PDA

View Full Version : Hockey East - Who's in, who's out, who's home: by the numbers - 2009-10 edition



Pages : [1] 2 3 4 5 6 7

Todd
02-05-2010, 01:46 AM
--- Home Lock - 38 (UMA/UML) ---
--- In - 34 (NU) ---
UNH 27 - 45 [1-9]
UMA 22 - 38 [1-10]
BC 22 - 40 [1-10]
ME 20 - 40 [1-10]
UML 18 - 38 [1-10]
BU 16 - 34 [1-10]
NU 15 - 33 [1-10]
UVM 15 - 35 [1-10]
MC 12 - 34 [1-10]
PC 9 - 27 [2-10]
--- Home Eligible - 22 (ME/UML/BU) ---
--- Out - 15 (NU/UVM) ---

Remaining LEAGUE schedules:
UNH - @MEx2, PC, UVMx2, @NU/NU, BC/@BC
UMA - BC, NU/@NU, @MC/MC, @BC, @MEx2
BC - @UMA, @UML/UML, @NU/NU, MC, UMA, @UNH/UNH
ME – UNHx2, @BUx2, UMLx2, @MCx2, UMAx2
UML - BU, @MC, BC/@BC, @MEx2, PC/@PC, UVMx2
BU - @UML, MEx2, @PC/PC, @UVMx2, NU/@NU
NU - @MC, @UMA/UMA, BC/@BC, UNH/@UNH, @BU/BU
UVM - @PCx2, MCx2, @UNHx2, BUx2, @UMLx2
MC - NU, UML, @UVMx2, UMA/@UMA, @BC, MEx2, @PC/PC
PC – UVMx2, @UNH, BU/@BU, @UML/UML, MC/@MC


Around this time of year for the past few years, I've started a thread that lays out the math behind the playoffs in Hockey East. The impetus for starting the thread each year was when someone clinched... something. These are all manual calculations, so, as always, if anyone notices an error or oversight, feel free to let me know.

Well it’s about that time, so... where are we? Has anyone clinched anything?

In contrast to last year, when the kick-off post had everyone still eligible for every slot from 1-10, ... there is at least something settled, if not clinched. MC can no longer catch NU.

In other ways, this kick-off post is remarkably similar to last year’s. While it seems like UNH took 1st place by running away and hiding with it last year, as of the first post they were only three points up.

So that was the kickoff to last year's thread. A lot of where we are now is similar, but this year, it is PC that can no longer pass UNH as of last weekend (they can tie, but UNH has the H2H tb, and it's too far out to do the math on multi-team tbs that will likely be irrelevant by next weekend). Instead of being three points up, as they were two years ago, this year's kickoff post has UNH up by five.

One thing of note this year that is unusual is that with the standings so close, but so many variations in the games in hand, we have a lot of counterintuitive cross-overs between points in the bank (minimums) and points available (maximums). For example, MC is in 9th with 12 points, but they have the same max as 6th place BU (34) and actually have a higher max than tied-for-7th NU (33), which is why NU sets the "In" benchmark for the moment. Similarly, 4th place Maine has a higher max (40)than tied-for-2nd UMA (38). Much of that should disappear after this weekend's irregular Beanpot-accommodating schedule wraps up.


Explanatory notes:

For those that haven't read one of these threads before, I'm basically following a couple of key benchmarks and comparing each team's min and max points to see where they might fall. As results come in, the benchmarks themselves change and the teams above or below them change.

The keys are:
"In" - what point threshold needs to be reached to guarantee a playoff spot (can't fall to 9th or 10th).
"Home Eligible" - can still reach home ice (4th seed),
"Home Lock" - can't fall below 4th,
"Home Ice" - once the top four spots are settled,
and "Out" - can no longer reach 8th.

If you're below a line, your minimum (points already earned) has not yet reached that benchmark, or in the case of "Out", cannot. If you are above a line, your maximum keeps you in play. At this point, no one is yet In or, of course, locked up Home Ice. Conversely, no one is yet Out and everyone is still Home Eligible.

As results come in, In and Out will draw closer together and, at some point, merge once the top eight (or bottom two) are set in stone. The same is true of Home Eligible and Home Lock, which will merge into Home Ice once the top four are set.


Here's an example of how this works:

With a variable number of games remaining on the schedule for each team, the two lowest maximums are 27 (PC) and 33 (NU), then 34 points puts you in the playoffs. That puts the "In" benchmark at 34 (the last two years used 32 at this point). As the teams at the bottom fail to get points, their maximum falls, and the benchmark drops with it. If NU wins, but MC, BU and/or UVM (34, 34, 35 max, respectively) don't, once their maxes cross over, the benchmark moves over to the lower maximum, as those are points that team can no longer get. At the same time, as teams win, their minimum will increase, as those points are already in the bank.

If PC gets swept this weekend, that would put their max at 23 (9 pts earned with 7 games – or 14 points – left undetermined). With BC and UMA facing off on Friday and both teams at 22 going in, that would put at least one of those teams out of reach of PC even if PC won every remaining game. Since both BC and UMA own the tie-breaker over PC, a UMA/BC tie would put both teams at 23 and unpassable by PC. Since Maine (20 pts) also has the tb w/ PC, taking 3+ points this weekend would also put them ahead of a swept PC, meaning that PC would no longer be eligible for Home Ice and would drop below the Home Eligible line.

With this many games to go, I haven't factored in the remaining schedule to the benchmarks and maximum points too deeply yet. ME, BU and UML all play each other in every combination (each plays the other two), so even though Maine only has 20, the Home Eligible mark is 22 because of the schedule among those three meaning that at least one of them has to hit that mark. However, the bottom five still have to play each other quite a bit, and I did not factor in that interaction into the other benchmarks. When the permutations are fewer and those fine points become more relevant, I'm sure they'll get added in either by myself or fellow posters.

SPCSMAN
02-05-2010, 07:47 AM
Todd, thanks for taking the time to do this....I enjoy your analysis and look forward to it each year. :)

Todd
02-06-2010, 01:37 AM
--- Home Lock - 36 (UMA/UML) ---
--- In - 32 (MC) ---
UNH 27 - 43 [1-9]
BC 24 - 40 [1-10]
UMA 22 - 36 [1-10]
ME 22 - 40 [1-10]
BU 18 - 34 [1-10]
UML 18 - 36 [1-10]
NU 17 - 33 [1-10]
UVM 17 - 35 [1-10]
MC 12 - 32 [1-10]
PC 9 - 25 [3-10]
--- Home Eligible - 22 (ME/UML/BU/...) ---
--- Out - 17 (NU/UVM) ---

Remaining LEAGUE schedules:
UNH - @ME, PC, UVMx2, @NU/NU, BC/@BC
UMA - NU/@NU, @MC/MC, @BC, @MEx2
BC - @UML/UML, @NU/NU, MC, UMA, @UNH/UNH
ME – UNH, @BUx2, UMLx2, @MCx2, UMAx2
UML - @MC, BC/@BC, @MEx2, PC/@PC, UVMx2
BU - MEx2, @PC/PC, @UVMx2, NU/@NU
NU - @UMA/UMA, BC/@BC, UNH/@UNH, @BU/BU
UVM - @PC, MCx2, @UNHx2, BUx2, @UMLx2
MC - UML, @UVMx2, UMA/@UMA, @BC, MEx2, @PC/PC
PC – UVM, @UNH, BU/@BU, @UML/UML, MC/@MC

This weekend, the league's schedule kind of breaks into three segments. The top four play each other (1 goes to 4, 2t v 2t), 5 hosts 6 (and then 9) and the bottom four match up.

BC's win over UMA not only breaks the tie for second in the Eagles' favor, but also gives them the season series and tb since they also won the prior meeting. BC also has a game-in-hand advantage over the Minutemen.

Maine's win over visiting UNH tightens things up for the top four and also splits their season series (1-1-0) with the rubber match on tap for Saturday.

BU took the deciding third game at UML to knot the pair in a tie for 5th (UML has a game-in-hand). For the fifth time in the last six games, it was a one-goal difference, so if these two somehow meet again in the playoffs, it should be tightly-contested.

Where does the Home Eligible benchmark stand? Well, I've been able to arrange a six-way tie for 4th at 22, with the 10th place team a point behind at 21. In that case, that benchmark stays as is for another night at least.

With tonight's wins by NU and UVM, if the regular season were to end right now, the four road teams would all be within a single point of each other.

MC's loss (combined with NU's win) drops their max below the Huskies' and drops the In line to 32. UMA and UML losing drops the Home Lock benchmark as well.

With tonight's BC win and PC loss and the remaining schedule, I can no longer get PC into 2nd place, but I can get the Friars into 3rd with 25 points - in front of a seven-way tie at 24. BC can still end up 10th, in a number of ways, so no one else's seeding range moves tonight.

Todd
02-07-2010, 02:17 AM
--- Home Lock - 36 (UMA/ME) ---
--- In - 32 (MC) ---
UNH 27 - 41 [1-9]
BC 24 - 40 [1-9]
ME 24 - 40 [1-9]
UMA 22 - 36 [1-10]
BU 18 - 34 [1-10]
UML 18 - 34 [1-10]
UVM 18 - 34 [1-10]
NU 17 - 33 [1-10]
MC 14 - 32 [1-10]
PC 10 - 24 [4-10]
--- Home Eligible - 23 (UMA/UML/BU/...) ---
--- Out - 17 (NU) ---

Remaining LEAGUE schedules:
UNH - PC, UVMx2, @NU/NU, BC/@BC
BC - @UML/UML, @NU/NU, MC, UMA, @UNH/UNH
ME – @BUx2, UMLx2, @MCx2, UMAx2
UMA - NU/@NU, @MC/MC, @BC, @MEx2
BU - MEx2, @PC/PC, @UVMx2, NU/@NU
UML - BC/@BC, @MEx2, PC/@PC, UVMx2
UVM - MCx2, @UNHx2, BUx2, @UMLx2
NU - @UMA/UMA, BC/@BC, UNH/@UNH, @BU/BU
MC - @UVMx2, UMA/@UMA, @BC, MEx2, @PC/PC
PC – @UNH, BU/@BU, @UML/UML, MC/@MC

With the PC/UVM tie, PC's max is now 24. Losing both tbs to BC and Maine, PC can no longer reach 3rd and both BC and Maine can no longer fall to 10th. PC could still end up in 4th ahead of a logjam a point behind, so they are still Home Eligible by a slim margin.

Part of the issue for the Friars is that they already have played 20 league games, where most of the teams they're chasing have an extra game remaining. In fact, MC, the team directly ahead of PC, has two games in hand in addition to a four point lead on them. The final weekend of the season could put either of these teams in a position to bury the other, with PC making up the four-point gap and taking the tb, or the Warriors could keep or expand their lead. Both teams are hoping that they do well enough in the interim that the series matters for more than avoiding last place.

In Merrimack's favor for moving up in the standings are two things: a win in their game in hand would currently narrow the gap on 8th place to a single point while leaving them only two behind the tie for 5th, and four of their remaining games allow them to take advantage of their astounding record tis season at Volpe - where they are 10-2. Of course, the flip side of that schedule is that 5 of the games left are on the road, where they are, comparably astoundingly, winless (0-12).

The other impact of the tie at Schneider, combined with UML's loss and BU and NU off and awaiting the Beanpot, is that the four road teams as of today are a Huskies point away from a four-way tie. Looking at the final two weekends, BU and UVM will match-up for a Fri-Sun pair at the Gut before leading into the final week where the 'Cats come south to Tsongas for two and BU and NU face off in a home-and-home. If none of this group are able to overcome the four-(or five-)point deficit on home ice, it could be interesting to see the jockeying for position on quarterfinal road trips and match-ups that final weekend.

The top four had an interesting weekend primarily due to Maine's sweep reeling UNH back in to the pack, while also leapfrogging the Black Bears over the loser of Friday's BC/UMA game (UMA) and into a tie with that game's winner (BC). The final weekend may also prove pivotal here as UMA wil travel to Orono and BC and UNH will battle home-and-home.

Lastly, as noted previously, this weekend's balancing out of most of the games in hand leaves us a more intuitive arrangement of min-max ranges than we had leading in to the weekend.

sterlippo1
02-07-2010, 03:13 AM
kudos! great stuff

WildShawn
02-08-2010, 10:46 AM
Thanks again Todd!

JB
02-08-2010, 01:33 PM
Some other interesting number bits about the remaining schedule. All my numbers are league games only so MC is only 7-2 at home as there 3 other wins get pitched. Likewise UNH is 6-0-2 at home as the games against Miami and Cornell get tossed. I did this because of the disparity of the schedule UNH has Miami and Cornell removed and MC has Holy Cross, Army and Connecticut removed. I did this to try and compare apples to apples down the stretch.

H.E. Road in (Average Opponent Winning Percentage, In League Record only):
BU 45.1%
Maine 47.2%
PC 47.8%
UNH 47.9%
MC 49.4%
UML 49.7%
UVM 50.3%
UMA 51.0%
BC 51.6%
NU 58.3%

Winning % at Home (games remaining)
UNH 88% (5)
Maine 85% (4)
MC 78% (4)
BC 75% (5)
NU 60% (4)
UML 60% (4)
BU 56% (4)
UMA 55% (2)
UVM 44% (4)
PC 36% (3)

I find it interesting that all but two teams are better than .500% at home. Just shows how hard road points have been to come by this year. Also it notes home ice in the quarters should be huge.

JB
02-08-2010, 01:49 PM
So with all that non sense, and a bunch more at that, I have a prediction on the finish.

UNH 37
BC 33 (tb)
Maine 33
UMA 27
BU 25 (tb)
UML 25
UVM 24 (tb)
NU 24
MC 23
PC 15

This is a tough prediction; 5-9 are separated by 1 win. Like stocks past performances do not predict future results, so yikes. Take UNH - macro looks pretty good, but micro 3-3 in the last 6 with two wins in OT and now has a habit of blowing 2 goal leads. 5 times in the last 6 games up by 2 goals and came out of those games with: 1 W, 2 OT W’s, and 2 L’s.

I am still trying to find a better way to represent opponents, right now I look at opponents winning% and "your" winning% and take into account “your” Home wining %. Take UNH again 5 games remaining at home were they win 88% and 2 on the road @BC (75% at home) and @ NU (60% at home).

Thing is I only find this interesting about 4 weeks a year…

ScottK
02-08-2010, 03:28 PM
So with all that non sense, and a bunch more at that, I have a prediction on the finish.

UNH 37
BC 33 (tb)
Maine 33
UMA 27
BU 25 (tb)
UML 25
UVM 24 (tb)
NU 24
MC 23
PC 15



Your point numbers need adjusting. You have the teams getting a total of 74 points the rest of the way when there are 39 league games left. Those 4 other points should make a different in the standings.

JB
02-08-2010, 04:36 PM
Your point numbers need adjusting. You have the teams getting a total of 74 points the rest of the way when there are 39 league games left. Those 4 other points should make a different in the standings.

I know. I rounded to the nearest whole number. I didn't go game by game I took more of a marco average opponent point of view. The little program I wrote will assign .12876112356 of a point. The order was surprising that the tie breaks and the points all worked out (spooky actually). The lost points are consumed in the fractional points. Also suprising is all rounded down.

Probably should have commented on this with the rest of the non sense.

Todd
02-08-2010, 04:43 PM
I know. I rounded to the nearest whole number. I didn't go game by game I took more of a marco average opponent point of view. The little program I wrote will assign .12876112356 of a point. The order was surprising that the tie breaks and the points all worked out (spooky actually). The lost points are consumed in the fractional points. Also suprising is all rounded down.

Probably should have commented on this with the rest of the non sense.
"The little program I wrote will assign .12876112356 of a point." For what?

So in an example where two teams play 10 games and, going into the series, Team A wins 60% of the time and Team B wins 40%, do you:
A) Assign 12 points to Team A and 8 to team B by win percentage, or
B) Assign every game to Team A because Team A is more likely to win each individual game by a 3:2 ratio?

bottomdweller
02-08-2010, 07:53 PM
So with all that non sense, and a bunch more at that, I have a prediction on the finish.

UNH 37
BC 33 (tb)
Maine 33
UMA 27
BU 25 (tb)
UML 25
UVM 24 (tb)
NU 24
MC 23
PC 15

This is a tough prediction; 5-9 are separated by 1 win. Like stocks past performances do not predict future results, so yikes. Take UNH - macro looks pretty good, but micro 3-3 in the last 6 with two wins in OT and now has a habit of blowing 2 goal leads. 5 times in the last 6 games up by 2 goals and came out of those games with: 1 W, 2 OT W’s, and 2 L’s.

I am still trying to find a better way to represent opponents, right now I look at opponents winning% and "your" winning% and take into account “your” Home wining %. Take UNH again 5 games remaining at home were they win 88% and 2 on the road @BC (75% at home) and @ NU (60% at home).

Thing is I only find this interesting about 4 weeks a year…

Sending Lowell to Orono makes me very angry. You're going to have to adjust your points so that does not happen...

UMLWeatherman
02-09-2010, 04:28 AM
Sending Lowell to Orono makes me very angry. You're going to have to adjust your points so that does not happen...

Yes yes .. please just move 1 pt here or there ...

sandiegoblkbr
02-09-2010, 05:21 AM
As far as the Top 3 go.....remaining games and final standings predictions:

1. Maine (7-1.....split with BU) 38 points
2. BC (6-2.....split with Lowell and UNH) 36 points
3. UNH (4-3....split with Vermont, NU and BC) 35 points

:)

btb91
02-09-2010, 07:09 AM
As far as the Top 3 go.....remaining games and final standings predictions:

1. Maine (7-1.....split with BU) 38 points
2. BC (6-2.....split with Lowell and UNH) 36 points
3. UNH (4-3....split with Vermont, NU and BC) 35 points

:)

I can see this happening.

JB
02-09-2010, 09:33 AM
"The little program I wrote will assign .12876112356 of a point." For what?

So in an example where two teams play 10 games and, going into the series, Team A wins 60% of the time and Team B wins 40%, do you:
A) Assign 12 points to Team A and 8 to team B by win percentage, or
B) Assign every game to Team A because Team A is more likely to win each individual game by a 3:2 ratio?

In my "analysis" (or shot in the dark if you prefer) nothing is based on a game or a match-up, it is all averages.

It all started with who has the toughest road in, UNH for example plays PC(10th), UVM(7th), NU(8th) and BC(2nd). On paper that doesn't appear to be a tough end of the year in reality it is the 4th easiest of the 10. I am still playing with strength of the remaining schedule. Right now I just look at opponents record but haven't factored home road splits. Again for UNH 5 home to 2 road seems friendly.

So looking at the average opponent left toss in home vs road records of each specific team, I figured out the average number of points each specific team would get. In this "method" (word is probably bit strong, but a dart board is a method) nobody get 2 points from any game because no team has a perfect record (overall, home, road etc).

Clear as mud?

Priceless
02-09-2010, 11:09 AM
Northeastern 27 33
Merrimack 27 32
Mass.-Lowell 27 30
Vermont 27 29
Boston College 27 28
Massachusetts 27 28
Boston University 27 27
New Hampshire 27 27
Maine 27 24
Providence 27 12

BUT09
02-09-2010, 11:14 AM
Northeastern 27 33
Merrimack 27 32
Mass.-Lowell 27 30
Vermont 27 29
Boston College 27 28
Massachusetts 27 28
Boston University 27 27
New Hampshire 27 27
Maine 27 24
Providence 27 12

You have two of the top three dropping to 8th and 9th? This seems a little strange.

Priceless
02-09-2010, 11:20 AM
You have two of the top three dropping to 8th and 9th? This seems a little strange.

You're right. This seems more likely.

Merrimack 27 32
Northeastern 27 31
Mass.-Lowell 27 30
Massachusetts 27 30
Vermont 27 29
Boston College 27 28
Boston University 27 27
New Hampshire 27 27
Maine 27 24
Providence 27 12

Congrats to Merrimack!

Todd
02-13-2010, 04:44 AM
--- Home Lock - 34 (UMA/BU/UML) ---
--- In - 31 (MC) ---
UNH 29 - 41 [1-9]
BC 24 - 38 [1-9]
ME 24 - 38 [1-9]
UMA 22 - 34 [1-9]
BU 20 - 34 [1-10]
UML 20 - 34 [1-10]
NU 19 - 33 [1-10]
UVM 19 - 33 [1-10]
MC 15 - 31 [1-10]
--- Home Eligible - 23 (UMA/UML/BU/...) ---
PC 10 - 22 [6-10]
--- Out - 19 (NU/UVM) ---

Remaining LEAGUE schedules:
UNH - UVMx2, @NU/NU, BC/@BC
BC - UML, @NU/NU, MC, UMA, @UNH/UNH
ME – @BU, UMLx2, @MCx2, UMAx2
UMA - @NU, @MC/MC, @BC, @MEx2
BU - ME, @PC/PC, @UVMx2, NU/@NU
UML - @BC, @MEx2, PC/@PC, UVMx2
UVM - MC, @UNHx2, BUx2, @UMLx2
NU - UMA, BC/@BC, UNH/@UNH, @BU/BU
MC - @UVM, UMA/@UMA, @BC, MEx2, @PC/PC
PC – BU/@BU, @UML/UML, MC/@MC

This weekend features three series where current home ice teams (BC, Maine, UMA) are playing a pair with teams that would be on the road, if the playoffs were set on Thursday (UML, BU, NU, respectively). After Round One, those "road" teams are 3-0.

Meanwhile, in the only game between teams in the bottom half, Merrimack got its first road point of the season by matching UVM in the first of two at the Gut.

These four results take a race that was tight among the four "road" spots and draws them all closer to the three teams ahead of them for home ice. There are now only five points separating 2nd place from 8th with seven games to play (six for UMA).

In the final pair, 1st place UNH was knotted at 2 after 2 at home with last place PC before the game broke open on the strength of a 3rd period Hat Trick by Bobby Butler (who also had the first UNH goal). That's the only game of the weekend for both teams and restores the five-point lead that UNH had on 2nd place before being swept by Maine last weekend, while shrinking PC's chances of getting out of the cellar.

In fact, PC's lost opportunity combined with the remaining schedule leaves them 6th at best, with a tiebreaker at 22 and the last three teams a point behind at 21. The top four can no longer be caught by PC, and even though any of the lower teams could be caught and end up 10th, at least one of them has to also finish ahead of PC.

Merrimack, however, could still come in first, so most of the ranges are still fairly flexible.

The Home Eligible line stays at 23 (albeit in at least a four-way tie) for the time being. The same is true for Home Lock at 34, for now. "In" drops to 31 because MC only picked up one of the two available points on the night.