Closing line value gets ruined long before anyone runs the calculation. The damage happens when the bet is placed and the accepted price is not saved. By the time the market closes, the screen has changed and memory has promoted a near-miss into an exact record.
Week 1 is the clean place to stop that habit. Start with a log that preserves the wager as executed, then append the closing reference later. You do not need a glamorous dashboard. You need evidence that survives a bad Sunday.
Define the record before the season
CLV compares an accepted entry with a later reference for the same market. The entry is not the number you considered, the number shown before clicking, or the best number another book displayed. It is the term the venue accepted.
The closing side needs a rule too. Choose the reference source and matching method before seeing results. A designated book, a declared consensus, or another reliable feed can work. Switching references per wager cannot.
The closing line value explainer covers the concept. The rest of this checklist is about preserving enough evidence to use it honestly.
Log the accepted ticket
Capture these fields from the confirmation:
- Event: the game or contest attached to the wager.
- Market: spread, total, moneyline, prop, or exact derivative.
- Selection: the side of the market you accepted.
- Line or threshold: the market term that defines the outcome.
- Price: the accepted odds or contract price.
- Venue: the book, exchange, or platform.
- Timestamp: acceptance time under one documented timezone.
- Decision source: model version, rule, or manual rationale.
Those fields identify the bet after the lobby has moved on. A screenshot can support the record, but the structured fields are what make matching and review possible.
Keep line and price separate
A spread can remain on the same number while its attached price changes. A prop can move its threshold, price, or both. If the log collapses those terms into one text field, the closing comparison becomes guesswork.
Store the market definition and the price independently. That lets the matcher reject a different threshold instead of pretending the two tickets are equivalent.
Use the confirmation, not the intention
Markets can reprice between click and acceptance. The confirmation controls. When a venue asks you to accept changed terms, save the final response, not the stale screen that made you click.
Make this part of the bet routine: place, confirm, capture, then move on. A later memory of what “should have” been accepted is not a field value.
Match the exact closing market
A full-game spread closes against the full-game spread. A first-half total closes against the first-half total. A player prop must match the player, direction, threshold, and relevant participation rule.
When the exact market disappears or closes under a materially different definition, mark CLV unavailable. Keep the wager and settlement in the log. Do not borrow a nearby market because an empty cell looks untidy.
Freeze the closing-source rule
Write down which venue or consensus counts as the close, how stale quotes are excluded, how vig is removed, and how ties between sources are handled. Version the rule when it changes.
If the source changes midseason, split the reporting window. A method change is not a footnote to bury inside one average; it changes what the series measures.
Keep result and price quality apart
A wager can beat the close and lose. It can trail the close and win. The log needs separate fields for CLV and settlement so one does not overwrite the other.
Preserve the original entry after the game. Append the close, settlement, and grading time. Corrections should be auditable rather than silently replacing the past.
Represent voids and missing data honestly
A void is not a win or a loss. A push is not a missing result. An unmatched closing market is not neutral CLV. Make those states explicit.
Do not delete awkward rows from the season summary. Missingness can reveal that one market is poorly supported, one venue is difficult to capture, or one branch of the workflow needs a new source.
Attach context without hiding structure
Notes can record injury news, weather, model changes, or why you acted. Keep notes separate from the official market fields. Another person should be able to identify the wager without interpreting your prose.
Link the model version or Studio run when a model produced the candidate. The Studio can keep model context near the decision, but the accepted ticket still controls the entry price.
What the Number Says
At the weekly review, start with missing prices, missing timestamps, duplicate entries, unmatched closes, and ambiguous markets. Fix verifiable mistakes while the source evidence is still available.
Only then summarize CLV or the graded record. An attractive average built from easy-to-match winners is worse than an ugly complete log.
Make the evidence portable
The core record should survive a change in app, book, or dashboard. Export the structured fields and preserve source captures under stable identifiers. The accepted ticket is evidence; the CLV calculation is analysis derived from that evidence.
Keeping both lets you correct a calculation without rewriting the wager. It also makes disagreements about the closing reference easier to audit.
The pre-kickoff checklist
- The event and market are exact.
- The selection is unambiguous.
- The accepted line or threshold is saved.
- The accepted price is saved.
- The venue and timestamp are saved.
- The model or decision source is identified.
- The closing source and devig rule are documented.
- Void, push, and unavailable states exist.
- The log can be exported independently of the interface.
When any field is missing, return the specific cure. “Add the accepted timestamp” is useful. Silently filling it from game time is not.
Bottom line
The Week 1 CLV habit is simple: save the bet as accepted, define the close in advance, and never let the final score edit either side. A complete season begins with one complete row.
Use the concept guide for the math and Studio for model context. The log’s job is more basic and more important: make every later comparison reproducible.
