Using ChatGPT to fill out a bracket...

[dahliaclone]

I saw CNET did this and was curious. I'm in a $5 bracket with some friends and family and asked ChatGPT to give me their picks and scores for the entire tournament with some well placed upsets and gonna go with exactly what they have and see how it ends up doing...some of these I don't buy but it's just an exercise so we'll see.

Some noteworthy picks by AI:

• Akron over Arizona by 4
• Drake over Missouri by 2
• Texas Tech over UNCW but only by 1
• Purdue over High Point by 25 (weird but whatever)
• Michigan over UCSD by 15 (popular upset pick)
• San Diego State over Ol' Miss by 10
• Marquette over New Mexico by 14
• St. Mary's beating Alabama by 2 in round of 32
• Memphis over Maryland in round of 32
• Drake over TT by 3 in round of 32
• Kansas over St. John's by 12 in round of 32
• ISU over SDSU by 3 in round of 32
• Marquette over Michigan State by 6 in round of 32
• Kansas over Drake by a lot in Sweet 16
• ISU over Marquette by 2 in Sweet 16
• Auburn over ISU by 4 in Elite 8
• FF: Auburn/Florida/Duke/Houston which is all #1 seeds but whatever
• Florida over Houston for title

 

[EnkAMania]

I asked ChatGPT to rank advanced statistics, here's what they gave me.

Tier 1 – Most Crucial​

1. Adj. Score Margin – The best single indicator of overall team strength, combining offense and defense efficiency. Elite teams often dominate this category.
2. Defense Rating – Defense wins championships. NCAA tournament games tend to be lower-scoring, making elite defense a game-changer.
3. Effective FG% (eFG%) – Accounts for the extra value of three-pointers, one of the strongest predictors of offensive success.
4. Turnover % (Offensive) – Taking care of the ball is crucial. Teams that limit turnovers give themselves more chances to score.
5. Opp. Turnover % – Generating turnovers leads to extra possessions and transition opportunities, a significant factor in March.

Tier 2 – Strong Contributors​

1. Offense Rating – Overall offensive efficiency is important but less predictive than defensive dominance.
2. Opp. True Shooting % – Defensive shot quality matters. Teams that force tough shots often advance far.
3. Turnover Margin – The difference between forcing and committing turnovers often decides close tournament games.
4. True Shooting % (TS%) – A more comprehensive version of eFG% that includes free throws, though not as predictive as eFG%.
5. Rebound % – Controlling the glass limits second-chance points and extends possessions.
6. Off. Rebound % – Second-chance points can swing tight games, especially against elite defenses.

Tier 3 – Situational Factors​

1. Assists / Turnover Ratio – Helps evaluate offensive flow and decision-making but is secondary to actual efficiency metrics.
2. Assist % – Ball movement is valuable, but teams can win in different ways (e.g., iso-heavy teams like 2024 UConn).
3. FT / FGA – Getting to the free-throw line is helpful, but efficiency matters more than volume.
4. Opp. FT / FGA – Fouling too much can be costly, but it's not as predictive as shot quality defense.
5. Pace – Less of a factor in tournament success; both fast and slow teams have won titles.

 

[Gonzo]

Ok but how many times did it use the word "delve"?

 

[bosco]

I asked ChatGPT to rank advanced statistics, here's what they gave me.

Tier 1 – Most Crucial​

1. Adj. Score Margin – The best single indicator of overall team strength, combining offense and defense efficiency. Elite teams often dominate this category.
2. Defense Rating – Defense wins championships. NCAA tournament games tend to be lower-scoring, making elite defense a game-changer.
3. Effective FG% (eFG%) – Accounts for the extra value of three-pointers, one of the strongest predictors of offensive success.
4. Turnover % (Offensive) – Taking care of the ball is crucial. Teams that limit turnovers give themselves more chances to score.
5. Opp. Turnover % – Generating turnovers leads to extra possessions and transition opportunities, a significant factor in March.

Tier 2 – Strong Contributors​

1. Offense Rating – Overall offensive efficiency is important but less predictive than defensive dominance.
2. Opp. True Shooting % – Defensive shot quality matters. Teams that force tough shots often advance far.
3. Turnover Margin – The difference between forcing and committing turnovers often decides close tournament games.
4. True Shooting % (TS%) – A more comprehensive version of eFG% that includes free throws, though not as predictive as eFG%.
5. Rebound % – Controlling the glass limits second-chance points and extends possessions.
6. Off. Rebound % – Second-chance points can swing tight games, especially against elite defenses.

Tier 3 – Situational Factors​

1. Assists / Turnover Ratio – Helps evaluate offensive flow and decision-making but is secondary to actual efficiency metrics.
2. Assist % – Ball movement is valuable, but teams can win in different ways (e.g., iso-heavy teams like 2024 UConn).
3. FT / FGA – Getting to the free-throw line is helpful, but efficiency matters more than volume.
4. Opp. FT / FGA – Fouling too much can be costly, but it's not as predictive as shot quality defense.
5. Pace – Less of a factor in tournament success; both fast and slow teams have won titles.

Ok..well we are waiting!?

 

[cedarstrip]

If I ask chatGPT to fill out a winning bracket and 10,000 other people ask that too do we all get the same bracket?

 

[2122]

If I ask chatGPT to fill out a winning bracket and 10,000 other people ask that too do we all get the same bracket?

Perhaps chatGPT's palms might be greased a bit?

 

[NATEizKING]

I'm still waiting for chatgpt to tell me how to make the best grilled cheese sandwich, hasn't beaten my mom's.

 

[dahliaclone]

If I ask chatGPT to fill out a winning bracket and 10,000 other people ask that too do we all get the same bracket?

Depends on what you ask it. CNET I think did three where the prompt was different each time. One normal. One with a few upsets. Then one with heavy upsets and a smart low seeded team to make a run.

I will say I did my prompt a few times and everything was the same including score predictions.

 

[nfrine]

There is a reason it's called artificial intelligence.

 

[Dgilbertson]

I asked ChatGPT to rank advanced statistics, here's what they gave me.

Tier 1 – Most Crucial​

1. Adj. Score Margin – The best single indicator of overall team strength, combining offense and defense efficiency. Elite teams often dominate this category.
2. Defense Rating – Defense wins championships. NCAA tournament games tend to be lower-scoring, making elite defense a game-changer.
3. Effective FG% (eFG%) – Accounts for the extra value of three-pointers, one of the strongest predictors of offensive success.
4. Turnover % (Offensive) – Taking care of the ball is crucial. Teams that limit turnovers give themselves more chances to score.
5. Opp. Turnover % – Generating turnovers leads to extra possessions and transition opportunities, a significant factor in March.

Tier 2 – Strong Contributors​

1. Offense Rating – Overall offensive efficiency is important but less predictive than defensive dominance.
2. Opp. True Shooting % – Defensive shot quality matters. Teams that force tough shots often advance far.
3. Turnover Margin – The difference between forcing and committing turnovers often decides close tournament games.
4. True Shooting % (TS%) – A more comprehensive version of eFG% that includes free throws, though not as predictive as eFG%.
5. Rebound % – Controlling the glass limits second-chance points and extends possessions.
6. Off. Rebound % – Second-chance points can swing tight games, especially against elite defenses.

Tier 3 – Situational Factors​

1. Assists / Turnover Ratio – Helps evaluate offensive flow and decision-making but is secondary to actual efficiency metrics.
2. Assist % – Ball movement is valuable, but teams can win in different ways (e.g., iso-heavy teams like 2024 UConn).
3. FT / FGA – Getting to the free-throw line is helpful, but efficiency matters more than volume.
4. Opp. FT / FGA – Fouling too much can be costly, but it's not as predictive as shot quality defense.
5. Pace – Less of a factor in tournament success; both fast and slow teams have won titles.

Looked back at a previous 3/14 upset and the only category the 3 team that lost was not better in was offensive rebounds and rebound %.

 

[CoachHines3]

I asked ChatGPT to rank advanced statistics, here's what they gave me.

Tier 1 – Most Crucial​

1. Adj. Score Margin – The best single indicator of overall team strength, combining offense and defense efficiency. Elite teams often dominate this category.
2. Defense Rating – Defense wins championships. NCAA tournament games tend to be lower-scoring, making elite defense a game-changer.
3. Effective FG% (eFG%) – Accounts for the extra value of three-pointers, one of the strongest predictors of offensive success.
4. Turnover % (Offensive) – Taking care of the ball is crucial. Teams that limit turnovers give themselves more chances to score.
5. Opp. Turnover % – Generating turnovers leads to extra possessions and transition opportunities, a significant factor in March.

Tier 2 – Strong Contributors​

1. Offense Rating – Overall offensive efficiency is important but less predictive than defensive dominance.
2. Opp. True Shooting % – Defensive shot quality matters. Teams that force tough shots often advance far.
3. Turnover Margin – The difference between forcing and committing turnovers often decides close tournament games.
4. True Shooting % (TS%) – A more comprehensive version of eFG% that includes free throws, though not as predictive as eFG%.
5. Rebound % – Controlling the glass limits second-chance points and extends possessions.
6. Off. Rebound % – Second-chance points can swing tight games, especially against elite defenses.

Tier 3 – Situational Factors​

1. Assists / Turnover Ratio – Helps evaluate offensive flow and decision-making but is secondary to actual efficiency metrics.
2. Assist % – Ball movement is valuable, but teams can win in different ways (e.g., iso-heavy teams like 2024 UConn).
3. FT / FGA – Getting to the free-throw line is helpful, but efficiency matters more than volume.
4. Opp. FT / FGA – Fouling too much can be costly, but it's not as predictive as shot quality defense.
5. Pace – Less of a factor in tournament success; both fast and slow teams have won titles.

I wanted to see what Iowa States #s on this were..​

​

Tier 1 – Most Crucial​

1. Adj. Score Margin – 11th at 12.3 pts.
2. Defense Rating – 9th at 92.4 (KenPom)
3. Effective FG% (eFG%) – 50th at 54.6% (KenPom)
4. Turnover % (Offensive) – 162nd at 16.9% (KenPom)
5. Opp. Turnover % – 14th at 21.8% (KenPom

Tier 2 – Strong Contributors​

1. Offense Rating – 20th at 119.4 (KenPom)
2. Opp. True Shooting % – 69th at 104.2% (TeamRankings)
3. Turnover Margin – 22nd at 0.6% (Team Rankings)
4. True Shooting % (TS%) – 29th at 116.9% (Team Rankings)
5. Rebound % – 51st at 52.7% (Team Rankings)
6. Off. Rebound % – 110th at 32.2% (KenPom)

Tier 3 – Situational Factors​

1. Assists / Turnover Ratio – 102nd at 1.262
2. Assist % –
3. FT / FGA –
4. Opp. FT / FGA –
5. Pace – 102nd at 68.6 (KenPom)

had trouble with 3 of the last ones and finding them. Assist % was only giving me assists per game.

 

[AirWalke]

I did this last year and my AI bracket was kind of mid. It did better than my own personal bracket, but it definitely did not win the office pool. Just goes to show how much of a crap shoot the tourney can actually be.

 

[EnkAMania]

I did this last year and my AI bracket was kind of mid. It did better than my own personal bracket, but it definitely did not win the office pool. Just goes to show how much of a crap shoot the tourney can actually be.

I let the robot talk me into UC San Diego and St. Mary's having a big run. I just can't pick Rick Patino . Here is AI assisted bracket