Estimated outcome probability. A likely winner can still be overpriced.
← Decision briefsBoxing · Sep 12, 2:00 PM PDT FORECASTAndres Cortes RELIABILITY4/4 model reports MARKET VALUENot established
Mark Magsayo vs Andres Cortes
51.7%Andres Cortes
LOCKED Sep 12, 9:29 AM PDTAll four probabilities, converted to the same fighter, average 51.73% for Andres Cortes. 2/4 models select this fighter. Low confidence; the house still publishes a winner forecast.
15 reported evidence limits. Agreement alone does not establish accuracy.
A probability difference alone does not account for forecast error, fees, spread, or available liquidity.
Why the models lean this way
These are reasons recorded by the models. They are not independently verified facts or measured causes of a probability change.
- Claude: Scheduled 10 rounds at lightweight; both inside limit (134.5 vs 135)
- Grok: No verified records or histories available
- Gemini: Scheduled for 10 rounds
- GPT: Only verified facts are the fighters' names, 10 scheduled rounds, lightweight division, and official weights of 134.5 and 135.
Read all four model reports
Claude · Andres Cortes · 57.0%
Evidence is limited to a weigh-in line: 10 rounds at lightweight, Magsayo 134.5, Cortes 135. No records, opponent quality, recent form, or style data are supplied, so this is near a coin flip. The only directional cue available is that both made weight comfortably and the bout is a 10-round non-title matchup, which does not establish skill advantage. I lean marginally to Cortes as the naturally larger-listed man at the division limit while Magsayo is a fighter campaigning up from lower weight per the 134.5 figure, but this is a weak inference. Confidence is low; probability kept close to even.
anthropic/claude-opus-5 · Sep 12, 9:29 AM PDTGrok · Mark Magsayo · 50.0%
No records, opponent histories, recent form, styles or past results are supplied in the evidence. Only bout details and official weights are available, providing no basis to favor either fighter. Probability set at minimum due to complete absence of comparative data.
xai/grok-4.3 · Sep 12, 9:20 AM PDTGemini · Mark Magsayo · 50.1%
There is no supplied evidence regarding the fighters' records, past opponent quality, recent form, or styles. Only the scheduled rounds and weigh-in weights are provided, which do not establish a skill advantage. Without historical data, a meaningful skill comparison cannot be made, resulting in a near-even probability.
google/gemini-3.1-pro-preview · Sep 12, 9:20 AM PDTGPT · Andres Cortes · 50.0%
Andres Cortes is selected only as a forced lean in an evidence-sparse matchup. The supplied pre-fight data provides just names, a 10-round lightweight booking, and near-identical official weights, which do not establish a skill edge. With no documented records, opponent quality, recent form, or style information, there is no reliable basis to separate the fighters conditional on a decisive result, so this is effectively a coin-flip.
openai/gpt-5.4 · Sep 12, 9:19 AM PDTWhat remains uncertain
- Only a weigh-in result is sourced; no fight records or histories
- No injury, camp, layoff, or judging/venue information
- No opponent-adjusted performance data to compare levels
- Weight and round count alone cannot establish a skill edge
- No verified injury, camp, judges, or detailed opponent-adjusted history feed is connected
- Round start times are not fixed
- All forecasts use the card's prelim start as the cutoff
- Winner probabilities conditional on decisive result only
What to watch next
Official result recorded. Review the original forecast and sources below.
Open the event and source context →What changed before the lock
- Probability unchanged
The pre-event probability state was locked atomically and is immutable.
Record hash
09d9ddf3f1be0d344fc5c6ea9a93e72e68da34a3e9b458881b06c13aef58d181 - First recorded forecast
First prospective probability state captured for this source and competition.
Record hash
866c31c7233ea1d3f8693c398758b3db5973d38056135cd24e35e571f8fa1034
Reasons are shown as they were recorded. The ledger does not establish that a specific fact caused a probability change.
The saved method comparison
This event has no eligible record in the new comparison cohort. Historical or late forecasts are not filled in after the event.
Inspect the methods and progress →Available source context
These links provide event context. They do not verify every claim in a model report.
Recorded outcome: Andres Cortes defeats Mark Magsayo via unanimous decision (97-93, 97-93, 97-93)
Finish your review
Record whether this brief helped, how long the review took, and your next action.
Use this brief for a review →