Now Available! EBook from CWM

Now Available! EBook from CWM
Order Now from Amazon

You can also get an epub copy


Tuesday, March 21, 2017

Some National Analysis

The Women's Worlds are chugging along and I haven't yet taken a Curl With Math look back at the Scotties final, much less the closing weekend of the Brier.  Not much time for preamble as I need to get editing another Curling Legends Podcast (check out Cathy King on the latest episode and Paul Savage is coming up next week), so without further ado...

Scotties Finals

An entertaining game that was an impressive win for Rachel Homan but an equally disappointing loss for Michelle Englot.  Manitoba played fantastic all week and deployed a sound strategy against Homan in all of their games, resulting in a 2-1 record against the favorites from Ontario, but that single loss took place in the one game that mattered most.  A few thoughts on their final contest...

2nd End:  Manitoba leads 1-0, Ontario has hammer, final shot

Homan is Yellow

Homan could draw for a single to tie things up.  But Rachel doesn't seem to hesitate, as I suspect she'd already drawn up this shot before the last rock had come to rest.  If we assume no worse than a steal of 1, at worst Rachel will head to the third end down 2 points, holding hammer, and retain at least an average Win Expectancy of 30%. Despite the two earlier loses that week to Manitoba, you could make the argument her odds are higher than average in this situation with a full 8 ends remaining. 

Win Expectancy (WE) Homan Takes 1 = 44%
WE Homan Takes 3 = 70%
WE Englot Steals 1 = 30%

There is still some chance of scoring 2, if the back yellow rock is pushed 6 inches, though a single seems unlikely.   Even if they never score two, if Rachel can make the shot for 3 at least 35%, or just over one in three attempts, then it's a good call.  I believe Rachel expects to make good contact with the back red stone well over half the time, so the call appears to be worth the risk.

5th End:  Homan up 3-2 with hammer, final shot

Rachel calls for a draw to the back four with potential to sweep it to bump Englot's shot stone for two.  It looked like she could pick it out and ensure a single.  Strange that she came well short, a pick or strange ice path, but not something you expect in the 5th end.

During the end, Vic Rauter mentions that Randy Ferbey started using the points system to classify weight. Actually, Ferbey fourth David Nedohin picked this concept up from Arnold Asham while playing with him years earlier.  Arnold is famous for inventing the red brick slider and starting Asham Curling Supplies, and I'm looking forward to recording a podcast with him soon to get the history of his many contributions to the sport.

7th End: 3-3, Homan with Hammer, final shot

Rachel chooses to come through a port at an attempt for 4 or 5 points, rather than draw for a single to carry a 1 up lead into the 8th. 

Homan is Yellow

As the photo indicates, it wasn't a simple pick shot because there was a risk of jamming and possibly scoring only 1.  I'm not certain their confidence in the ice on the shot (the higher second guard appeared to add to the difficulty), but it is an attempt to win the game right there, at a risk of only being down 1.  In the moment, I wasn't sure I liked the call, but that could have been swayed by the emotion of how the game seemed to be shifting to their opponent.  In a vacuum, it's the correct decision if Rachel makes it greater than 33% of the time.  An alternative shot that they may have considered was the runback.  It introduces a risk of double jam (and a steal of two) but I suspect Rachel's percent success with that shot is higher than the soft weight hit she attempted.

10th end:  6-4 Homan, Englot has Hammer, final shot
The thin double by Rachel to hold Englot to a deuce and force an extra end was close to the difficulty of the Hackner double.  Having said that, she had no need to stick her shooter.  Small margin for error and a great shot that ultimately won the Scotties (zoom ahead to 15:56

Game saving 10th end double at 15m56
Brier Semifinal

A tough loss for Team McEwen from Manitoba.  It reminded me of a similar defeat that Kevin Koe had in the 2013 Olympic Trials against Kevin Martin.  Up two with hammer and three ends to go, Koe surrendered a steal of two in the 8th end, was forced to one in the 9th and Martin scored a deuce to win in the 10th.  Koe never recovered at the Trials (from the 0-3 start) but he did at the Brier a few months later.

8th End: 5-3 McEwen, with Hammer, final shot

Mike could attempt an in-turn draw for 1 (or even a tap on the red stone in the top eight foot) but instead chooses a run-back attempt to double the yellow stones on the out-turn side...

McEwen is Red

Most posters on CurlingZone claimed McEwen trying the run-back rather than drawing for a single was a mistake.  I think there's an argument to be made for both decisions.  With a draw for one, McEwen goes up 6-3 (a 93% Win Expectancy).  But the draw was no picnic, as he needed full four foot and perfect draws appeared to be dicey in the later ends (check out Gushue's draw to the eight foot in the finals).  

The way Team McEwen were playing the shot, it appeared to be only for two but at this point 2, 3 or 4 is essentially the same result, a win (99.1% WE or greater).

The argument is Mike's confidence in making that angle run-back versus the draw.  It was full four foot on a specific path, and only he knows how comfortable he felt at that moment on that call.  Let's say it was 90%.  Straight runbacks are usually around 80-85%, in Mike's case possibly higher.  As you move towards an angle versus straight back, your odds are reduced (see my previous CWM article "Aggression is the Better Part of Valor").

If the draw is 90%

WE = .9 x .93 = .837
WE up 1 with hammer and three ends remain = .835

He's ok to try the runback 100% of the time, it's roughly a wash if he misses every time.

If the draw is 95%
WE = .95 x .94 = .894

Assuming no chance at steal of 2, Mike needs to make the raise for 2 pts only 35% of the time to be equal to the draw.  

If the draw is 100%, he needs to make the raise 65% of the time.  Having re-watched the shot, I think Mike usually makes that raise at least 60%, if not higher, and the draw is no gimme.   

Extra End:  Tied 6-6, McEwen with Hammer, Mike's first shot

Mike is facing this with his first rock...

McEwen is Red

Wozniak's hit earlier in the end landed in the worst possible spot, right behind the button.  With a peel, Koe could attempt a freeze and force Mike to blast and head to another extra end (or much less likely, Koe steals the win).  Mike seems convinced he wants to attempt a draw, getting in first and leaving Kevin no shot.  Mike comes up perhaps a foot light and curls perhaps 3 more inches than necessary, and Kevin has a sliver of hope.  Koe plants his rock around the McEwen stone and corner freezes, leaving Mike with a miracle double raise for the win, which he misses, and Koe goes on to the final.  
The decision comes down to what you believe are Koe's odds of making a freeze that will require a blank.  If you think Kevin makes it 80% of the time, and your Win Expectancy (WE) in an EE is 80%, Mike needs to have a better than 84% probability of winning right there with the draw.

If you think Kevin's freeze leads to a blank only 50% of the time, then Mike needs to be over 90% sure his draw will be successful.

If Kevin's freeze is successful 90%, Mike needs to be 82% sure of the draw.

This all assumes an 80% success rate when tied with hammer and one end remains.  Historically this was 75%, and remains close for all major events, but Slams trend up towards 79% or higher and the better teams are better at winning in this situation (see past CWM "Giving Away Points at the Canadian Open")

It's my assessment that the entire scenario of Mike's last shot was a brain fart. From the poor call decision, to the judging error on weight by the sweepers to the confusion with the line call (the rock was better to be open than buried short), it was a complete team mistake.  These things can happen in rushed moments after a long week and a grueling game (remember Kevin Martin's loss at the 2009 Worlds?).  They were close, and I suspect will be close again.  In any case, Destiny appeared to be working against everyone but Gushue the entire week, and you could speculate that a trip to the finals would have simply swapped their Bronze medal for Silver. Still sucks.

Brier Final

6th End: 5-1 Gushue, Koe with Hammer, Gushue's last shot

Gushue is Red...Before his last shot
After his last shot
Before playing his last shot, I was surprised Gushue didn't see the potential for Kevin's shot for 3. Or perhaps he did and was attempting to roll slightly across, instead of high side which he did, leaving the possible shot for three.  After Brad's shot Russ Howard even says "great shot" but in retrospect, it wasn't.  If Brad draws top four instead, Koe has no real choice but to draw for two.  Then again, it was a great shot by Koe that had very little margin for error. Unfortunate for Brad the rocks ended up where they did, but he should shoulder some responsibility.

7th End:  5-4 Gushue, with Hammer, Kennedy's last shot

On Marc Kennedy's last shot, Koe decides to put up a second centre guard. 

Koe is Yellow

It pulls Gushue into hitting to sit three, and provides cover of the four foot which eventually leads to a force.  If Koe instead hits to sit two, even with a roll he provides many options for Gushue to blank or even score two, and the single guard is too long to be much protection for later in the end.  Granted, Brad missed a runback on his first shot that could have changed the outcome as well.

9th End: 6-5 Gushue, Koe with Hammer, final shot

Good decision to not attempt the triple for three.  The shot really did not look be there and to take that chance here it needs to be automatic.  They had stolen a win the night before, no reason to assume there wasn't at least a 20% chance of it happening again.

Koe is Yellow

10th End: 6-6, Gushue with Hammer, final shot to win the Brier

In the last 15 years I don't recall having seen a draw to the eight foot that looked to be so difficult.  

Brier winning draw at 2h36m11s 
Certainly the enormity of the moment, and being under the watch of a large waiting-to-burst-into-cheers Newfoundland crowd, had some impact.  Brad even discussed afterwards that he was told it was 5 feet heavier than usual, and admitted to only throwing it two feet more.  But it also seemed more difficult than similar shots in other situations.  I'm interested to hear from players about the frosty and/or flat conditions that seemed to creep into some games during the event.  Hair brooms are no longer permitted and I wonder if limiting players to only the use of the new broom head may be detrimental in certain conditions, particularly late in the season when outside temperatures reach above zero with high humidity.  I'm not a meteorologist or even an ice technician, and I could be way off track here, but someone should investigate.  Just because we appear to have "solved" the broom problem from last season doesn't mean we haven't created a different one going forward.  I might vote for a return to a (albeit restricted model) of horse hair broom, but I'm not certain that's being considered.

Until next time...

No comments:

Post a Comment