Is the NCAA tournament a “crapshoot”

[Uncle Rico]

Seriously guys.  I'll allow you each one more reply to each other if you choose.  After that, I've got the banhammer ready.

The tournament is a crapshoot.  Ball knowers know it

[FartyEightHours]

Those are pretty good odds.  I wish I could find those odds in a game at poto

[5DollarPitcher]

The tournament is a crapshoot.  Ball knowers know it

No

(My 1 reply token)

[Hards Alumni]

No

(My 1 reply token)

Incorrect

[FartyEightHours]

This is boring.  Nobody has even defined what the parameters of “crapshoot” is.  It might mean different things to different people.  This is a giant waste of time

[TAMU, Knower of Ball]

This is boring.  Nobody has even defined what the parameters of “crapshoot” is.  It might mean different things to different people.  This is a giant waste of time

Farty for the win!

[DoctorV]

Some would argue it’s crapcrap

[lostpassword]

This is boring.  Nobody has even defined what the parameters of “crapshoot” is.  It might mean different things to different people.  This is a giant waste of time

Agreed, but this got my thinking and I did some math.

Scenario:
Team #1 - head and shoulders the best
Teams #2 thru #64 - equally crummy
Head-to-head, Team #1 will takes 9 of 10 games against any of #2 thru #64.

With 6 rounds:
Single-Elimination tourney: Team #1 wins 53% (upset every other year)
Best of 3: Team #1 wins 84.3% (upset once every 6 years)
Best of 7: Team #1 wins 99.2% (upset once a century)

Team #1 is far and away the favorite - Teams 2 thru 64 each have well under a 1% shot - but it's basically a toss up versus the field with lose and go home.  If I change the scenario to Team #1 wins 19 of 20 instead of 9 out of 10, it's still under 75% likely to win it all.

[5DollarPitcher]

Agreed, but this got my thinking and I did some math.

Scenario:
Team #1 - head and shoulders the best
Teams #2 thru #64 - equally crummy
Head-to-head, Team #1 will takes 9 of 10 games against any of #2 thru #64.

With 6 rounds:
Single-Elimination tourney: Team #1 wins 53% (upset every other year)
Best of 3: Team #1 wins 84.3% (upset once every 6 years)
Best of 7: Team #1 wins 99.2% (upset once a century)

Team #1 is far and away the favorite - Teams 2 thru 64 each have well under a 1% shot - but it's basically a toss up versus the field with lose and go home.  If I change the scenario to Team #1 wins 19 of 20 instead of 9 out of 10, it's still under 75% likely to win it all.

Even in this ridiculous scenario, the most probable outcome out of 64 mathematically possible outcomes happens 53% of the time in a single-elimination format.

Now compare this with the most probable outcome out of 12 mathematically possible outcomes happening only 16.67% of the time in a literal crapshoot.

Mathematically, you cannot argue that this tournament is a crapshoot.

You can certainly emotionally call it that if that feels right. Math and facts can't really stop anyone from doing that.

[wadesworld]

Definite crapshoot or the best team would win every year.

[5DollarPitcher]

Definite crapshoot or the best team would win every year.

You need to compare apples to apples.

In craps you have 12 possible outcomes. You need to pool teams by ranking into 12 tiers to create an honest comparison.

Statistically, the champion comes from the top tier (of 12) 86% of the time.

[lostpassword]

You need to compare apples to apples.

In craps you have 12 possible outcomes. You need to pool teams by ranking into 12 tiers to create an honest comparison.

Statistically, the champion comes from the top tier (of 12) 86% of the time.

Ruh roh... I think you have exceeded your post quota in this thread.

[5DollarPitcher]

Ruh roh... I think you have exceeded your post quota in this thread.

I was told we get one more reply *to each other*

Haven’t replied to Rico as far as I know

[FartyEightHours]

Chicos would have loved this thread 10 yrs ago when everyone called him out for saying it was a crapshoot.  P.S.  I believe there is no way in hell your odds for ANY team to win is synonymous with a “crapshoot” of two dice.

[wadesworld]

My favorite part is that the original post says definitions will not be provided, answer as you wish. Then when the poll doesn’t go his way he tries to give definitions. Lol.

[brewcity77]

You need to compare apples to apples.

In craps you have 12 possible outcomes. You need to pool teams by ranking into 12 tiers to create an honest comparison.

Statistically, the champion comes from the top tier (of 12) 86% of the time.

Okay...but effectively, the 12-16 seeds are the equivalent of rolling a 1 or 13 in craps. It never happens. And as I said earlier, you basically have the top-25 as your potential Championship tier. Because the number of permutations in craps are not equal, you need to break those 25 teams into 11 tiers (not 12, there are only 11 different numeric totals you can roll in craps).

7 - #1 team
6/8 - #2-3 teams
5/9 - #4-7 teams
4/10 - #8-13 teams
3/11 - #14-21 teams
2/12 - #22+ teams

Next are the combinations.

7 - 6 combinations
6/8 - 10 combinations
5/9 - 8 combinations
4/10 - 6 combinations
3/11 - 4 combinations
2/12 - 2 combinations

So comparing the permutations to the ratio, here's what we get expectation vs reality (considering that 21 tournaments for 11 outcomes is a very small sample size):

#1 - 14.3% actual vs 16.7% expected
#2-3 - 33.3% actual vs 27.8% expected
#4-7 - 38.1% actual vs 22.2% expected
#8-13 - 0.0% actual vs 16.7% expected
#14-21 - 9.5% actual vs 11.1% expected
#22+ - 4.8% actual vs 5.6% expected

The disparities for 4-13 are stark, but everything else is almost exact. In terms of crapshoot, with the idea that your actual contenders are the top 25 or so teams, with some margin for error given in the 2/12 roll range for the outliers like 2021 UCLA or 2023 Miami or the hypothetical 12-16 winning it all, it sure feels like crapshoot is a pretty accurate depiction. And this is factoring UConn in as the 2023 winners already, so the 4-7 number would come down while the 14-21 or 22+ range would go up if it's anyone but UConn.

But ultimately...it doesn't really matter. It's just a mathematically apropos statement to make.

[Uncle Rico]

Okay...but effectively, the 12-16 seeds are the equivalent of rolling a 1 or 13 in craps. It never happens. And as I said earlier, you basically have the top-25 as your potential Championship tier. Because the number of permutations in craps are not equal, you need to break those 25 teams into 11 tiers (not 12, there are only 11 different numeric totals you can roll in craps).

7 - #1 team
6/8 - #2-3 teams
5/9 - #4-7 teams
4/10 - #8-13 teams
3/11 - #14-21 teams
2/12 - #22+ teams

Next are the combinations.

7 - 6 combinations
6/8 - 10 combinations
5/9 - 8 combinations
4/10 - 6 combinations
3/11 - 4 combinations
2/12 - 2 combinations

So comparing the permutations to the ratio, here's what we get expectation vs reality (considering that 21 tournaments for 11 outcomes is a very small sample size):

#1 - 14.3% actual vs 16.7% expected
#2-3 - 33.3% actual vs 27.8% expected
#4-7 - 38.1% actual vs 22.2% expected
#8-13 - 0.0% actual vs 16.7% expected
#14-21 - 9.5% actual vs 11.1% expected
#22+ - 4.8% actual vs 5.6% expected

The disparities for 4-13 are stark, but everything else is almost exact. In terms of crapshoot, with the idea that your actual contenders are the top 25 or so teams, with some margin for error given in the 2/12 roll range for the outliers like 2021 UCLA or 2023 Miami or the hypothetical 12-16 winning it all, it sure feels like crapshoot is a pretty accurate depiction. And this is factoring UConn in as the 2023 winners already, so the 4-7 number would come down while the 14-21 or 22+ range would go up if it's anyone but UConn.

But ultimately...it doesn't really matter. It's just a mathematically apropos statement to make.

Yup, it’s a crapshoot

[#UnleashSean]

You need to compare apples to apples.

In craps you have 12 possible outcomes. You need to pool teams by ranking into 12 tiers to create an honest comparison.

Statistically, the champion comes from the top tier (of 12) 86% of the time.

As already stated in the previous thread, you're taking the definition of crapshoot too literally.

[5DollarPitcher]

As already stated in the previous thread, you're taking the definition of crapshoot too literally.

So what you’re saying is it’s not a crapshoot

[5DollarPitcher]

Okay...but effectively, the 12-16 seeds are the equivalent of rolling a 1 or 13 in craps. It never happens. And as I said earlier, you basically have the top-25 as your potential Championship tier. Because the number of permutations in craps are not equal, you need to break those 25 teams into 11 tiers (not 12, there are only 11 different numeric totals you can roll in craps).

7 - #1 team
6/8 - #2-3 teams
5/9 - #4-7 teams
4/10 - #8-13 teams
3/11 - #14-21 teams
2/12 - #22+ teams

Next are the combinations.

7 - 6 combinations
6/8 - 10 combinations
5/9 - 8 combinations
4/10 - 6 combinations
3/11 - 4 combinations
2/12 - 2 combinations

So comparing the permutations to the ratio, here's what we get expectation vs reality (considering that 21 tournaments for 11 outcomes is a very small sample size):

#1 - 14.3% actual vs 16.7% expected
#2-3 - 33.3% actual vs 27.8% expected
#4-7 - 38.1% actual vs 22.2% expected
#8-13 - 0.0% actual vs 16.7% expected
#14-21 - 9.5% actual vs 11.1% expected
#22+ - 4.8% actual vs 5.6% expected

The disparities for 4-13 are stark, but everything else is almost exact. In terms of crapshoot, with the idea that your actual contenders are the top 25 or so teams, with some margin for error given in the 2/12 roll range for the outliers like 2021 UCLA or 2023 Miami or the hypothetical 12-16 winning it all, it sure feels like crapshoot is a pretty accurate depiction. And this is factoring UConn in as the 2023 winners already, so the 4-7 number would come down while the 14-21 or 22+ range would go up if it's anyone but UConn.

But ultimately...it doesn't really matter. It's just a mathematically apropos statement to make.

What you’re doing here is weighting the dice in the NCAA tournament field to reflect the fact that better teams are much more likely to go far than worse teams. And worse teams have a virtual 0% chance to go far.

This would be equivalent to acknowledging the tournament not a crapshoot, but rather a large field with a very focused group that actually has a chance to win.

If the argument is that the tournament is a crapshoot ONLY amongst the top 20-25 teams - I would actually agree with that. But that’s not the argument people are making. People are saying the NCAA tournament, on the whole, is a crapshoot. Which is neither mathematically or logically true.

[Uncle Rico]

As already stated in the previous thread, you're taking the definition of crapshoot too literally.

It’s a crapshoot

[Uncle Rico]

Jay Wright on if the tournament is a crapshoot:

https://www.inquirer.com/philly/sports/colleges/villanova/villanova-march-madness-2018-ncaa-tournament-east-region-20180311.html?outputType=amp

[Uncle Rico]

Bill Self crapshooting:

https://247sports.com/college/kansas/Article/KU-basketballs-Bill-Self-talks-unpredictability-of-NCAA-Tournament-145489019/

[Scoop Snoop]

I don't give a crap. Just enjoy the tourney.

[Uncle Rico]

I don't give a crap. Just enjoy the tourney.

Precisely.  The charm of a single-elimination tournament is the volatility.  It sucks when that volatility hits your team. 

Personally, I’d prefer a Final 4 with the 4 best teams.  Upsets opening weekend are fine but I’d rather see UConn with Houston, Bama, Texas, UCLA, Creighton, Marquette or Kansas but March surprises abound like 2011