Tuesday, November 27, 2012

NFL Week 12 Projected Records - The picture gets muddier

The projected records after week 12 are now posted and listed below as well.

What was a pretty clear playoff picture has become a little cloudier.

The division leaders are still very solid.  With Green Bay's loss, Chicago is now projected as the winner in the NFC North and with Dallas' loss, the Giants control the NFC East.  And Pittsburgh's loss gives control of the AFC North to Baltimore.  But the wildcards are a different story.

In the NFC where the two teams were expected to be the Green Bay/Chicago non-division winner and Seattle comfortably, it is now Green Bay still comfortably but Seattle with just a 1 game cushion over 4 teams at 8-8.  This could get interesting.

And in the AFC, where it was fairly clear it would be the Baltimore/Pittsburgh non-division winner and a surprising Indianapolis, those two teams now find themselves in a dead heat with Cincinnati all at 9-7.

I am also tracking the projected playoff teams by week so take a look at that to see how things change from week to week.

TeamRecord% chance% +1% -1
TeamRecord% chance% +1% -1
San Francisco11-436.427.821.7
New England11-533.633.216.4
Green Bay10-635.932.517.3
NY Giants10-637.023.226.1
Tampa Bay8-836.135.712.4
New Orleans7-936.924.324.9
NY Jets7-934.221.227.2
San Diego7-935.018.130.4
St Louis6-935.723.725.4
Kansas City2-1441.925.924.7

Monday, November 26, 2012

NFL Week 12 Ratings and Rankings - Houston and New England remain 1 and 2

The ratings after week 12 are now posted and listed below as well.

Houston stays #1 but New England has narrowed the gap, and the 49ers edge ahead of the Broncos.

1Houston88.82310-180.056+0, -0.515
2New England88.5198-380.706+0, +0.525
3San Francisco87.6318-2-181.243+1, +0.128
4Denver87.3398-380.995-1, -0.481
5Chicago85.5428-380.752+3, +0.583
6Atlanta85.35710-179.101+0, +0.073
7Green Bay85.0767-482.259-2, -2.072
8NY Giants84.8527-480.781+2, +1.943
9Baltimore84.3689-279.585+0, +0.379
10Seattle84.0316-582.051-3, -1.030
11Tampa Bay82.7196-579.350+0, -0.181
12Washington81.3885-681.261+5, +1.200
13Cincinnati81.3126-578.913+5, +1.176
14Pittsburgh81.2096-579.372-2, -0.734
15Dallas80.8395-682.426-1, -0.647
16Minnesota80.6856-580.416-3, -0.965
17New Orleans80.4655-680.275-2, -0.651
18St Louis80.2034-6-182.196+4, +1.041
19Detroit79.8774-781.759+0, -0.205
20NY Jets79.8624-782.903-4, -0.757
21Miami79.5285-680.071+2, +0.374
22San Diego79.4884-779.892-2, -0.323
23Carolina79.4873-882.363+1, +0.605
24Indianapolis78.7167-479.599+2, +0.445
25Buffalo78.2264-782.079+0, -0.607
26Arizona77.8524-781.836-5, -1.670
27Cleveland77.3523-880.494+1, +1.357
28Tennessee76.0814-781.588-1, -0.728
29Jacksonville73.8582-981.224+2, +0.724
30Philadelphia73.7393-880.913-1, -0.452
31Oakland72.8093-880.876-1, -0.700
32Kansas City70.9901-1080.889+0, +0.400

USTA 2012 year end ratings are out - How did my Estimated Dynamic NTRP do?

It turns out that despite the lack of a notice on TennisLink, year-end ratings are in fact being published today and they have even done so well in advance of the traditional 5pm.  With that, TennisLink is now enduring its peak load for the year as we all check to see if we or our prospective teammates have been bumped up/down so we can know how to assemble our teams for 2013.

I've been generating Estimated Dynamic NTRP reports for anyone interested in a preview of what their year-end rating may be and the details behind it and with a number of reports done, it is now time to compare what I estimated with what year-end ratings ended up being.

Of all the reports I did, 79% estimated the player to be at the right level.  But I consider being within 0.05 to be an acceptable margin for error considering it is just an estimate and an additional 9% or 88% total fell within this range.  Of the remaining 12%, I had 3% estimated too high and 9% estimated too low.

All those I estimated too high were self-rates that went to playoffs, so since I don't attempt to estimate the benchmark part of the calculation, I don't feel too bad about missing here.  And I did correctly estimate a self-rate being bumped down so I don't think I have a general problem with overrating self-rated players.

Of those I estimated too low, one went to playoffs (benchmark calculation likely kicked in) and one was just 0.07 from being estimated to be bumped up and also went to playoffs (benchmark calculation kicked in), so I don't feel too bad about these.

So I'll take 88% accuracy where most of the misses have explanations that I actually commented on being possible in the report I generated.

Even though ratings are out, if you want to know why your rating is where it is, I am still generating reports so contact me if you are interested.

Update 12/20: More reports have been generated and the accuracy is now at 92%.

Sunday, November 25, 2012

When will USTA NTRP Year End Ratings be out? Get a report while you wait

[this post was for 2012 year-end ratings, click here to see information for 2013]

Many people are expecting that the USTA will publish year-end ratings tomorrow at 5pm, but TennisLink doesn't seem to be announcing that like prior years.  The speculation is based on it being the Monday after Thanksgiving which is when the ratings typically come out, but this year, Thanksgiving is the earliest it can be making the Monday after only November 26th.  So, it could be that they won't be ready until the following Monday December 3rd.

If that is the case, we'll all have to wait an extra week, but that also gives everyone another 7 days to get a preview and details of what goes into their rating from an Estimated Dynamic NTRP Report.

An example of a report is below.  It shows the current NTRP level of the player, what their Estimated Dynamic NTRP is (what I calculate), some match/set/game statistics, the best and worst match results and dates of the match, and the highest and lowest Estimated Dynamic NTRP and corresponding dates.

But the cool part is the chart that shows each match rating and what the Estimated Dynamic NTRP is as a result of the match.  This allows you to see specifics about matches and trends throughout the year.

Jane Doe
Current NTRP: 3.0
Estimated DNTRP: 2.76
Match Record: 4-8
Sets Won-Lost: 13-16
Games Won-Lost: 112-117
Best Match Result: 3.04 on 2012-03-18
Worst Match Result: 2.53 on 2012-03-11
Highest Estimated DNTRP: 2.86 on 2012-01-29
Lowest Estimated DNTRP: 2.73 on 2012-03-11

I've done reports for players in the Pacific Northwest, Southern Cal, Northern Cal, Inter-Mountain, Texas, Midwest, Southern, Florida Mid-Atlantic, and Northeast sections but should be able to do for most any section.

Contact me to learn more or if you have any questions.

The voters who dropped Oregon after 24 point win on road against #16 - East Coast or anti-Oregon bias?

In analyzing the AP poll, I discovered that Oregon lost 4 points while Florida gained 91 and moved ahead of them.  This after both won on the road against similar opponents (Oregon State was and still is #16, Florida State was #10 and is now #13), Oregon winning far more convincingly (by 24 vs 11 and Oregon game wasn't in doubt late, FSU had a chance in the 4th quarter).

To understand how this happened, I looked for the voters that flipped the two teams and here they are:

  • Axe - Moved Florida from 7 to 5 and dropped Oregon from 6 to 7.
  • Bohls - Moved Florida from 7 to 6 and dropped Oregon from 3 to 7!
  • Cripe - Moved Florida from 6 to 3 and dropped Oregon from 4 to 7.
  • Emerson - Swapped the teams at 5 and 6.
  • Fowler - Moved Florida from 6 to 4 and dropped Oregon from 5 to 7.
  • Giglio - Moved Florida 5 to 2 and dropped Oregon 4 to 5.
  • Gimino - Moved Florida 6 to 4 and kept Oregon at 5.
  • Hlas - Moved Florida 7 to 2 and kept Oregon at 4.
  • Hurt - Moved Florida 7 to 4 and dropped Oregon 4 to 5.
  • Lesmerises - Moved Florida 5 to 4 and dropped Oregon 4 to 6.
  • Marcello - Swapped the teams at 5 and 6.
  • Murray - Moved Florida 6 to 3 and dropped Oregon 4 to 5.
  • Staples - Moved Florida 12(!) to 4 and dropped Oregon 4 to 5.
Some of these moves are to put it bluntly, baffling.  Bohls thought Oregon was still #3 last week, and they go out and beat Oregon State by 24 on the road and drop them 4 spots?  And he actually moves Oregon state up from #19 to #18 with the loss?

Staples thought FSU was a top-5 team (had them #5 last week) despite all the evidence to the contrary, and when proved wrong with Florida beat them, moves Florida, who he had #12 all the way to #4, higher than the overrated FSU team.

It was pointed out to me that Oregon's game was televised only regionally while Florida's was nationally televised, and I can understand the general public being swayed by that, but I thought the AP voters could be held to a higher standard.

Analyzing the 11/25/12 AP Poll - Bias against Oregon?

Here is week 13 of analyzing the AP poll.  Thanks as always to the AP publishing their voting.

All 13 charts are displayed below to make it easy to see the changes.

The top-4 remain the same this week with very little change for each.  Alabama gained 1 point, Georgia 15, and Ohio State 2.  Where it gets interesting is behind them.

Last week Oregon was #5 with a 75 point lead over #6 Florida.  This week, Florida turned that around to a 20 point lead over Oregon.  Oregon beat a then and still #16 Oregon State on the road by 24 and actually lost 4 points.  That isn't to say that Florida didn't do well, they did win by 11 on the road against a then #10 and now #13 FSU.  But both win on road against similar competition, and the team that wins by more loses points while the one that wins by less gains 91?  There seems to be some serious bias against Oregon with these voters.  Oregon even got a 2nd place vote this week, they didn't last, and lost points.

Ohio State still has the largest spread of the top teams garnering 2nd to 12th place votes.  You have to drop all the way to Stanford to get to the next team getting a 12th place vote.  And like last week, Texas A&M is getting a 2nd place vote ahead of a bunch of 1-loss (and Ohio State with no losses) teams.

College Football Week 13 Ratings and Rankings - Oregon stays #1

This week's ratings and rankings are now posted.

Oregon stays #1 this week, in fact increasing their lead over Alabama who also improved slightly.  Florida improves the most of any team in the top-10 and moves ahead of Texas A&M to #3.

Oregon's improvement is largely from their win over a good Oregon State team while Alabama played a weak Auburn team but benefited from the SEC winning in general over some non-SEC opponents (Florida over FSU, Georgia over Georgia Tech, South Carolina over Clemson).  And while Florida improved their rating the most, Georgia made the biggest move up to #7, up 4 spots.

What about Notre Dame?  Their rating did improve after their win over USC and they are up to #5, but my computer still says they aren't the best team in the country.

The BCS compatible algorithm ratings and rankings do have Notre Dame on top and they are followed by Florida and Ohio State, with Oregon up to #4 ahead of Alabama at #5.  A computer that can't take margin of victory into account has to ding Alabama for a weak Auburn opponent while Oregon benefits from a strong Oregon State team.  We'll see what the actual BCS computers do later today.

The top-25 is below, full ratings on the site.

1Oregon93.98311-173.073+0, +1.273
2Alabama91.67211-171.655+0, +0.408
3Florida89.28811-175.872+1, +2.240
4Texas A&M88.39310-274.031-1, +1.033
5Notre Dame87.97712-073.981+1, +1.527
6Kansas St87.36910-173.988-1, +0.341
7Georgia87.07911-171.780+4, +1.798
8Stanford86.41910-277.222+1, +0.892
9South Carolina86.36810-273.698+3, +2.063
10LSU85.70910-274.194-2, +0.075
11Oklahoma85.2259-273.541-1, -0.186
12Oregon St84.3988-378.310-5, -1.845
13Oklahoma St83.7377-472.357+2, -0.014
14Ohio State83.35412-069.689+0, -0.444
15Southern Cal82.8317-576.039-2, -1.393
16Utah St81.93610-266.110+2, -0.504
17Florida St81.47510-264.626-1, -1.868
18Texas81.2288-374.529-1, -1.503
19Nebraska81.04710-270.493+2, -0.494
20Clemson80.22810-267.731+0, -2.070
21Arizona St80.1767-574.041+4, +0.890
22Arizona80.1387-575.060-3, -2.163
23Baylor80.0886-576.182+0, +0.663
24Michigan79.7758-472.775+0, +0.429
25UCLA79.6779-373.924-3, -1.745

Tuesday, November 20, 2012

The ever colorful NFL Week 11 Projected Records Chart

I introduced a colorful wins chart last week and I've updated it with this week's projected records.  It is a bit busy and has a ton of info in it, but find the segment for your favorite team and get a feel for how many wins they might get to.

NFL Week 11 "What If" Ratings, Rankings, and Projections - Giants could be the team left out

Below you will find my computer's ratings, rankings, and projections calculated using the week 3 Green Bay at Seattle game as a 12-7 win by Green Bay.  This is being done to see how things would be different if the officials had made the right call.  I've done this a few times this season for this case because I don't recall there ever being a single controversial call that was the clear deciding factor in the game, and this has clear playoff implications.

For the ratings and rankings, the change column is the difference from the actual week 11 ratings.  You'll see that Green Bay would be up 3 spots to #2 and Seattle would drop just one spot to #8.

1Houston89.4489-180.050+0, +0.110
2Green Bay87.9928-281.772+3, +0.844
3New England87.8997-380.803-1, -0.095
4Denver87.7937-381.759-1, -0.027
5San Francisco87.4897-2-181.385-1, -0.014
6Atlanta85.2479-178.673+0, -0.037
7Chicago85.0447-381.268+1, +0.085
8Seattle84.3225-582.437-1, -0.739
9Baltimore83.9668-279.091+0, -0.023
10NY Giants82.8566-480.417+0, -0.053
11Tampa Bay82.8446-478.847+0, -0.056
12Pittsburgh81.9096-479.043+0, -0.034
13Minnesota81.5656-479.683+0, -0.085
14Dallas81.3825-582.204+0, -0.104
15New Orleans81.1265-580.080+0, +0.010
16NY Jets80.5334-682.566+0, -0.086
17Washington80.1434-680.775+0, -0.045
18Cincinnati80.1155-579.153+0, -0.021
19Detroit80.0834-681.730+0, +0.001
20San Diego79.7884-679.821+0, -0.023
21Arizona79.5014-682.577+0, -0.021
22St Louis79.1543-6-182.615+0, -0.008
23Miami79.1434-680.274+0, -0.011
24Buffalo78.8054-682.126+1, -0.028
25Carolina78.7752-882.760-1, -0.107
26Indianapolis78.3456-480.137+0, +0.074
27Tennessee76.8084-682.203+0, -0.001
28Cleveland75.9702-880.349+0, -0.025
29Philadelphia74.1563-781.218+0, -0.035
30Oakland73.4973-780.559+0, -0.012
31Jacksonville73.2081-982.388+0, +0.074
32Kansas City70.5681-980.709+0, -0.022

And here are the projected records where you see Green Bay winning the division outright (in the current actual projections they are projected to win the tie-breaker with Chicago at 11-5), and Seattle would fall to 9-7 and be tied with Dallas or the Giants for the last wildcard rather than winning it outright.  Should they tie with Dallas, they have the advantage with a head to head win, but should it be with the Giants, they'd likely lose out on conference record.

So, depending on how the NFC East plays out, there may be no change in the playoffs teams and seedings due to the bad call, or the Giants could be the team that loses a playoff spot because of it.

TeamRecord% chance% +1% -1
TeamRecord% chance% +1% -1
Green Bay12-434.631.218.2
San Francisco11-432.318.828.1
New England11-533.424.024.1
NY Giants9-733.225.622.2
Tampa Bay9-735.925.423.3
New Orleans8-831.117.229.8
NY Jets7-931.326.720.2
San Diego7-932.926.621.0
St Louis5-1033.025.721.9
Kansas City2-1440.127.522.1

I'll continue to calculate these each week so we can see how things would be different.

NFL Week 11 Projected Records - Things get clearer, especially in the AFC

The projected records after week 11 are now posted and listed below as well.

Atlanta and  Houston continue to lead the way as you might expect.  But behind them, it is tight with Denver at 12-4 then five teams projected to finish 11-5.

In the NFC, Atlanta is projected to have a first round bye and the 49ers have the inside track to the 2nd one with Green Bay and Chicago close behind with the Packers getting the nod for the division based on current tie-breakers.  The Giants would beat out the Cowboys, both 9-7, for the NFC East based on current tie-breakers and the Bears and Seahawks would pick up the wilcards.  Seattle does have a 21% chance of being 9-7 and Tampa Bay and Dallas have decent shots at 10-6, so this could all certainly change and is not a lock by any stretch.

The AFC is a bit clearer with all four division winners alone at 11-5 or better but Pittsburgh is close at 10-6 but really would need some help to get the division given the current tie-breakers.  The Colts, despite their bad loss, still hang on to the last wildcard and have only a 15% chance of 8-8, but the Bengals do have a 26% chance of 9-7 so this too isn't a lock.

I am also tracking the projected playoff teams by week so take a look at that to see how things change from week to week.

TeamRecord% chance% +1% -1
TeamRecord% chance% +1% -1
San Francisco11-4-132.218.428.4
Green Bay11-535.228.220.9
New England11-533.524.323.8
NY Giants9-733.126.121.6
Tampa Bay9-735.925.523.1
New Orleans8-831.016.930.0
NY Jets7-931.326.920.0
San Diego7-932.926.521.1
St Louis5-10-133.225.422.4
Kansas City2-1440.127.522.1

NFL Week 11 Ratings and Rankings - A few moves near the top

The ratings after week 11 are now posted and listed below as well.

Houston stays #1 while New England and Denver swap spots and San Francisco makes a strong move into 4th.

Kansas City has a stranglehold on last with a 2.5 ratings point "advantage" over Jacksonville.

1Houston89.3389-179.960+0, -1.100
2New England87.9947-380.892+2, +0.723
3Denver87.8207-381.782-1, -0.476
4San Francisco87.5037-2-181.394+3, +1.828
5Green Bay87.1487-381.815-2, -0.266
6Atlanta85.2849-178.712+0, -0.711
7Seattle85.0616-482.405+1, +0.149
8Chicago84.9597-381.182-3, -2.118
9Baltimore83.9898-279.118+0, +0.468
10NY Giants82.9096-480.471+0, +0.036
11Tampa Bay82.9006-478.902+1, +0.495
12Pittsburgh81.9436-479.076-1, -0.576
13Minnesota81.6506-479.755+1, +0.287
14Dallas81.4865-582.308-1, -0.893
15New Orleans81.1165-580.032+1, +0.657
16NY Jets80.6194-682.642+8, +1.686
17Washington80.1884-680.817+4, +0.764
18Cincinnati80.1365-579.172+4, +0.879
19Detroit80.0824-681.727-2, -0.250
20San Diego79.8114-679.843-2, +0.031
21Arizona79.5224-682.598+2, +0.474
22St Louis79.1623-6-182.623-7, -1.605
23Miami79.1544-680.282-3, -0.368
24Carolina78.8822-882.862-5, -0.807
25Buffalo78.8334-682.152+1, +0.961
26Indianapolis78.2716-480.060-1, -0.388
27Tennessee76.8094-682.204+0, -0.285
28Cleveland75.9952-880.375+0, +0.762
29Philadelphia74.1913-781.251+0, -1.003
30Oakland73.5093-780.574+0, -1.183
31Jacksonville73.1341-982.270+0, +0.470
32Kansas City70.5901-980.733+0, -0.864

Monday, November 19, 2012

BCS Math 11/19/12 Edition - Does Oregon or Florida have a shot?

Two weeks to go in the season and we have a clear #1 in the BCS so the fight is on for making the case for #2.

My computer's performance charts show that Oregon and Alabama have the strongest resumes, but the BCS doesn't agree slotting in Georgia and Florida between them.  And Kansas State is getting little love but Florida State even less.  What would it take for any of these teams to get to #2?

Alabama has a solid hold on #2 this week, but how might that change?  Despite the fact that Auburn is very weak this year (3-8 my #77), the pollsters aren't likely to ding Bama much at all even if they have a close win because it is a rivalry game, so their poll numbers aren't likely to slip.  Their computer numbers could though.  Here are the details:

  • Anderson & Hester - At #5 they have a slim lead over Stanford, 0.781 to 0.774 and not that far ahead of Oregon at 0.765.  Stanford just went up 0.019 with their win over Oregon, so with both Stanford (UCLA) and Oregon (Oregon State) playing good opponents, it isn't out of the question that one or both get ahead and drop Alabama to #7 or #8.
  • Billingsley - Alabama is #2 but close behind is Kansas State and Oregon.  K-State doesn't play so won't change (Billingsley doesn't change a rating at all during bye weeks, silly) but as mentioned above, Oregon has a chance to improve.
  • Colley - Alabama is #4 just ahead of Stanford and Oregon again and could slip at least one spot.
  • Massey - Alabama is #3 just ahead of LSU and Stanford but we don't know how close since his ratings aren't updated on his site.  LSU plays Arkansas (50s) so there is less chance to jump ahead, but LSU was just 0.01 back last week so it is possible and Stanford does get UCLA so more opportunity.
  • Sagarin - Alabama is #3 fractionally ahead of Stanford and Oregon so again could drop.
  • Wolfe - Has Alabama #3 over Georgia and Stanford, but perhaps enough margin to not drop.
This means that what they have this week (four 3rds and a 4th after throwing out the high and low of 2nd and 5th) could slip to 6th, 3rd, 5th, 4th, 4th, 3rd dropping their 0.910 computer average to 0.890 and with no change in their polling numbers their BCS rating to 0.9267.  That isn't near enough for anyone behind to get close, but may still be an interesting twist to see.

And of course, a win in the SEC title game would turn the computers around so it is really all moot.  Win out, and Alabama is in the BCS title game.

Georgia is in a similar situation to Alabama even though they trail them significantly, mainly because they get to play them.  They play a not so great Georgia Tech team (but not as bad as Auburn) so their computer numbers could take a minor hit similar to Alabama's, but if they win the SEC title game that is all moot and they get ahead of Alabama and likely aren't caught by anyone behind them.

Florida is pretty close to Georgia in the BCS ratings, and has a much tougher opponent in FSU this week, but they are already #2 in the computers and aren't going to catch Notre Dame there so won't see any BCS improvement from that.  It would have to come from the polls.  A win would certainly help, and they could get very close to Georgia in the BCS, but they don't get a title game meaning they don't get that last opportunity to improve their numbers and can't get to #2.  Their only hope is one or both of Auburn / Georgia Tech pull the upset, and the better team loses in the SEC title game leaving them as the best 1-loss team from the SEC.  Possible, but not likely.

Oregon still has room to improve in the computers, and gets BCS #15 Oregon State to finish, but that likely isn't enough given that they are 4th in the polls.  Even if their computer numbers improve a spot across the board, they at best get close to Florida but not all the way to Georgia let alone Alabama in the BCS ratings.  Couple that with being left out of the Pac-12 title game (for now) and they simply can't get there.

Now, if FSU beats Florida, their computer and poll numbers have the chance to improve a bit more and they could perhaps even get past Georgia in next weeks BCS ratings, but they'd still not improve enough to be ahead of the Alabama/Georgia winner.  If UCLA were to beat Stanford however, that gets Oregon into the title game against a hot UCLA team that may give them the opportunity to improve their computer and poll numbers even more.  It is highly unlikely they'd pass Alabama should they win the SEC, but it would be very interesting if it game down to Oregon or Georgia in this situation.

Kansas State is just too far back right now and doesn't get a title game, so they will be on the outside looking in.

Florida State?  They can make a statement by beating Florida, but then they get Georgia Tech in the ACC title game which won't help and the computers already (justifiably) hate them so even with a win over Florida, won't be in the running.

So, we can probably expect a Notre Dame vs Alabama/Georgia BCS title game, especially with Matt Barkley out for USC giving Notre Dame a great shot to win that game, but stranger things have happened.  If it isn't Alabama or Georgia, Oregon is probably next in line but they definitely need help. But should they get it, it would be very close and make for some BCS drama in 13 days.

Sunday, November 18, 2012

Comparing the 1-loss Contenders - Alabama, Florida, Georgia, Oregon, Kansas State

So we are down to an unbeaten Notre Dame and a bunch of 1-loss teams that want to make their case for being the #2 team and make it to the BCS title game.  So let's take a look at each team, and the best way to do so is with my computer's performance charts.

We'll go in the order the teams will possibly be in tonight's BCS ratings.

First, Alabama.

Alabama has been pretty consistent, but in their most meaningful game lost to Texas A&M.  Their schedule has been reasonably tough, getting a bit stronger as they played more games within the SEC but has dropped to #32 after this week's game and will go down a bit more after Auburn.  Beating a likely 11-1 Georgia or Florida in the SEC title game would certainly bring that back up.

Next, Georgia.

Georgia has not be as consistent nor dominant as Alabama with several somewhat meaningful games well below their current rating.  They do benefit from losing earlier in the year and do have a decent schedule strength at #38.  In my opinion, Alabama is pretty clearly better than Georgia and my computer agrees having Alabama #2 and Georgia way down at #11 because of their loss and other not so great performances.

Next, Florida.

The next SEC team looks similar to Georgia on the surface, some great results but a loss and a couple below par games, but their loss was not as bad, their high is higher, and their schedule, even with playing Jacksonville State this week is stronger at #17.  For this reason, in my opinion, and my computer's analysis too, they are higher rated than Georgia but not as high as Alabama.

Next, we leave the SEC to look at Oregon.

Oregon is very similar to Alabama, good to great results and pretty consistent, but a glaring big red dot in their most meaningful game.  They are hurt by their loss coming a week later than Alabama's, but despite all the pundits saying otherwise, their schedule is about the same as Alabama's.  It wasn't early in the season as the light blue line indicates, but their schedule is mostly at the end and will improve with Oregon State and possibly UCLA.  For this reason, my computer has them slightly ahead of Alabama.  I personally can make a case either way, it is very close.

Next, Kansas State.

Kansas State looks similar to Alabama and Oregon, but their highs were slightly higher consistently mid-season and that garnered them the #1 spot in my computer for several weeks.  However, their loss is noticeably worse than Alabama's or Oregon's and frankly Florida's too, and that pulls them down to #5 in my computer.  But their schedule is tougher than Alabama and Oregon's at #21, similar to Florida's.  The result is a rating still better than Georgia by a significant margin because they didn't have the close wins Georgia has had.

What do you think?  The media consensus seems to be Alabama is the clear #2 but the data and charts above would seem to indicate it isn't as clear cut as they say.  That said, the two teams I'd argue should be considered instead, Oregon and Florida, may not get to play championship games and if so will likely lose out by not being forefront in pollster's minds in 2 weeks.  Worse, the team that would play Alabama would be enhancing their resume by beating them and while I may not think Georgia is as good as Oregon or Florida, they'd likely get the nod.

Analyzing the 11/18/12 AP Poll - Notre Dame clear #1, but then what?

Here is week 12 of analyzing the AP poll.  Thanks as always to the AP publishing their voting.

All 12 charts are displayed below to make it easy to see the changes.

Where there was clarity before, there is confusion now.  Notre Dame is the clear #1 according to the writers garnering all the 1st place votes, but after that it isn't as clear.  Alabama is #2 but they get votes as low as #6!  Georgia ranges from 2nd to 8th, and Ohio State ranges from 2nd to 14th!  Oregon doesn't get any 2nd place votes and is as low as 10th, but is ahead of Florida who does get as high as 2nd but as low as 12th.  South Carolina is #13 overall but gets a 4th place vote.  Opinions certainly vary!

So, what about some of these on the edge votes?

One voter that has Alabama 6th has Ohio State, Georgia, Oregon, and Florida ahead of them.  The other has Texas A&M ahead instead of Florida.  A voter that has them 5th has Georgia, Florida, and Texas A&M ahead.

Ohio State is 2nd on 20 ballots, but 10th or lower on 5.  Florida is 2nd on one ballot by the voter from South Bend and he also has South Carolina 4th ahead of Kansas State, Oregon, and Georgia.

The Coaches and Harris polls will be interesting.

Will the SEC East play the role of last years SEC West in the BCS?

Last year, we had Alabama and LSU play during the regular season, then with other teams falling by the wayside, the BCS gave us a rematch in the championship game.  Could we have the same thing happen with the SEC again this year, just with the SEC East instead?  It is certainly possible.

Florida and Georgia played the last week of October with Georgia winning 17-9.  Both teams have won out since and could quite possibly be #3 and #4 in the BCS after the rankings come out later today.  Should Notre Dame lose to USC (possible, although perhaps unlikely should Barkley be out), then they move up to #2 and #3 and should Georgia beat Alabama in the SEC championship, you likely have Georgia #1 and Florida #2 in the BCS standings.

Now, a lot has to happen, Georgia beating Georgia Tech, Florida beating FSU, Notre Dame losing, and Georgia beating Alabama, but it is certainly possible.  And after what happened yesterday we can never say the unlikely won't happen.

Of course, we could have the more likely scenario of Alabama beating Georgia and Alabama and Florida being the two SEC teams (my computer at least likes Alabama a lot more than Georgia), but I found it interesting that a reasonable scenario exists where two SEC East teams play for the title rather than two SEC West as we did last year.

Of course, the gotcha in this will be how far K-State and more importantly Oregon fall in the computers and polls.  Both teams have a chance to recover a bit with Texas and Oregon State next week, and if Oregon gets help from UCLA they'd play at 10-2 UCLA in the Pac-12 title game which would also help them in the computers, so don't write them off yet.

College Football Week 12 Season Projections - Only Ohio State Undefeated?

The projected records for the regular season using results through week 12 are now available.  These projections now include projected conference championship games.

And then there was 1!  And it isn't even one of the teams from last week.  Instead, Oregon and K-State fall by the wayside, and Ohio State by virtue of winning at Wisconsin is now projected to finish at 12-0.  Notre Dame's chances at being undefeated go up to 46% with their win over Wake Forest, but more importantly USC looking a little weaker than they did.  Still, that game is at the Coliseum so my computer still favors the Trojans.  It of course can't factor in Barkley's availability.

The teams with one or fewer losses below, full list on the site.

TeamRecord% chance% +1% -1
TeamRecord% chance% +1% -1
Ohio State12-
Notre Dame11-153.946.10.0
Kansas St11-
Northern Illinois12-

College Football Week 12 Ratings and Rankings - Oregon still #1?

This week's ratings and rankings are now posted.

And after "what happens in November" happened, who is #1?  My computer says Oregon!

Oregon had such a big lead in the ratings last week (5.5 points over Alabama) and the team right behind them K-State also lost, so even though they lost to Stanford and gave up nearly 4.5 ratings points, their rating is still ahead of Alabama's.  Alabama was not helped by playing Western Carolina, and won't really be helped by playing Auburn next week while Oregon will be helped by playing Oregon State, so unless Oregon continues to falter, they could very well stay at the top of the heap in my computer.

What about Notre Dame?  They won convincingly for once, but against a very weak opponent so simply did what they were supposed to do.  Their rating improved ever so slightly and they move up 2 spots, but that was more due to LSU and Oklahoma making their games closer than they should have been and falling.

The SEC does have 4 teams in my top-10, but the Pac-12 now has 3 as well.

It is interesting to note that Ohio State is undefeated and would ordinarily be getting lots of BCS love and national attention, but because they aren't eligible for the post season they are being largely ignored.  My computer says they should be ignored just based on results too as despite being 11-0, are ranked only #14, due in large part to clearly having the weakest schedule of any team in the top-15.

The BCS compatible algorithm ratings and rankings do have Notre Dame now on top and they are followed by Florida and Ohio State, with Alabama down at #4.  Oregon and K-State fall all the way to #7 and #8 respectively.

The top-25 is below, full ratings on the site.

1Oregon92.71010-172.007+0, -4.429
2Alabama91.26410-172.030+1, -0.396
3Texas A&M87.3609-273.808+2, -0.418
4Florida87.04810-174.442+0, -0.919
5Kansas St87.02810-173.706-3, -5.794
6Notre Dame86.45011-072.749+2, +0.292
7Oregon St86.2438-277.567+5, +1.022
8LSU85.6349-273.696-1, -1.202
9Stanford85.5279-277.113+4, +2.240
10Oklahoma85.4118-272.875-4, -1.639
11Georgia85.28110-171.269-1, -0.451
12South Carolina84.3059-272.255-1, -1.027
13Southern Cal84.2247-475.594-4, -1.787
14Ohio State83.79811-069.212+3, +1.441
15Oklahoma St83.7517-370.975-1, +1.024
16Florida St83.34310-163.571-1, +0.646
17Texas82.7318-274.297-1, +0.093
18Utah St82.4409-267.703+1, +0.734
19Arizona82.3017-475.533+1, +0.927
20Clemson82.29810-167.080-2, +0.409
21Nebraska81.5419-270.585+0, +1.253
22UCLA81.4229-273.352+0, +1.685
23Baylor79.4255-575.871+16, +3.676
24Michigan79.3468-371.466+4, +1.119
25Arizona St79.2866-573.490-2, -0.324

Wednesday, November 14, 2012

What matches count towards your USTA year-end rating?

The end of November is coming and with it the USTA releases their year-end ratings.  These are important because it determines what level you can play at in 2013.  Some people want to get bumped up to achieve a goal or increase the challenge, others want to get bumped down so they can play on or assemble a team with championship aspirations.

I've done quite a bit of research and tinkering the past few years (my wife says far too much) and can calculate pretty accurate Estimated Dynamic NTRP Ratings and generate a pretty informative report.  But to do this requires analyzing the proper matches as the USTA does not include every match you play in calculating your rating.  In fact, each section is given some leeway in determining what to include.

There are a few givens though.

If you play enough gender specific matches (I believe it needs to be 3 or more), any mixed matches you play are not included.  If you play mixed but not enough gender specific matches, your mixed matches will be used to calculate a Mixed Exclusive rating.

Then I believe that all gender specific leagues (regular leagues) that have National Championships are included.  This means that your Adult, Senior, etc. leagues that are usually played in the spring (some are early start and begin in the prior fall or winter) with playoffs/districts/sectionals mid-year and nationals in the fall count.

But then a section gets to choose.  For example, In the Pacific Northwest where I play, tournament matches are not included, but a summer "One Doubles" league is.  In the Southern section I'm told that only the regular leagues are included, no Combo or Tri-Level or Singles league matches count.  Inter-Mountain does include tournament matches.  So it varies.

But I'd love to build up a list of what each section counts, so please leave a comment or contact me to let me know what the rules are for your section.

Monday, November 12, 2012

NFL Week 10 Ratings and Rankings - Houston stays on top

The ratings after week 10 are now posted and listed below as well.

Houston stays #1 and Denver moves up to #2 with Green Bay close behind.  Atlanta, now with a loss falls to #6.

1Houston90.4388-181.044+0, +1.298
2Denver88.2966-382.567+3, +1.158
3Green Bay87.4146-381.856+1, +0.131
4New England87.2716-381.257+3, +0.189
5Chicago87.0777-280.715-3, -0.857
6Atlanta85.9958-179.374-3, -1.446
7San Francisco85.6756-2-181.168-1, -1.441
8Seattle84.9126-482.287+1, +0.564
9Baltimore83.5217-278.684+4, +1.400
10NY Giants82.8736-480.415-2, -2.779
11Pittsburgh82.5196-378.913-1, -1.047
12Tampa Bay82.4055-478.790+2, +0.382
13Dallas82.3794-583.726-2, +0.018
14Minnesota81.3636-479.455+4, +0.928
15St Louis80.7673-5-183.075+4, +0.562
16New Orleans80.4594-580.739+6, +1.116
17Detroit80.3324-581.679-2, -0.883
18San Diego79.7804-578.898+2, +0.050
19Carolina79.6892-783.424-3, -1.367
20Miami79.5224-580.091-8, -2.685
21Washington79.4243-682.036+0, -0.150
22Cincinnati79.2574-579.587+3, +2.057
23Arizona79.0484-581.888+0, -0.287
24NY Jets78.9333-682.599-7, -1.545
25Indianapolis78.6596-378.774-1, +0.895
26Buffalo77.8723-682.380+0, +0.867
27Tennessee77.0944-682.483+3, +2.655
28Cleveland75.2332-779.659+1, +0.597
29Philadelphia75.1943-681.040-2, -1.325
30Oakland74.6923-681.017-2, -1.156
31Jacksonville72.6641-881.655+0, -1.167
32Kansas City71.4541-881.114+0, +1.195

Sunday, November 11, 2012

New chart for showing projected records

I had an aha moment today with an idea of how to visually display the projected record my computer has for NFL teams.  For each team, I calculate the chances of each possible record, and then list the most likely record as the projected record along with the chance of finishing one game better or worse.

But that doesn't tell the whole story.  If you graph an individual teams possible projected records, you get something of a bell curve with the most likely record in the middle.  But it would be a pain to have 32 different charts to have to look at, so I've come up with a way to display all for a given conference on a single chart.

It is a little overwhelming at first, but take a look at this and let me know what you think.

What this is showing is a segment for each team for each possible win total indicating the likelihood of that record.  The size of the segment indicates a higher likelihood.

So for example, Atlanta in blue has a small chance of going 15-1 and a smaller chance of 10-6, the most likely record in between being 13-3.  Then Chicago is in red, the 49ers in orange/yellow, etc.

I used a stack bar chart to be able to fit it all on a single chart so it is a little complex, but I think it is somewhat useful.  Let me know what you think.

Oh, and the above is showing the current projected records after the Sunday night game.

Will we have a BCS title game between undefeated teams? Or can the SEC find their way back in?

With Alabama's loss, we are down to 3 undefeated teams vying for participating in the BCS title game; Oregon, Kansas State, and Notre Dame.  There are only a few games left, so there is a good chance 2 will stay that way, but what are the exact chances of that occurring?

The 3 teams in question have the following chances of finishing perfect:
  • Oregon - 81.7%
  • Kansas State - 78.7%
  • Notre Dame - 37.8%
It is safe to assume that any 2 undefeated results in them facing each other.  It is also probably safe to assume that if Oregon and K-State do it, Notre Dame gets left out.

So, there is a 64.3% chance of Oregon and Kansas meeting as undefeated teams.  But if one of them falters, Notre Dame could fill in.  Running the different permutations results in a 75.7% chance of some combination of these teams facing each other as undefeated teams.

If you look at that number, there is a good gap between it and 100%.  So what happens then?  It is probable that should the SEC champ finish with 1-loss, they would be the choice if 2 of the aforementioned teams have a loss.  So what are the chances of a 1-loss SEC champ?

Alabama has the best chance at 70.4%, but Georgia has a 27.5% chance too, and should Georgia falter and let Florida in the title game, they have a 20% chance.  If we put these permutations together, we have a 99.2% chance of a 1-loss SEC champ.

That means that there is still a 24.1% chance of an SEC team making it to the BCS title game.  Trust me, stranger things have happened.

College Football Week 11 Season Projections - Only 2 projected to be undefeated

The projected records for the regular season using results through week 11 are now available.  These projections now include projected conference championship games.

And then there were 2!  With Alabama's loss, and Notre Dame rating dropping a bit, just Oregon and Kansas State are projected to finish the season undefeated.  And both teams are definitely projected to do so at 82% and 79%.  This puts the chances of both occurring at 64% so a very good chance we have a match up of undefeated teams in the BCS Championship game.

Notre Dame still has a chance at a perfect season at 36%, but a 1-loss season is more likely at 64% due to having to visit USC.

Ohio State actually has a better shot at an undefeated season at 38%, but the computer still says 11-1 is more likely at 51%.

The teams with one or fewer losses below, full list on the site.

TeamRecord% chance% +1% -1
TeamRecord% chance% +1% -1
Kansas St12-
Ohio State11-150.637.811.6
Notre Dame11-163.535.90.6
Northern Illinois12-