Monthly Archives: January 2018


First look at 2018 tournament likelihoods

This week’s PWR by wins forecast is now available. See last week’s post, First 2018 PWR forecasts available, for an explanation of the forecasts and how to interpret the results.

For those who would rather just know the bottom line, this post will go into deeper detail interpreting the forecast table.

Is anyone safe yet?

No one is completely, mathematically, guaranteed a tournament appearance yet.

Neither #1 Notre Dame nor #2 Clarkson fell below #12 in any of the simulations (and thus would almost be guaranteed an at-large bid). But, note that in those same simulations neither team ever dipped down to 0-1 wins, so falling to the bubble is mathematically possible, but would take a highly improbable collapse.

Ok, then who’s likely to make the tournament?

#1 Notre Dame, #2 Clarkson, #3 St Cloud St, #4 Cornell, #5 Ohio St, and #6 Denver are all most likely to finish in the top 12 if they win at least half their remaining games.

Is anyone out?

Much like the at the top, very little is completely mathematically settled at this point. But, by looking at the near perfect seasons some lower ranked teams would require to get an at-large bid, you can guess at the low likelihood of that outcome.

From #35 Mercyhurst down need a near perfect remaining season to get in position for an at-large bid. Those teams include the following:

#35 Mercyhurst
#36 Merrimack
#37 Air Force
#38 Army
#39 New Hampshire
#40 Holy Cross
#41 Niagara
#42 Princeton
#43 Robert Morris
#44 Quinnipiac
#45 Bentley
#46 RIT

Getting an at-large bid, even with a perfect remainder of the season, seems very unlikely for #47 Ferris St and down.

#47 Ferris State
#48 Arizona State
#49 Dartmouth
#50 Connecticut
#51 American Internationl
#52 Brown
#53 Alaska
#54 Alabama-Huntsville
#55 Rensselaer
#56 St. Lawrence
#57 Sacred Heart
#58 Lake Superior
#59 Vermont
#60 Alaska Anchorage

What next?

I’ll keep updating the forecasts weekly, so you can always browse them yourself. I’ll also try to post interesting interpretations here, with increasing frequency as we near the end of the regular season. Meanwhile, you can explore more of the data yourself:

First 2018 PWR forecasts available

With the regular season about half over, it’s a good time to start paying attention to PWR. Remember that PWR is a ranking whose calculation mimics the NCAA’s tournament selection process, so the final PWR ranking perfectly predicts the teams that will be selected for the NCAA tournament. The 2014 article, When to start looking at PWR (revisited), examines how today’s PWR is reasonably predictive of the final PWR.

Current PWR Ranking

But, if what we’re really interested in is knowing what PWR is going to be at the end of the regular season, can we do better than a table of the PWR as if the season ends today? In last year’s article, New forecast presentation—PWR by wins, I introduced a new tool that can be used to answer questions such as:

  • How many wins does my team need to make the tournament?
  • Can my team make the top 4?
  • What are some unlikely tournament seeding outcomes that could occur?

PWR By Wins (What does it take for each team to finish at each PWR ranking?)

The table on the PWR By Wins page shows you how many wins each team needs to likely finish at each PWR ranking. If you want more detail on a specific team, you can click a team name to see the probability curves of how likely that team is to end the regular season with each PWR ranking with a given number of wins in its remaining scheduled games.

The forecasts are usually updated in the first half of the week. You can always browse all the data any time, but I’ll also scour the data and post interesting results and observations in this space in coming weeks.

How it works

The page notes when the forecast was last run (assume that it includes all games that had been completed as of that time).

Each forecast is based on at least one million monte carlo simulations of the games in the described period. For each simulation, the PairWise Ranking (PWR) is calculated and the results tallied. The probabilities presented in the forecasts are the share of simulations in which a particular outcome occurred.

The outcome of each game in each simulation is determined by random draw, with the probability of victory for each team set by their relative KRACH ratings. So, if the simulation set included a contest between team A with KRACH 300 and team B with KRACH 100, team A will win the game in very close to 75% of the simulations. I don’t simulate ties or home ice advantage.