rules, and what remains open. Choose a single missing field that would change your answer. Explain where you would look for it without treating the current display as more complete than it is.
Compare timezones and units before reading the total. Use that check here, then name the exact part of the example that supports your reading. Record whether the condition affected the result, the display, or the strength of the evidence.
If a more complete view appears, add a separate dated note instead of rewriting the earlier impression. That provides a record of why the answer changed. A useful review should make the earlier uncertainty and the later evidence understandable.
3. A summary with a parent identifier
In this hypothetical case, look at “A summary with a parent identifier” and suppose that the column uses a different unit. Your job is to write a note that can be checked later. Decide what the screen supplies directly and what you need from a second source before answering.
Compare two records. Create two rows with the same column headings. Add the time or version to each row before you enter any values. Put the records side by side and circle the field that changes the interpretation. A name or number that matches can help you locate a record, but confirm what each row actually represents.
If a relationship is unclear, mark it as unconfirmed and ask the account service to verify its own record. Use that check here, then name the exact part of the example that supports your reading. Record whether the condition affected the result, the display, or the strength of the evidence.
As a final check, ask whether you can explain why these rows belong together without using the artwork or a similar-looking label as proof. If you cannot, describe the two observations separately and identify the field you still need.
4. A row with a detail identifier
In this hypothetical case, look at “A row with a detail identifier” and suppose that the record omits the relationship label. Your job is to write a note that can be checked later. Decide what the screen supplies directly and what you need from a second source before answering.
Write a note someone else can reproduce. Start the note at the point a second reader would see on the screen. List the route to that view, the exact caption, and the first visible change. A note about what happened should stay separate from your guess about why it happened. Keep the details that let somebody else find the same record.
Compare timezones and units before reading the total. Use that check here, then name the exact part of the example that supports your reading. Record whether the condition affected the result, the display, or the strength of the evidence.
Read your note back without the original screenshot. Replace vague references such as “here” and “that amount” with the heading or identifier. If the label cannot be recovered, mark it as unavailable rather than assigning it a likely value.
5. A row with a detail identifier
In this hypothetical case, look at “A row with a detail identifier” and suppose that one screen shows a wider date range. Your job is to write a note that can be checked later. Decide what the screen supplies directly and what you need from a second source before answering.
Review the sequence. Write the events in order using the times shown. Give each entry a label that tells the reader whether it is a rule, a status, a result, or a later correction. If two sources update at different times, keep both timestamps with their own entries. Avoid filling in a missing step based on an earlier example.
If a relationship is unclear, mark it as unconfirmed and ask the account service to verify its own record. Use that check here, then name the exact part of the example that supports your reading. Record whether the condition affected the result, the display, or the strength of the evidence.
Before drawing a conclusion, check whether changing the order would change your reading of the event. If it would, identify the timestamp or reference you need to verify first. Keep the original record available while checking.
6. A row with a detail identifier
In this hypothetical case, look at “A row with a detail identifier” and suppose that the session status changes before the export. Your job is to write a note that can be checked later. Decide what the screen supplies directly and what you need from a second source before answering.
Review the sequence. Write the events in order using the times shown. Give each entry a label that tells the reader whether it is a rule, a status, a result, or a later correction. If two sources update at different times, keep both timestamps with their own entries. Avoid filling in a missing step based on an earlier example.
Compare timezones and units before reading the total. Use that check here, then name the exact part of the example that supports your reading. Record whether the condition affected the result, the display, or the strength of the evidence.
Before drawing a conclusion, check whether changing the order would change your reading of the event. If it would, identify the timestamp or reference you need to verify first. Keep the original record available while checking.
7. Two entries with the same amount
In this hypothetical case, look at “Two entries with the same amount” and suppose that the record omits the relationship label. Your job is to write a note that can be checked later. Decide what the screen supplies directly and what you need from a second source before answering.
Identify the missing evidence. Split the observation into what the screen confirms, what you can reasonably infer from the rules, and what remains open. Choose a single missing field that would change your answer. Explain where you would look for it without treating the current display as more complete than it is.
If a relationship is unclear, mark it as unconfirmed and ask the account service to verify its own record. Use that check here, then name the exact part of the example that supports your reading. Record whether the condition affected the result, the display, or the strength of the evidence.
If a more complete view appears, add a separate dated note instead of rewriting the earlier impression. That provides a record of why the answer changed. A useful review should make the earlier uncertainty and the later evidence understandable.
8. Two entries with the same amount
In this hypothetical case, look at “Two entries with the same amount” and suppose that the column uses a different unit. Your job is to write a note that can be checked later. Decide what the screen supplies directly and what you need from a second source before answering.
Identify the missing evidence. Split the observation into what the screen confirms, what you can reasonably infer from the rules, and what remains open. Choose a single missing field that would change your answer. Explain where you would look for it without treating the current display as more complete than it is.
Compare timezones and units before reading the total. Use that check here, then name the exact part of the example that supports your reading. Record whether the condition affected the result, the display, or the strength of the evidence.
If a more complete view appears, add a separate dated note instead of rewriting the earlier impression. That provides a record of why the answer changed. A useful review should make the earlier uncertainty and the later evidence understandable.
9. Two entries with the same amount
In this hypothetical case, look at “Two entries with the same amount” and suppose that the session status changes before the export. Your job is to write a note that can be checked later. Decide what the screen supplies directly and what you need from a second source before answering.
Compare two records. Create two rows with the same column headings. Add the time or version to each row before you enter any values. Put the records side by side and circle the field that changes the interpretation. A name or number that matches can help you locate a record, but confirm what each row actually represents.
If a relationship is unclear, mark it as unconfirmed and ask the account service to verify its own record. Use that check here, then name the exact part of the example that supports your reading. Record whether the condition affected the result, the display, or the strength of the evidence.
As a final check, ask whether you can explain why these rows belong together without using the artwork or a similar-looking label as proof. If you cannot, describe the two observations separately and identify the field you still need.
10. An adjustment recorded the next day
In this hypothetical case, look at “An adjustment recorded the next day” and suppose that one screen shows a wider date range. Your job is to write a note that can be checked later. Decide what the screen supplies directly and what you need from a second source before answering.
Write a note someone else can reproduce. Start the note at the point a second reader would see on the screen. List the route to that view, the exact caption, and the first visible change. A note about what happened should stay separate from your guess about why it happened. Keep the details that let somebody else find the same record.
Compare timezones and units before reading the total. Use that check here, then name the exact part of the example that supports your reading. Record whether the condition affected the result, the display, or the strength of the evidence.
Read your note back without the original screenshot. Replace vague references such as “here” and “that amount” with the heading or identifier. If the label cannot be recovered, mark it as unavailable rather than assigning it a likely value.
11. An adjustment recorded the next day
In this hypothetical case, look at “An adjustment recorded the next day” and suppose that the column uses a different unit. Your job is to write a note that can be checked later. Decide what the screen supplies directly and what you need from a second source before answering.
Review the sequence. Write the events in order using the times shown. Give each entry a label that tells the reader whether it is a rule, a status, a result, or a later correction. If two sources update at different times, keep both timestamps with their own entries. Avoid filling in a missing step based on an earlier example.
If a relationship is unclear, mark it as unconfirmed and ask the account service to verify its own record. Use that check here, then name the exact part of the example that supports your reading. Record whether the condition affected the result, the display, or the strength of the evidence.
Before drawing a conclusion, check whether changing the order would change your reading of the event. If it would, identify the timestamp or reference you need to verify first. Keep the original record available while checking.
12. An adjustment recorded the next day
In this hypothetical case, look at “An adjustment recorded the next day” and suppose that the session status changes before the export. Your job is to write a note that can be checked later. Decide what the screen supplies directly and what you need from a second source before answering.
Write a note someone else can reproduce. Start the note at the point a second reader would see on the screen. List the route to that view, the exact caption, and the first visible change. A note about what happened should stay separate from your guess about why it happened. Keep the details that let somebody else find the same record.
Use a row for each unique event and an indented note for its details. Use that check here, then name the exact part of the example that supports your reading. Record whether the condition affected the result, the display, or the strength of the evidence.
Read your note back without the original screenshot. Replace vague references such as “here” and “that amount” with the heading or identifier. If the label cannot be recovered, mark it as unavailable rather than assigning it a likely value.
Related guide: Document a disconnected session
