CEBL Summer Series Analytics: Previewing the Tournament by the Numbers

Kingston,ON- Professional basketball will be back in Canada on Saturday (July 25). And even better, it will be extremely competitive, played at a very high level, and after fifteen days we will have crowned a national champion. If those two sentences don’t get your blood pumping you may have clicked on the wrong article.

The basketball I’m referring to is the CEBL Summer Series. For those of you who aren’t aware, the CEBL began play in 2019. The league was founded with a very intelligent concept – all rosters would be comprised of at least 70% Canadian players. The first season involved six teams:

  • Edmonton Stingers
  • Fraser Valley Bandits
  • Guelph Nighthawks
  • Hamilton Honey Badgers
  • Niagara River Lions
  • Saskatchewan Rattlers

In the inaugural season each team played twenty games. The Niagara River Lions earned first place with a 15-5 record, finishing a game ahead of the Edmonton Stingers. The playoffs were a single-elimination tournament involving four teams and were hosted at a pre-determined site (Saskatchewan). Rounding out the playoff field were the Saskatchewan Rattlers (11-9) and the Hamilton Honey Badgers (10-10).

Both semi-finals were upsets and instant classics, with Hamilton knocking off Niagara 104-103 and Saskatchewan edging Edmonton 85-83. In the finals, third-seeded Saskatchewan (the tournament host) claimed the first-ever CEBL championship, dispatching Hamilton 94-83.

League-champ Saskatchewan Rattlers to defend title at CEBL ...
(Devin Heroux/CBC Sports)

Below is a table summarizing the 2019 season. All stats, with the exception of wins and losses, include playoff statistics.

We’ll work our way from left to right in the table. Wins and losses – pretty self-explanatory, right?

Next comes points per game, and points allowed per game. Points per game differential is simply PPG minus PPGA. Not surprisingly, Niagara, the team with the best record in the league, also had the best point differential. What is interesting though, is that if you sort the teams by point differential, two sets of teams would swap places. Saskatchewan would jump Edmonton into second place, and Fraser Valley would pass Guelph for fifth. That is especially interesting because point differential has been proven to be a better measure of team quality, and better predictor of future success, than wins and losses. That’s due to the fact that close games often come down to luck. In a season of only twenty games, a team of average quality could very easily go 4-0 in close games. Unfortunately for them, winning that percentage of close games is not sustainable in the long run.

Another reason that point differential is a good predictor, is that good teams win in blowouts quite often, but very rarely lose in blowouts. The opposite is true of bad teams. We can look at those two statistics for every team in the league.

Close games and blowouts have arbitrary cut-offs in the above table, but it illustrates the point. I defined close games as any game in which the final margin was five points or fewer, and big wins or losses as any game in which the final margin was fifteen points or more. The reason Fraser Valley has a better point differential than Guelph is immediately clear – Fraser Valley was winless in close games, but actually above .500 in blowouts. That’s the profile of a team that was unlucky to go 4-16.

The table isn’t as clear cut on Edmonton and Saskatchewan. Saskatchewan was slightly better (luckier) than Edmonton in close games but also had one more blowout. The conclusion is that these two teams were of very similar quality last year and likely should have had records more similar than 14-6 (Edmonton) and 11-9 (Saskatchewan).

Moving farther to the right on the season summary table, we see expected wins and expected losses. These two columns reflect a lot of the concepts we were just talking about. Expected wins and losses (also called Pythagorean wins and losses) adapts a concept to basketball that Bill James first used for Baseball. It uses a team’s points per game and points allowed per game to predict what a team’s record would have been if a portion of luck was removed. If a team’s actual wins are greater than expected, they may have been lucky, and vice versa – if a team’s actual wins are lower than expected. As you might expect, we see Edmonton’s actual wins are higher than expected, and Saskatchewan’s are lower. Fraser Valley’s wins are also significantly lower than expected, for the reasons discussed above.

Fraser Valley Bandits winless streak hits six games – Peace Arch News
(Peace Arch News)

Next, we see offensive efficiency rating and defensive efficiency rating. These columns are simply points scored and allowed per 100 possessions. Scaling points scored or allowed to a set number of possessions allows us to eliminate pace, and see which teams are actually good (or poor) on offence or defence. A team that plays at a fast pace may score a lot of points simply because they have more possessions per game, not necessarily because they are efficient on offence. Based on ORtg and DRtg, we see that Niagara was actually the best offence, and the best defence, in the league. If we simply looked at points scored and allowed per game, we would have assumed that both Edmonton and Saskatchewan were better defensive teams than Niagara, when in fact that’s not the case.

Finally, on the far right, we have pace – the number of possessions a team averages per game. We can see that Guelph played at the fastest pace, while Edmonton and Saskatchewan were the slowest.

So now that we’ve recapped the 2019 season, let’s move onto 2020.

A lot has changed since the first season concluded in August 2019. For starters, a seventh franchise (Ottawa) joined the league. You may have even heard of their General Manager, Dave Smart. After thirteen National championships as the Head Coach of Carleton University, he’s ready to try his hand at the CEBL.

Dave Smart (Chris Young/ The Canadian Press)

The other teams can be split into two distinct groups – those that kept a core group of players from the 2019 season and attempted to build around them (Edmonton, Guelph, and Niagara), and those that blew up essentially their entire roster (Fraser Valley, Hamilton, and Saskatchewan). Due to the latter group, and the Ottawa expansion franchise, more than half the league is coming into 2020 with minimal CEBL experience.

The 2020 season was scheduled to begin in May. However, due to the COVID-19 pandemic, it had to be postponed. In June, the CEBL announced a plan to return to play with the CEBL Summer Series.

The series will be played at a bubble site, hosted at the Meridian Centre in St. Catherine’s. From July 25 to Aug. 5, each team will play six games (one against every other team in the league). After the round- robin, the team that finishes in seventh place will be eliminated. The remaining six teams will enter a single-elimination tournament, with the top two seeds earning byes into the semi-finals. In the quarter-finals, the third seed will play the sixth, and the fourth seed will play the fifth. The lower-ranked winner will match-up with the top seed in one semi-final, while the higher seeded survivor will play the second-ranked squad in the other. The two winners will duel on Aug. 9, playing for the 2020 CEBL championship.

New to the CEBL this year is the Elam Ending. The Elam Ending is intended to keep the game flowing, and to avoid the situation in which the team that’s behind in the fourth quarter begins fouling. At the first stoppage in play with 4:00 or fewer remaining in the fourth quarter, a target score will be set. The target score is equal to the leading team’s point total plus nine points. So, for example, if the clock is stopped with 3:55 remaining, and Edmonton leading Fraser Valley 80-75, the target score would be 89. At that point, the game clock is turned off, and the game continues until one of the two teams reaches 89 points.

“The Basketball Tournament,” played in the U.S., has used the Elam Ending in all their games since 2018. They adopted an additional rule in 2020, whereby a team that is in bonus, and is fouled on a non-shooting foul, receives one foul shot and the ball. If that rule is also used for the Summer Series, it will benefit teams that can defend well without fouling. The table below shows 2019 team defensive efficiency and fouls committed per game.

In 2019, Saskatchewan had a very strong defensive efficiency rating and also fouled at the lowest rate in the league. It remains to be seen whether Saskatchewan’s defensive proficiency will continue, with many of their key pieces gone from the 2019 roster. Whether that rule is adopted, or not, the Elam Ending is likely to hurt teams that shoot free throws well (relative to regular rules), as normally a team with a lead very late in the game is in good shape if they can hit their free throws. Last year Fraser Valley led the league in free throw shooting (although their roster has undergone a massive overhaul in 2020). Niagara shot free throws at the second- best clip.

Got all of that? Good, because there will be a quiz later.

So, what can we expect from the 2020 Summer Series? It’s very cliché, but I would start with expecting the unexpected. Sports, in general, even when a full season is played, involve a lot of randomness.  Now we are talking about a six-game season, in a league in which four out of seven teams have essentially entirely new rosters. So yeah, I would say we are going to see some surprising results. But surprising results make for lots of excitement, and a fantastic tournament.

This is reflected in the Summer Series title odds, shown below.

The odds are based on a model I developed which attempts to project the quality of each team based on the 2020 training camp rosters. Based on the model developed ratings, I then calculated the probability each team had of winning a match-up versus any other team in the league. Those probabilities are then used to simulate the Summer Series 10,000 times. The results of the simulations are shown in the above table as the team’s title odds.

We can also calculate the probability that each team makes the playoffs or gets a first-round bye (finishing first or second).

Not surprisingly, given that six of the seven teams will make the playoffs, every team has a playoff probability well above 50%. Niagara leads the league in playoff probability (95.5%) and bye probability (53.2%).

Given the extreme turnover in the league, any kind of projection will need to attempt to predict how players in leagues all over the world will stack up against each other. How good is one of the best players in the Cyprus Division A, compared to someone that had a cup of coffee in the NBA G-League? That’s the type of question you need to try to answer if you want to build a model to predict this year’s Summer Series. Suffice it to say, it’s a challenging, but interesting task.

My model is based on win shares. Win shares is a player statistic that attempts to divide up credit for team success to the individuals on the team. It accounts for both offensive and defensive performance. As an example, we can look at the win shares for the 2019 Niagara River Lions season.

The total win shares for the 2019 River Lions was 15.6, compared to their win total of 15. Total win shares is generally fairly close to the team win total. In this case, some of the difference may be due to the player win shares being rounded to one decimal place. In any case, the higher a player’s win share total, the more he contributed to his team’s success. In this example, Dorian Pinson was responsible for approximately three of his team’s fifteen wins.

Getting back to the modelling, the task for players that played in the CEBL in the 2019 season was fairly simple. I rolled forward their win shares to the 2020 season, making modifications if they joined a new team, if I expected them to play significantly fewer or more minutes this season, or if their 2019 win shares seemed significantly out of line with their previous college or professional performance.

For players that had no CEBL experience the task was significantly more difficult. I scraped individual player data (including win shares) for most of the top European leagues, and the G-League, from realgm.com. This allowed me to get statistics on the majority of the players in the CEBL this season. Manual look-ups on the CEBL website provided data on the remaining players. I won’t get into all the details, but I attempted to assign win shares for each CEBL rookie, based on his past performance, his current team, and his likely role within that team.

Once that was completed, summing the win shares for all the players on each team produced an expected wins total for that team. Converting a team’s expected wins to an expected point differential allowed me to produce the win probabilities mentioned earlier in the article. The win probabilities were then used to simulate the Summer Series 10,000 times, and voila, title probabilities were created.

This method is not without flaws. Win shares are highly dependent on the team you play on, your role within that team, and that team’s success. Because of that, it is not easy, or always reliable, to attempt to predict win shares for a player on a new team or in a new league. With that being said, I think in the end this modelling produces reasonable playoff and title probabilities. As mentioned earlier, anything can happen at the Summer Series – with every team having title odds between 5% and 30%.

Based on my model, Niagara is the Summer Series favourite. They bring back most of the key players from a team that finished first in the regular season and welcome a number of other good players to the roster, including Tyrone Watson, Ryan Ejim, and Daniel Mullings.

Image may contain: 1 person, playing a sport, basketball court and shoes
Daniel Mullings (Niagara River Lions)

The expansion Ottawa Blackjacks come in ranked second. Their roster has an extreme Ottawa flavour, with the majority of their players having played at Carleton at one point in their careers. The Scrubb brothers are likely to be among the best players in the league, and recent addition Olivier Hanlan bumped the Blackjacks ahead of our next team.

The Edmonton Stingers have the third-highest title odds. They bring back a very strong core with Xavier Moon, Travis Daniels, and Brody Clarke. The additions of Mambi Diawara and Kirk Williams Jr. are good ones, and should position the Stingers to be among the better teams at the Summer Series.

Canadian Elite Basketball League MVP Xavier Moon excited for ...
Xavier Moon (Lary Wong/Postmedia)

The Hamilton Honey Badgers are up next based on title odds, although separating the next three teams (Hamilton at 10.6%, Fraser Valley at 10.5%, and Saskatchewan at 10.4%) is splitting hairs. Their final Summer Series roster, which is missing Justin Jackson and MiKyle McIntosh, doesn’t look quite as impressive as the original training camp roster. Despite that, Hamilton still has significant G-League, and even NBA, experience. A team with Briante Weber, Derek Cooke Jr., and Duane Notice, will be tough to beat.

Fraser Valley is up next, and if you saw them play last year you may not recognize them this year. Their only returning player from last year’s roster is Marek Klassen. New Head Coach, Kyle Julius, has brought in an all-new cast, including Kyle Johnson, Cameron Forte, and Junior Cadougan.

In sixth based on title odds, we have the defending champion, Saskatchewan Rattlers. The original training camp roster suggested they would be bringing back the majority of their top players. Based on that information, the model projected Saskatchewan to be the second-best team at the Summer Series. However, recent news suggests the Rattlers will be without Alex Campbell, Tavrion Dawson, Chad Posthumus, and Armoni Brooks. After including the losses of Bruce Massey, and Marlon Johnson, the Rattlers have lost the top four players (measured by win shares) from their 2019 title team. Saskatchewan did add Denzell Taylor and Kris Joseph (on loan from Niagara), among others.

Finally, we have the Guelph Nighthawks. Guelph will be led by returning players Kimbal Mackenzie, Jamal Reynolds, and Olu Famutimi. Reinforcements for the Nighthawks include Joel Friesen and Tre’darius McCallum.

With that quick (or not so quick!) rundown out of the way, we still have a lot to get through. Part two of the preview will be posted tomorrow and will have a section dedicated to each team. So if you’ve made it this far, you only have seven sections left to go. If you plan ahead, keep meals, hobbies, and sleep to a minimum, you should be able to make it through part two in time for the first game on Saturday.

For more info on the U SPORTS players in the CEBL, CLICK HERE!

*Summer Series Analytics Parts 1 and 2 will be updated periodically throughout the week as new roster information becomes available. Most recent update was July 24, 4:30 PM EST.

One thought on “CEBL Summer Series Analytics: Previewing the Tournament by the Numbers

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s