NBA – Journalists, managers, fans… Millions of them gather information on this site specializing in the world of statistics. A platform that has become essential, which continues to grow.
This online site, which seduces with its ease of use, has established itself as a must for all fans of the orange ball, eager for statistical navigation between the NBA of today and that of yesterday. ESPN considers him ” the best basketball history site in the world “. He is “ became the first place I go when looking for NBA stats “says David Leonhardt of the New York Times, making himself the spokesperson for an entire profession.
« I believe in terms of NBA history, WNBA… it’s hard to compete “, abounds Mike Lynch, director of data for Sports Reference (about thirty employees in total), the parent company of the site, joined by Basketball USA. He too used the site on a daily basis at the time, about ten years ago, when he played at ESPN and other US sports media: “ I’ve always found it to be the best thing to do for me to do my job, to find things that are complete and well organized, with the tools to dive into them. I had a lot of respect for the site. I really believe in their mission. It is a pleasure to work there every day. »
The same year as Facebook
In the beginning, there is a man, Sean Forman, at the origin of the first version of the brand: “Baseball Reference”, launched in 2000. The site was born, as the New York Times said, while Forman was looking for… to avoid working on his doctoral thesis in applied mathematical and computer sciences at the University of Iowa. This enthusiast had begun his compilation of data thanks to the CD-ROMs that accompanied the “Total Baseball” encyclopedias. Without suspecting that his initial idea was going to make children.
Justin Kubatko, a graduate in mathematics and holder of a master’s degree in statistics at the University of Ohio State, wants to resume his model for basketball. The site was born in 2004, the same year as Facebook. ” It started out as a hobby, but as the site quickly grew in popularity, I realized it could grow into something much bigger. “, testifies to Basketball USAthe founder, who originally did not imagine such a development of his database, now regularly mentioned by the American media.
After a few years of existence, the site obviously does not have the graphic appearance of today but already gives a good overview of what it will be able to provide information on each player (example here with the Baron Davis). Above all, the platform is already emulated enough for Justin Kubatko to leave his post as a statistics teacher at Ohio State in 2007 to devote himself full-time to its creation.
A Hobbyist’s Collection Wasn’t an Urban Legend
Finding current data is one thing; distant data in time, another. ” Many online newspapers add new pages very regularly. It’s not always helpful. Sometimes we get there on our own. We also have people who come to us saying ‘Hey I found this, could you be interested?’ “, reports Mike Lynch about this hunt for “box scores”, real treasures of the site.
It is rather within the framework of the second scenario that the platform will experience a major advance. In early 2012, Justin Kubatko announced on the site’s blog that he now has a game sheet… for each game played in NBA history, stating that he had collected scans for the seasons from 1946-47 to 1984 -85: “ I would be remiss if I didn’t thank Dick Pfander, who took the lion’s share of this project. Several years ago Dick was recommended to us by a third party. At first, it sounded like an urban legend. Seriously, someone had scanned every box from every game in history? »
Much to the founder’s surprise, the answer is yes. Also passionate about the NBA and statistics, Dick Pfander began cutting out game results from newspapers when he was a teenager in Michigan in the late 1940s. Alone, by hand, and at a time when the computer does not exist. His work becomes obsessive. On each trip, he searches the local press. In his crazy quantified epic, this amateur crosses paths with a certain Harvey Pollack, director of statistical information for the 76ers, a future major figure in the discipline, whose own collection of ‘box scores’ does not go back as far as that of Dick Pfander.
A typing champion in demand!
« We talked for several weeks, we made a deal and we bought what he had “, described Justin Kubatko to FiveThirtyEight, who devoted a paper to the collection of Dick Pfander. The site manager thus recovers a disk containing folders for each year and sub-folders for months and days. Problem: Results are saved as images, not spreadsheets or databases. Not the easiest, therefore, to add them to a statistical database. Unless you ask for a champion outside the typing category. Capable of typing up to 200 words per minute (!), Sean Wrona contributes to entering numbers.
Ten years after this painstaking work, it is now possible to go back to the beginnings of the Great League and have access to any game sheet, sometimes incomplete. ” At some point, we will undoubtedly reach a ceiling where all resources will have been exhausted. remarks Mike Lynch, recalling that certain categories were not counted in the past. Thus in the early 1950s, the total number of rebounds was referenced but without distinguishing between offensive or defensive catches. Interceptions, blocks or lost balls do not appear either.
Today and for several years, to aggregate the current “box scores” as well as the “play by play” of each match, the site calls on its partner, SportRadar, exclusive provider of NBA statistics. This sports technology company recently extended its partnership with the Great League, over several years, to expand the statistical tracking of players and “ revolutionize » the world of sports betting.
Not the same definitions
“Basketball Reference” has no official link with the NBA, whose statistics site has grown significantly in recent years. The data found on one or the other platform can also sometimes differ slightly. Example with advanced data such as the “Offensive Rating” (points scored on 100 possessions) of a team or a player, simply because the definition of “possession” is not the same.
« Unlike the NBA, we define how we calculate our possessions. Our method allows us to compare like-for-like each season until the mid-1950s, when what the NBA did was always a bit of a blur.
“, judge Mike Lynch whose database delimits this notion on the basis of an undrinkable mathematical formula, where the site of the league is much more vague in the matter.
The site, which now makes it possible to follow the evolutions of the main international championships (Spain, Greece, France, China, Australia…), has also created its own formula to evaluate its famous probability for each player to integrate the Hall of Fame. Thus, today, Kyle Lowry has an 85% chance of being elected to the basketball pantheon, in particular because of his title won with the Raptors in 2019, where a player like Derrick Rose, yet MVP of the league, n only 10% of being.
« In general, this ranking underestimates the case of international players since it only takes into account accomplishments in the NBA “, also notes Mike Lynch. He says he has never heard of players complaining to the site. ” On the other hand, I believe that some fans have complained about the low percentage of Ben Wallace “Adds the manager in allusion to the 45% of the former pivot, yet inducted in 2021.
Part of the tool has become paid
In addition to the galaxy of figures collected, the daily work on the site consists of manually updating the race for the various trophies (players of the week, etc.), the numbers according to transfers, the player files by adding links to articles dealing with an injury…” There are not only statistics on the site but we do not come to obtain opinions or articles. We have quite long articles, which explain concepts such as the ‘win share’, the ‘plus minus’… We could probably go a little more in depth in this register, go in this direction rather than making ‘news’ “, Executive Mike Lynch, who himself keeps a precise follow-up of the NBA “buzzer beaters”.
Don’t overdo the numbers
According to him, ” the more stats the better. But this encrypted data must also be combined with what is seen in the game. “Because basketball is not baseball. The stat sheet cannot contain the entire match. There’s so much going on on the court that doesn’t show up in the ‘box’ or ‘play by play’. “So rely on the stats to form an opinion yes, but not” an abuser ” neither.
Justin Kubatko « hated for example when certain highly random statistical indicators are combined to give the illusion of an authentic achievement. The founder, who left the site ten years ago due to differences of opinion with Sean Forman on the future, takes as an example this tweet from Elias Sports, another well-known player in the world of statistics.
A year ago, Josh Hart, for example, had fun when he learned that he was the first player in history to compile more than 40 points, 6 rebounds, 6 assists, 6 3-point shots. and 4 interceptions at over 70% shooting success. ” What the fuck can they get out of their ass (sic). I imagine they’ll also talk about the game being on a Saturday, in Portland and stuff like that. […] It’s cool. I’m gonna have to put it all on a little card that says, ‘Hey, I’m Josh Hart’ “, had quipped the Blazers player.
« Come on, it’s this kind of thing that puts people off “, annoys Justin Kubatko, who also fears that we become ” more and more insensitive “to certain statistical performances, such as the triple-double, “ because they are too common “. Mike Lynch reminds him that ” sport is meant to be entertaining, stats can be too. Someone will estimate that (this kind of encrypted compilation) is insane. But that doesn’t have to mean anything. Sport is entertainment. “In which “Basketball Reference” occupies a place of choice.
