Scoping it out - 2018

[SWC75]

It’s time for that annual exercise in self-flagellation, the filling out of our annual NCAA basketball tournament prediction sheets. We are handicapped by two things: the awesome gap between our supposed knowledge of college basketball and our actual knowledge of it and the blank screen that is the future. I long ago realized that there is little relationship between the amount of time you spend on figuring things out and the accuracy of the results so I want to waste as few of the precious seconds of my remaining time on this earth doing my sheet so I’ve come up with a systematic method not of picking the correct winners but just of predicting something so I can get it done in a finite period of time.

For several years I’ve been using the procedure I outlined in this post: NCAA Seed Stats - updated through 2017
…and then picked the wrong upsets. I’m going to supplement that procedure by using the information in this post to help me pick the winners:
Against Ranked teams - FWIW

I figure that will be at least as valid as my normal mental gymnastics and will save even more time. So here goes:

Step One – Pick all the #1 seeds to win in the first round.

Step Two- Pick all the #2 seeds to do the same, even though they are historically a little vulnerable.

Step Three: Pick one #14 seed to beat a # seeds, then pick the other #3 seeds to win their games.
Tennessee has +52 points. Wright State has 0. Michigan has +161 points, Montana 0. Texas Tech has +176, Stephen F. Austin 0. Michigan State has +118, Bucknell #5. Bucknell is the best of the 14s but Tennessee is the worst of the 3’s and the differential is only +52 between them and their opponent, easily the lowest of the four match-ups. So, I’ll choose Wright State to upset Tennessee at the #3 vs.#14 line.

Step Four: Pick one #13 seed to beat a #4 seed but the other #4 seeds all win.
Arizona Is +42, Buffalo +10. Gonzaga is +65, UNC-Greensboro 0. Wichita State is +64, Marshall +24. Auburn is +28, Charleston +2. So I’ll go with Charleston to upset Auburn.

Step Five: Pick one #12 seed to beat a #5 seed but have the other #5 seeds win.
Kentucky is +65, Davidson +1. Ohio State is +50, South Dakota State -2. West Virginia is +111, Murray State 0. Clemson is +59, New Mexico State +16. So I’ll go with New Mexico State to beat Clemson.

Step Six: Pick one #11 seed to beat a #6 seed but the other #6 seeds win, (this one has bene pretty shake in recent years as the #11 seeds are 6-2 in the last two tourneys)
U of Miami is #71, Loyola-Chicago 0. Houston is +93, San Diego State +15. Florida is +117, St Bonaventure +13 and UCLA +19. Texas Christian is +84, Arizona State +66 and Syracuse +48. By the method I’ve decided to use, TCU is the #6 that will go down but, unfortunately, Arizona State is the team that will beat them. Three of the four upsets so far are in the Midwest regional.

Step Seven: Pick two #10 seeds to upset #7 seeds but the #7s win the other two games.
Nevada-Reno is only +2 and Texas +129, so that’s low hanging fruit. Texas A&M is +105 and Providence +92. Arkansas is +86, Butler +68. Rhode Island is +7 and Oklahoma +129 so that’s also pretty juicy. Texas and Oklahoma are the obvious choice for “upsets”. The committee apparently wasn’t that impressed with the Big 12.

Step Eight: Pick two #9 seeds to beat #8 seeds but pick the other #8 seeds to win.
Creighton is +85, Kansas State +19. Missouri is +22, Florida State +105. Virginia Tech is +87 but Alabama is +168. Seton Hall is #+73 but North Carolina State is +114. I’ll go with Florida State and Alabama to pull off the upsets here, but I don’t see them as upsets.

I have now picked 32 of the 63 games without my mind exploding.

Step Nine: Pick Your Final Four, consisting of two 1 seeds, a #2 seed and somebody else. Have one of the #1 seeds win the national championship. The top #1 seed against ranked teams is Kansas +122, followed by Virginia +119, Villanova +108 and Xavier +98. I should pick Kansas to make the Fianl Four and win it about I just don’t have confidence in that pick. I think the big 12 had a lot of good teams but no great ones and that this actually isn’t one of Kansas’ more impressive teams. I’ll pick the two ‘V’ teams and Virginia to win it all. “But they have a lousy record in the NCAA tournament.” So did Connecticut and even Duke- until they won it, again and again.

Among the 2 seeds, Cincinnati is only +32. North Carolina is +117, Purdue +94 and Duke +154. I’ll go with the Dukies to knock off the Jayhawks in the regional final. (Kansas’ route there is paved with the 1st round upsets.)

That means that the lower seeded Final Four team must come from the West. Florida State is +105, Ohio State +50, Gonzaga +65, Houston +93, Michigan +161 and Texas A&M +105. I like the way the Wolverines are playing so they are my other Final Four team.

I’ve no picked 17 more games for a total of 49 out of 63.

Step Ten: I need to pick a 10 seed to get to the Sweet 16 and stop there. Oklahoma runs into Duke so they can’t get there. It’s got to be Texas. The other 1st round upsetters are to lose: Wright State to Miami who will then beat Texas before losing to Virginia. Florida State will lose to Xavier. I already have Alabama losing to Villanova. New Mexico State is +16 and Charleston +2. One of them has to win so I’ll pick the Aggie to win that one and then lose to Kansas. That’s 54 of 63 games.

Step Eleven: Do the best you can with the other 9 games. Kentucky is +65 and hot. Arizona is +42. Ohio State is +50 but Gonzaga is +65. Texas A&M is +105 but North Carolina is +117. West Virginia is ++111, Wichita State +64. Florida is +117 but Texas Tech is +176. Arkansas is +86 but Purdue is +94. Arizona State is +66 but Michigan State is +118.

Xavier is +98, Gonzaga +65. Texas tech is +176, Purdue +94.

Tiebreaker? The average score of the NCAA championship game in the last ten years has been 70-63. So I’ll guess the total points at 133.

Done. Correct? Who knows? But it's done and I’ve still got a life.

 

[bballbeadle]

Okay, this made my head hurt but I found it intriguing. I actually wrote you a longer post because I couldn't quite understanding the number differential, but I gather that the smaller the number between two teams in a group, the more likely the upset. I'm going to give this a try. I always do two brackets so this can be my scientific bracket and the other will be my emotional one.

However, any system that has SU losing in the first round I am afraid I have difficulty respecting. Why is ASU 66 and SU 48? Please explain gently! Obviously, I have grasped the smaller picture of the number comparison but not the larger picture of how the numbers are arrived at.

 

[orangehomer]

It’s time for that annual exercise in self-flagellation

it's easier if we stick to predictions of self flatulation

 

[JrodCuse]

Can you explain what the +# means?

 

[SWC75]

Can you explain what the +# means?

The + or - numbers are from my "Against Ranked Teams" post which i linked in the above post. Here is the explanation I provided at the beginning of the ART post, (which I've posted once a week for the entire season as the numbers changed):

"This year I will be again charting how teams have done “against ranked teams”, which is something that always comes up at the end of the year when the arguments over who should be in the NCAA tournament start. As usual, this is not meant to end those arguments, just to give those making them something to look at in forming their opinions. This is just one thing to look at. But I think it’s a big thing.

I’m going to do it a bit differently this year. The old procedure was that I would first grant teams playing ranked teams a certain number of points based on who they are playing. If you played the #1 team, (I use the AP- writer’s poll for continuity), you get 25 points. If you are playing the #2 team, you get 24 points, etc. down to getting 1 point for playing the #25 team. I call these “schedule points”. Here’s a chart to make it simple:
1-25, 2-24, 3-23, 4-22, 5-21, 6-20, 7-19, 8-18, 9-17, 10-16, 11-15, 12-14, 13-13, 14-12, 15-11, 16-10, 17-9, 18-8, 19-7, 20-6, 21-5, 22-4, 23-3, 24-2, 25-1.
(Of course you can always just subtract the ranking from 26. That will work every time.)

Then I look at the point differential in the actual game. If you defeated the ranked team, you add the number of points you won by to the schedule points. If you lost to that team, you subtract the margin you lost by from the schedule points. (I call these ‘game’ points.) The difference between this year and last year if that I didn’t count any deficit: if you got to zero, you got zero. This year I’m going to try to record negative points, too, so a team can get full ‘credit’ for a really bad performance."

Thus the numbers are the cumulative + or - numbers in each team's games against ranked teams. I figure that's as good a measure as any for how they might do in the tournament. I usually sweat out arguments about how one team's strengths, weaknesses and playing style match up against another's but the results over the years have not been impressive so I went with this approach this year. It will probably do just as well, maybe better and it sure was quicker.

 

[SWC75]

Okay, this made my head hurt but I found it intriguing. I actually wrote you a longer post because I couldn't quite understanding the number differential, but I gather that the smaller the number between two teams in a group, the more likely the upset. I'm going to give this a try. I always do two brackets so this can be my scientific bracket and the other will be my emotional one.

However, any system that has SU losing in the first round I am afraid I have difficulty respecting. Why is ASU 66 and SU 48? Please explain gently! Obviously, I have grasped the smaller picture of the number comparison but not the larger picture of how the numbers are arrived at.

Here are the team's results against ranked teams this season, (the ranking are from the time they played them, hnot the final ranking):

Week #3: Arizona State beat #15 Xavier 102-86 = +27 points
Week #4: Syracuse lost to #2 Kansas 60-76 = +8 points
Week #5: Arizona State beat #2 Kansas 95-85 = +34 points
Week #8: Arizona State lost to #17 Arizona 78-84 = +3 points
Week #10: Syracuse lost to #3 Virginia 61-68 = +16 points
and: Syracuse lost to #23 Florida State 90-101 = -8 points
Week #13: Syracuse lost to #2 Virginia 44-59 = +9 points
Week #15: Arizona State lost to #17 Arizona 70-77 = +2 points
Week #16: Syracuse lost to #10 North Carolina 74-78 = +12 points
and: Syracuse lost to #5 Duke 44-60 = +5 points
Week #17: Syracuse beat #18 Clemson 55-52 = +11 points
Week #18: Syracuse lost to #12 North Carolina 59-78 = -5 points

That adds up to +66 for Arizona State, (based primarily on their early season wins over Xavier and Kansas) vs. +48 for Syracuse. if you want to you can say that since ASU's points were early and they've not performed well of late, you are going to pick Syracuse and then have them be the #11 seed that gets to the Sweet 16. Sounds good to me.

 

[SWC75]

12-4 the first day. I had three upsets predicted: Wright State was Wrong State and Rhode Island road over the Sooners. But I had Alabama over Va Tech. I didn't see Loyola over Miami comings or Buffalo's devastation of Arizona. Our win over them is looking pretty good right now. (I had the Canes winning two more games so I guess I'm actually 12-6.)

Back at it tomorrow.

 

[721Comstock]

12-4 the first day. I had three upsets predicted: Wright State was Wrong State and Rhode Island road over the Sooners. But I had Alabama over Va Tech. I didn't see Loyola over Miami comings or Buffalo's devastation of Arizona. Our win voer them is looking pretty god right now.

Back at it tomorrow.

Wonder if that will help our NCAA seeding?

 

[SWC75]

Wonder if that will help our NCAA seeding?

(It didn't help my typing.) It does suggest the fact that we had a seeding was more appropiate than some people thought.

 

[SWC75]

Fly the friendly skies!

That is my NCAA prediction sheet as of right now. I went 8-8 tonight with my national champion going down. Nothing like drawing six straight horizontal lines in March.

I had predicted 8 first round upsets and got 2 of them. Actually, the SU game was an upset I could have gotten but i predicted Arizona State to get it, so that's a loss. There were 6 other upset I didn't have. I had the wright number but, as usual, the wrong teams.

 

[SWC75]

Incredibly, I went 7-1 today, with the one loss being a game I couldn't win because neither of the teams I had playing in it won their first round games, (Loyola beat Tennessee). But I've got no shot at such a game tomorrow after going 8-8 on the second night. 5-3 is my glass ceiling. And thanks to UMBC and Loyola, I've got five more losers in subsequent rounds.

 

[SWC75]

I went a predictable 2-6 Sunday and 1-3 tonight for an unimpressive total of 30-22 but I still have 3 of my Final Four still alive: Michigan, Duke and Villanova. But the retrievers made my sheet irretrievable.

 

[SWC75]

I got all four games Friday night, (wish I hadn't) and went 3-1 over the weekend, (with the South already being a shell hole), so I've actually got a pretty good record at 37-23 with 3 of the four Final Four teams. Unfortunately, my favorite to win it all, Virginia, went down hard in the first round and killed my chances right out of the gate. If they had beaten UMBC and made ti to the FF, I'd be 41-19 with all four FF teams. That's better than I've done in years. My plan of using my "against ranked teams" ratings has worked out better than all the theorizing I used in past years and it enabled me to fill out my sheet in about 1/2 hour. I think I'll do the same thing in future years.

 

[bballbeadle]

I got all four games Friday night, (wish I hadn't) and went 3-1 over the weekend, (with the South already being a shell hole), so I've actually got a pretty good record at 37-23 with 3 of the four Final Four teams. Unfortunately, my favorite to win it all, Virginia, went down hard in the first round and killed my chances right out of the gate. If they had beaten UMBC and made ti to the FF, I'd be 41-19 with all four FF teams. That's better than I've done in years. My plan of using my "against ranked teams" ratings has worked out better than all the theorizing I used in past years and it enabled me to fill out my sheet in about 1/2 hour. I think I'll do the same thing in future years.

I filled out one bracket using my head and heart. Bracket #2 I used your system for the most part. In the first two days, my bracket #2 was a disaster and in #1 I looked like a genius, despite the ridiculous upsets. But by the second round, #2 was beginning to pay off for me, particularly the pick of Michigan, which I NEVER would have done. And after this past weekend, I will win my pool with bracket #2 IF Michigan beats the Little Sisters of Redemption.

So, I might use your system again next year although it pained me terribly to do it. Ha ha!