A Winning Strategy for Choosing Players in Your Hockey Pool

November 17, 2005, vol. 34, no. 6
By Howard Fluxgold



Document Tools

Print This Article

E-mail This Page

Font Size
S      M      L      XL

Related Stories

The probability that Amy Summers could play hockey for the Vancouver Canucks is likely zero since she is only five feet tall.

But the probability she might win her hockey pool by choosing a few Canuck players is considerably higher now that she has completed her thesis called: Hockey Pools for Profit: A Simulation Based Player Selection Strategy.

The thesis attempts to show statistically the best method for selecting players for a hockey pool for the Stanley Cup playoffs.

“I tried to come up with a strategy for choosing players based on statistical methods, rather than just loading up on players from one team that is favoured to win,” she says in an interview from her Prince George home.

With no Stanley Cup playoff last spring when she was working on her project, Summers performed a simulation instead.

Juggling the probability that a team might play many playoff games with the probability that some players would score many goals, she came up with a computer program that either won or finished second or third in simulations more than 50 per cent of the time.
And she did such a good job that she earned her master of science in statistics as well.
Her supervisor, Tim Swartz, professor of statistics and actuarial science, says, “Her method worked well in simulations. It was sound statistically.”

Summers' computer program assesses the value of each player based on the number of points scored during the regular season divided by the games played. It also uses the odds of a team winning their series developed by Sportsbook, an online oddsmaker.
“Her thesis is based on probabilities,” explains Swartz who is planning to author a journal paper with Summers and another professor, Richard Lockhart. “In any one year a pool participant might not win using Summers' strategy, but over the long run the probabilities are high that they will.”

Swartz says her thesis shows that her strategy is superior to various “ad hoc strategies” such as choosing many players from a favoured team.

Summers, a soccer player but “not a diehard hockey fan,” will try out her theory on her father and some friends this year.

However, the computer program she developed will remain out of the public domain for now.

Search SFU News Online