Tuesday, September 24, 2013

Does winning close games help in the playoffs?

A while back, I added the "Big Win" stat to the Py Win Rankings.  The idea behind this stat is that it counters the notion that "good teams win close games", and that winning close games prepares a team for success in the playoffs.

The idea for the stat, and the logic behind it, comes from Jim Glass of advancednflstats.com.  In his article, he analyzes playoff records for the best playoff teams, and groups them by their record in close games and blowouts.  I'm following his example with the CFL data.

Test Data

Like Mr. Glass, I will filter the data to the top tier playoff teams, those which made the playoffs with at least 10 wins..  More wins means of course that they will have more "big wins" or "close wins" respectively.  It will filter out some Grey Champions which didn't finish in the top of the league, but the goal here is simply to determine if big wins or close wins are more closely related to playoff success, not analyze each Grey Cup winner in detail.

Filtering to 10+ win teams since 1990 leaves us with 93 teams.

I'm using 9 points or more to represent the cut off between blowout and close win.  The original NFL article used 10 points, but he later amended it to 9 points, which represents the cutoff between a 1 and 2 possession game.


Most close wins:

  • Of the 93 teams, 8 won 8 or more close games.  Their record in the playoffs was 5-7.
  • The 15 teams with the best records in close games combined to be 91-14 (87%) in those close games.  In the post-season, they had a record of 17-11 (60%), 6 Grey Cup appearances, and 4 winners.
 Fewest close wins:
  • 20 of the 93 teams has losing records in close games.  The 15 with the worst record in close games combined for a record of 34-66 (34%).  In the post season, they had a record of 19-11 (63%), 9 Grey Cup appearances, and also 4 winners.
It doesn't appear that a team's record in close games matters very much at all, as both the best and worst teams in close games have very similar records come playoff time.

 Grey Cup Winners:

Looking at things from the perspective of the 20 champions (the 2012 Argos, 2000 Lions and 2001 Stamps didn't make the 10 win cut), they're win-loss records shake out as follows:
  • The playoffs: 47-0 (of course)
  • The regular season: 252-106-2 (70%)
  • Close games during the regular season: 81-52 (61%)
  • Big wins/losses during the regular season: 171-54 (76%)
A 61% win rate in close games does show some ability to win close games, though not much better than a coin flip.  They were slightly better than the rest of the 10 win teams, however, as the average for all 93 teams was 60%.

Playoff results by win cohort:

Here's how winning close games matches up with winning playoff games, grouped by W-L record.  For the purposes of this exercise, I'm considering a tie to be a loss (it's not "clutch" to tie, right?).  Only 2 teams finished with ties.  (Value in brackets is winning percentage in close games).

15-3 (7 teams)
Top 3 (83%): 6-1, 2 GC winners
Low 4 (68%): 7-2, 2 GC winners

14-4 (4 teams)
Top 2 (86%): 1-2
Low 2 (69%): 1-2

13-5 (17 teams)
Top 8 (82%): 8-6, 2 GC winners
Low 9 (51%): 10-5, 4 GC winners

12-6 (20 teams)
Top 10 (72%): 9-1, 3 GC winners
Low 10 (49%): 15-8, 2 GC winners

11-7 (24 teams)
Top 12 (69%): 11-10, 2 GC winners
Low 12 (46%): 11-10, 2 GC winners

10-8 (21 teams):
Top 10 (61%): 10-9, 1 GC winner
Low 11 (40%): 9-11

The two groups win at nearly the same rate.
  • The "higher halves" have a much better record in close games, but a 45-29 record in the playoffs (61%).
  • The "lower halves" are much worse in close games, but have a similar record at 53-38 (58%), and more Grey Cup winners (10 vs 9).

The data isn't as cut and dry in the CFL as it is in the NFL (where the lower half of each group clearly has a better winning percentage), but the numbers are extremely close.  Close enough to suggest that a team's record in close games may not have much to do with playoff success, either positively or negatively.

Big wins and big losses:

Perhaps then, big wins and losses are a better indicator of playoff success than close wins?
  • 5 teams had 11 or more big wins in a season.  Their playoff record was  6-3 (67%), with 2 Grey Cup wins.
  • 15 teams had 10 or more big wins.  Their record in the playoffs was 22-8 (73%) with 7 Grey Cup wins and 3 Grey Cup losses, meaning 10 of those 15 teams made the Grey Cup.


It seems clear that close wins do not equal playoff success.  Of those 8 teams with 8 or more close wins, only the '95 Stallions had a successful play off run - they went 3-0 and won the Grey Cup.  The remaining 7 only appeared in 1 Grey Cup, with no wins.

However, while "big" wins do appear to correlate better with Grey Cup wins than close wins, they don't appear to correlate any better than straight up wins and losses.  I think I will explore this in better detail in a later post, but I believe this comes down to the CFL having less scheduling variance than the NFL.

For now, I will continue to include the "big win" stat on my rankings table, as I think it is interesting, but I suspect that more analysis will show that it simply lines up quite closely with overall win-loss records, and doesn't give us much useful information.  In fact, it may be more beneficial to include "close wins" instead, as an indicator of teams which may do poorly in the playoffs, vs using big wins to indicate those which will do well.

No comments:

Post a Comment