A crunch of the numbers by a pretty sophisticated math formula shows the Senators have a 0.2 per cent chance of making the playoffs.
It was right about the time last week when I was reading the wildly speculative stories about the Ottawa Senators’ chances of making the playoffs this season that my good friend from sportsclubstats.com sent out his annual press release slapping me back into reality.
The Senators in the playoffs?
Sportsclubstats.com calculated -- and it is a sophisticated calculation -- Ottawa had about a 0.7 per cent chance for the postseason. Tracking through games of March 25 and including Buffalo's win over Florida and Ottawa's loss to Carolina, that number drops to 0.2 per cent.
Now, I never completely rule out anything until it's a mathematical impossibility and even then I usually wait a day or two because, well, I'm not that good at math.
But Ken Roberts, the mind behind sportsclubstats.com, is good with numbers and he's been doing it for years. His work is especially good come crunch time, the period we are in now and it works something like this: each night Roberts takes the latest results and simulates the rest of the season by randomly picking results for each remaining game. Though he uses the word "random," it is actually based on historical content. Roberts uses a weighted method that takes the opponents' records and home-ice advantage into account so the historically-better team is more likely to win. Employing a 50/50 method he gives each opponent an equal chance of winning (or tying if the sport allows it) in each game. He feeds it all into a computer and when it is finished "playing" all the remaining games it applies the league's tie-breaking rules to see where everyone finished. It repeats this random playing-out of the season millions of times, keeping track of how many "seasons" each team finishes where.
The result: a reasoned statistical analysis of what a team's odds are of making the playoffs; a projection as to how the daily remaining schedule will impact those odds followed by a projection as to how well a team has to finish in order to reach its goal.
Roberts, a software writer by trade and someone who has been bitten by the hockey bug ever since the Hartford Whalers became the Carolina Hurricanes, said the basic idea came from a model used by Baseball Prospectus.
In the case of the Senators, as of Thursday morning the calculations were, frankly, horrible.
Does that mean they can't do it? In a word, no. The way the Eastern Conference bottom contenders for the No.8 spot -- essentially Montreal, Carolina, Buffalo and Florida -- have been choking their way to the finish line, anything is still possible. But from the calculations, Wednesday night's 2-1 loss to the 'Canes was a killer for the Senators.
At 0.2 per cent even I can do the math.
The system is not foolproof and one would be wise not to use if for betting purposes, but Roberts argues it works pretty well.
"When (the computer) plays a game it takes home-field advantage and the 'strength' of each team into account," he said in an e-mail interview Wednesday. "The trick is finding an accurate value for 'strength.' I use a formula developed by Bill James (baseball) and refined for the NHL by Alan Ryder that is based on a team's goals for/goals against ratio.
Roberts says this gives a good baseline of the probabilities for saying "if the season played out in 30 million parallel universes the Oilers make the playoffs around five per cent of the time. It doesn't tell you what will happen this year. It doesn't know about trades or injuries. It doesn't know if a team has started believing in themselves."
It does, however, give some mathematical probabilities even with the so-called "three-point" game.
"I love the three-point game," Roberts said. "It makes the race harder to predict in your head, but the computer doesn't have any problems with it. It knows historically how often games go into OT."
After running the program, Roberts "slices and dices" the simulation data to show what happened the previous evening and who a fan should root for (or against) when tracking his favourite team's chances.
"It's part of what makes the site so addicting," Roberts said.
He's got a point there.
Tracking some of the Eastern contenders on Thursday morning, the Sabres' chances increased 4.3 per cent with the win over the Panthers, but they still have only a 12.6 per cent chance of making the postseason. The Panthers are at 23.5 per cent but the loss was big, dropping them some 15.4 per cent. Montreal did not play Wednesday, but the site shows the Habs chances went up 2.7 per cent to 80.3 per cent based on the impact of Buffalo's win, the Senators' loss and Pittsburgh's win over the now-struggling Calgary Flames.
The Flames, by the way, are still rated a 100 per cent lock to be in the playoffs.
The site also plots out the what-if scenarios of where each team would finish if it won all its remaining games, using a sliding scale based on points. That item could be useful to poolies and even real general managers who might want to calculate their playoff chances come the trade deadline.
All in all, no system is perfect, but Roberts' calculations make a pretty compelling case for the way the season will play out.
And when it comes to Ottawa, well, figure it this way: the Senators' chances are two-tenths of one per cent better than the Toronto Maple Leafs and the Leafs come in at 0.0480333333 per cent.
Do the math.