Quit Clownin'

Reviewing Charlie Danoff's
2018-19 Basketball Forecasts

Intro

CHICAGO - 21 February 2019 - 8:30 AM Central - When talking about what will come, most of us only have knowledge of what came before. To make the weather forecast for tomorrow Mr. Skilling can solely examine what happened today and days before. You would think it would be easier if he also knew something about what would happen in the future. Yet, even after Ms. Connor received information from the days to come courtesy of a time-traveler, it was still hard to precisely say when the droids disguised as the former Governor of California would attack.

So even for those of us with friends from the 2020s, making predictions is hard. Whether it be 'Will it snow tomorrow?' or 'Who will win tonight's basketball game?'. Looking only at the past to predict the future can make us look less like the star of a Hollywood action film and more like Bozo the Clown.

"The future doesn't mean predicting games-another form of entertainment that can make you look like a clown-but understanding how players will behave if they join a new team or why underdogs will start taking riskier strategies [emphasis added]."

Quote is from page 337 of 'Basketball on Paper' by Mr. Dean Oliver where he shared tools quantitative basketball analysis. Fourteen years of measuring the game in numbers and reading that data, more is understood about the game itself, but it is also clear there is still much to learn. As a result Mr. Ben Falk made the following observation last October:

"Here’s a dirty secret about the world of sports: We have no idea if people truly know what they’re talking about."

To his point, last June I was clowning around and wrote:

"3rd quarter game 2 prediction @cavs win Finals in 7 over @warriors".

I was quite wrong. The Cavs not only lost that game, but were swept in the series. Clearly I have room to improve as a forecaster. To reach that goal I want to review some predictions I made this 2018-2019 NBA season to evaluate my results and find out if there is anything to learn. Given we are the unofficial half-way point of the season following the trade deadline and All Star weekend, enough time has passed to generate a solid sample size.

Short-Term Predictions

Documenting your basketball predictions is easier now with the new 'Predict' feature from Ben Falk's Cleaning the Glass website where you answer questions with a percentage for how likely or not you think something is to happen. Your predictions are compared to your peers in the community and the actual results. You can see my work in Figure 1 below.


Figure 1 - https://www.cleaningtheglass.com/predict/leaderboard
Screenshot taken on 21 January 2019

Will the Warriors beat the Thunder on Oct. 16th, 2018?

I forecasted it was 75% likely to happen; I kept my reasoning terse:

"If Andre Roberson were playing, Thunder would have better chance."

I was right on the outcome, but not confident enough. The crowd thought it was five percentage points more likely at 80% so I lost -0.5 points.

Will the Celtics beat the 76ers on Oct. 16th, 2018?

I predicted it was only 25% likely and I was incorrect.

"Will take a little for Hayward and Irving to work their back to normalcy, while 76ers return their entire core."

Hayward and Irving weren't themselves combining for 17 points, but their team still beat the 76ers. The crowd thought this was 65% likely so here I was docked 8.1 points.

Will Markelle Fultz make 3 or more total three-pointers before Nov. 1st, 2018?

I thought it was 95% (extremely) likely to happen. My Explanation was he:

"Made 1 of his 5 preseason three pointers in 4 games. That is a big improvement than the 0 for 1 in 14 games [last season]. More importantly his year in college he made 52 of 126 threes (.413). Combine that with his reported offseason work and he should be fine to make three or more threes before November 1st."

He made four. I got 11.2 points for my confidence. The crowd average was 75%.

Will Jimmy Butler be traded before Nov. 1st, 2018?

Coming out of training camp I was confident he would be gone by the end of the month (despite Jimmy saying he requested a trade privately months earelier), stating it was 75% likely.

"Thibs wants to keep him, but the outside pressure will be so intense that they will need to move him."

Trade happened November twelfth, not by Halloween as I predicted. I got crushed here. The crowd said it was only 27% likely so I lost a staggering 156.1 points. All in all I was worse than the crowd three out of four times so far.

Season-Long Predictions

I made more predictions before the season started about what may happen over a longer time frame.

Will the Lakers make the 2018-19 playoffs?

I stated it was 50% likely on October 16th. I am a passionate LeBron James fan, to the point I feel it clouds my judgment (as you saw with my Tweet in the introduction). Growing up as a Bulls fan of MJ and then watching Kobe through my teenage years, I respected Mr. James' desire to win titles, which I felt he betrayed by his choice of franchise.

"Emotionally I am annoyed at LeBron for going to LA. After genuinely trying to win a NBA championship since he was 22 years old in the 2006-07 season, he made a choice to join a team clawing for a playoff spot. Watching him compete deep into the playoffs has been one of the perks of the spring season for a decade. Emotion aside, he is due for an injury. If he does go down though, I feel the rest of the Lakers young core combined with some leadership by Rajon Rondo can still make the playoffs. Last two years Rondo was on teams that got back to the playoffs. That said, I cannot go beyond 50% because beyond the 8 teams that made the playoffs last year in the West, the Nuggets, Clippers and Grizzlies all will be in the mix for the 8 seed. The injuries will hurt San Antonio and losing Butler will harm the Timberwolves, but they are both extremely well coached. They will compete and also have a great shot at the 8 seed. Taking all that into account, I feel like the Lakers making the playoffs is a coin flip."

I was wrong about the T'wolves and Grizzlies competing for the playoffs, but Mr. James did have the longest injury of his career, the Nuggets, Spurs, and Clippers are currently in the playoffs, and Mr. Rondo has become a leader, per Mr. Plaschke of the LA Times:

"For all his greatness, James is not this team’s veteran leader. That role has been taken by Rajon Rondo, who has become the most respected and trusted figure in the young locker room, something which was evident in the team’s joyous postgame celebration after his buzzer beater in Boston."

And even hinted at by LeBron himself:

"And, as James pointed out, these turnovers are happening with Rajon Rondo, their veteran floor general, sidelined with a broken hand."

Unfortunately for the Lakers, in addition to James, Rondo and others missed games due to injuries. With a healthy roster down the stretch, there is certainly a chance these Lakers can make the playoffs. The Kings and Clippers, however, will not give up their spots ahead of them easily, and it's even possible one of Minnesota (only one game behind in the loss column with 30 to the Lakers 29), Dallas (31 losses), or even New Orleans (33) goes on a crazy run and finishes ahead of the team aspiring for a return to Showtime. I am not as confident in my coin flip as I was in October, but I'm sticking with the prediction of a 50/50 chance for them to make the playoffs for now.

Will Hamidou Diallo make the All Rookie team?

I forecasted the 45th pick of the 2018 draft would during the Thunder preseason tilt with the Bucks.

"@hamidoudiallo also looked good against @ATLHawks last game ... if he keeps this up and gets minutes @kpelton mentions, he’ll be on all rookie team"

He came onto my radar during the summer league. I did not know a lot about him, but his two-way potential was tantalizing. Look at him taking it to hole, stealing it and making it from beyond the arc in these summer league highlights.


Figure 2 - Hamidou Diallo Full Highlights vs Nets (2018.07.07) Summer League - 19 Pts, 8 Reb via YouTube
by House of Highlights

In his weekly rookie rankings Drew Packham back in November commented:

"Diallo had become a fixture in the rotation, starting in OKC's previous two games. In Monday’s win over Sacramento, he scored 18 points, becoming the first Thunder rookie to go at least 7-for-7 from the field."

Unfortunately he hasn't made the list in the weeks since. Minutes wise he's played 515 total minutes in 45 of 57 Thunder games starting three. As a multi-positional wing he's played behind Terence Ferguson, Alex Abrines (who has since been waived) and when Westbrook plays with Schroeder together (and likely behind Roberson if he returns).


Figure 3 - raining rainy day GIF by Twitter via GIPHY tagged #hamidoudiallo

At the time of publication, he is 28th in minutes amongst rookies, 26th in Steals, and also tied for 26th in Blocks. Check out his two-way game with this block and coast to coast finish in Figure 4 below:


Figure 4 - Diallo swats shot then slams it on other end via ESPN

Using his FT rate (100 times Free Throws Attempted divided by Field Goals Attempted) as a measure of an efficient offensive player was strong in college 40.3 346th in the nation (his only top 500 college ranking). So far in pros it has dipped to 36.6. That said by another metric he is drawing fouls at a high percentage compared to his NBA peers, as he is in the 98th percentile for Shooting fouled percentage:

"Shooting fouled percentage measures how often the player draws fouls on their shot attempts. Note that this will be dependent on where the player typically shoots from: jump shooters will draw fewer shooting fouls than those that attack the rim."

Look at him draw the foul on Dirk December 30th after driving past him in this video:

Unfortunately he is only making 61% of his foul shots (just like the 61% last year studying in Lexington). He is showing potential but to be one of the ten rookies voted onto the All-Rookie team he'll need earn more minutes by improving his play and making it hard for Coach Donovan to keep him on the bench. If any of the coaching voters had not heard about him, they certainly did after his artistic endeavors Saturday night.

Will Kyle Anderson have the best year of his career?

I believed the St. Anthony's graduate would indeed have his best season. It has already been eventful; he's played his former coach, been called a scamp by Zach Lowe, and even had a #TrophyGame as the next video proves:

Heading into the year I made predictions about Anderson's plus minus rating.

"At the beginning the year it is negative, -1.76 then at the All Star break it is 2.09, and at the end is 3.68 ... The mean average of the three is 1.34."

We are at All Star break yet, but as shown in Figure 5 his current +/- is -1.4, which is far below 2.09.


Figure 5 - Kyle Anderson's Stats on NBA.com
Screenshot taken on 21 February 2019

In order for my prediction to be accurate, he needs to have a more consistently positive +/- for the remainder of the year. He is shooting well from two-point range at 58.3%, highest of his career, but it would help if he improved his three point (26.5%) and free throw percentages (57.8%), which are the worst of his career. His shooting chart in Figure 8 leaves a lot to be desired, look at all the blue where he is shooting below league average.


Figure 6 - Kyle Anderson Shooting Chart by Positive Residual
Screenshot taken on 21 February 2019

On separate notes I also predicted

"if he does not get injured, with more minutes and opportunities in Memphis Anderson ... will validate the investment the Grizzlies made."

After missing only one game early in the year, he has been injured more often lately. Yet even with that, his average minutes are 29.8, highest of his career. He is making an impact all over the floor as you can see in this video


Figure 7 - Kyle Anderson : All Possessions (11/07/18) via YouTube
by NF

Ultimately it is still too early to say whether or not the investment made by the Grizzlies has been validated, but I believe he is heading in that direction. He is still elite or near-elite defensively as shown by being 9th in the league in Defensive Box Plus/Minus at 3.5 (per Basketball Reference) and in the 90th percentile in block % and 83rd percentile in steals (per Cleaning the Glass) though he has been fouling more (in 38th percentile this year vs 84th last year). His passing skills also remain strong (88th percentile in Assist % per CTG), but his shooting from three and the line needs to get back to his career averages and then improve to above average from there for him to make Grizzlies management superbly happy with their decision to sign him.


Figure 8 - Kyle Anderson Defensive Stats from Cleaning the Glass
Screenshot taken on 21 February 2019


Figure 9 - Kyle Anderson on Leaderboards from Basketball Reference
Screenshot taken on 21 February 2019

Multi-Year Prediction

Are teams with bigger analytics staffs more likely to win?

I assumed they were, but the results of my investigation (kindly published by NBAstuffer) were inconclusive.

"Over the eight season sample, staff size was a positive predictor of regular season wins and SRS indicating growing your staff can at least help you get to the playoffs."

Heading into the year the 76ers (franchise that made Ben Falk [who was quoted in the introduction] their VP of Basketball Strategy during the Hinke-era) had the biggest staff (as defined by this awesome page from NBA stuffer) with nine members, based on my regression model I predict they will end up with 49 total wins. They had 52 last year.

38.33 + 1.19 * Analytics Staff Size = Regular Season Wins

38.33 + 1.19 * 9 = 49 Wins for the 76ers


Figure 10 - Using Regression to Predict Regular Season Wins based on Analytics Staff Size Results
Source Workbook on GitHub

However if we look only at a shorter time frame of only the last three seasons, staff size was not a statistically significant predictor of regular season wins. Perhaps in the past only certain teams were using analytics, which gave those franchises a substantial advantage. Now that everyone is to a certain degree the benefits are smaller. Even if the results were not significant, they still indicated having a bigger analytics staff was better than a smaller one.

This topic deserves more investigation, and Martí Casals‏ kindly gave some nice suggestions for additional questions to ask the data.

Lessons Learned

Since I want to improve at predicting future basketball outcomes, it is important to learn from my forecast work. I am going to use the five parts of the Paragogical Action Review framework review what I have done and determine next steps.

I hope to hear from you with criticism and ideas for how I can improve. With your help I hopefully will look less like a clown and more like someone who knows what he's talking about.