If this sounds somewhat familiar, that's because I did a 2019-2020 version
and posted it back in March.
In terms of changes from that post:
TL;DR Tracked tweets of Woj, Shams and Haynes from 2018-2020 to see whether any of them report on a certain team or a certain agent's players more than their counterparts. Here is the main graph concerning a reporter's percentage of tweets per team separated into three periods (2019 season, 2020 offseason, 2020 season). Here is a separate graph with the Lakers and Warriors, because Haynes's percentages would skew the first graph.
- I've expanded the timeline to tweets from September 27, 2018. This is the first official day where each of Shams, Woj and Haynes were at their own respective companies. Shams moved to the Athletic from Yahoo in August, and Haynes moved from ESPN to Yahoo in September.
- I have a new section on agents!
- thanks to Lambchops_Legion for the idea
- I've also expanded the criteria on when a tweet could possibly be linked to an agent
During times like the NBA trade deadline or the lifting of the NBA free-agency moratorium, it’s not uncommon to see Twitter replies to
(or Reddit comments about) star reporters reference their performance relative to others.
Woj is the preeminent scoop hound, but he is also notorious for writing hit pieces on LeBron
it’s been widely rumoured that the reason for these is that Woj has always been unable to place a reliable source in LeBron’s camp). On the other end of the spectrum, it has been revealed that in exchange for exclusive intel on league memos and Pistons dealings, Woj wrote puff pieces on then-GM Joe Dumars (see above Kevin Draper link). Last summer, Woj was accused of being a Clippers shill
on this very discussion board for noticeably driving the Kawhi Leonard free agency conversation towards the team.
This is the reason I undertook this project: to see whether some reporters have more sources in certain teams (and certain agencies) than other reporters.
First I’ll explain the methodology, then present the data with some initial comments.
To make this manageable on myself, I limited myself to tracking the 3 major national reporters: Shams Charania of the Athletic, Chris Haynes of Yahoo Sports and the aforementioned Adrian Wojnarowski of ESPN.
- I didn’t use beat reporters, as most (if not all) of their sources would be concentrated on their local team
- Others that I considered but ultimately decided not to track:
- Brian Windhorst of ESPN (double-dipping in ESPN)
- Zach Lowe of ESPN (I consider him more of an analyst)
- Marc Spears of ESPN (harder to sift through Twitter feeds, as he posts a lot more unrelated/non-news-breaking content)
- Marc Stein of the New York Times (same as Spears)
- Kevin O'Connor of The Ringer (same as Lowe)
The time period I initially tracked for was from January 1, 2020 to
the end of the regular season
March, but after finding a Twitter scraping tool on GitHub called Twint
, I was able to easily retrieve all tweets since September 27, 2018. However, a month ago, Twitter closed their old API endpoints, and Twint ceased to work. I used vicinitas.io
but the data loading became more time-consuming. Therefore, the tweets are up to the date of October 15 2020.
How I determined information was by manually parsing text tweets by the reporter (no retweets):
- This means I did not include images or multimedia appearances such as television, radio or podcasts. The rationale for this is that I simply don’t have the time to listen/watch and record all the instances of providing information through sources on these mediums.
Now, I didn’t take every single text tweet:
- I didn’t include direct statements, be they from players or front office folks
- I separated them, along with podcast guests in another tab
- I didn’t include the summary tweet that Woj & Shams love to do: “Story filed to/Story on [employer]:..” because it doesn’t add anything apart from a link to a story (also, I personally don’t want to be called an ESPN/Yahoo/Athletic shill)
- If the tweet added a reporter’s own analysis to someone else’s tweet, it was not included
- If it was new information, the tweet was retained
- Tweets that related solely to retired players were not included: mainly Haynes reporting Dwyane Wade joining CAA, as well as the unfortunate passing of Kobe Bryant on January 26
- I grouped multiple tweets about the same subject delivered around the same time frame (such as trades) into one, as doing otherwise would arbitrarily inflate totals
- There’s no hard and fast rule for whether or not to group tweets
- For example, the big 4-team trade that created the Pocket Rockets was grouped in full
- On the other hand, the Miami-Memphis trade was split up because the full details came like a day later
- Sometimes, I used my judgment to determine whether a tweet’s underlying information would have come from a source, and therefore whether I should include that tweet or not
- For example, consider the All-Star tweets: Haynes and Shams both posted the All-Star starters, but looking at the time signatures led me to believe that this was simply relaying the information from the TNT reveal
- On the other hand, both Shams and Haynes posted tweets disclosing the All-Star Reserves before the TNT reveal
Next, I had to assign possible teams to each tweet:
- Items such as changes to the league calendar, the naming of All-Star Reserves and salary cap projections were immediately attached to an NBA source
- Injuries and trades were fairly straightforward, assigning these tweets to the participating teams
- Items such as league mandated fines/suspensions, invitations to All-Star competitions and game protests were credited to both a general NBA source, as well as the related team(s)
- Direct sources from agents or mentions of specific agents were attributed as a catch-all “Agent”
- In the former, team was not included: examples include Matisse Thybulle’s agent on not being selected for the Rising Stars Game or Royce O’Neale’s agents confirming his contract extension with the Jazz
- In the latter, team was included: examples include two Knicks switching their agent to Rich Paul
- New addition: anything related to a player's status with a team were also attributed to agents (qualifying offers, extensions, option decisions, waivers, and contracts/deals)
- I then found which agents correspond to which players (big shoutout to realgm.com and the Wayback Machine)
- Rumours were slightly more difficult
- As we know very well, league sources is an exceedingly vague term
- Instead of attempting to pinpoint a rival executive with a motive to make a comment, I took the “Occam’s Razor” approach and assumed that the teams involved had someone talk to the reporter
- When it was impossible to even determine a participant team, it was the general “NBA” source to the rescue
With all the methodology out of the way, here’s the data! (Here’s a link to a full Google Sheet
Here's a graph of number of tweets per team per period
, with the colours denoting reporter.
On a quick glance, here's which teams saw a significant period-over-period increase in number of tweets:
- Shams: Brooklyn, Milwaukee, New York, Sacramento
- Haynes: Atlanta, Brooklyn, Memphis, New York, Portland
- Woj: Atlanta, LA Clippers, Orlando
And here's which teams saw decreases over a period-by-period basis:
- Shams: Chicago, New Orleans, Oklahoma City
- Haynes: Golden State, New Orleans
- Woj: LA Lakers, Memphis, Minnesota, New Orleans, Washington
The problem with just using number of tweets is that it's not close on totals between Haynes vs. Woj or Shams. Here's a graph showing total number of tweets in each period for all three reporters
. Haynes's most reported period doesn't even stack up to the least reported period of Woj or Shams.
Instead, let's look at percentage of tweets per team per period
- Chris Haynes has the highest percentage of tweets relating to the Detroit Pistons in all three periods. He also reports on far more Portland news than Shams or Woj.
- Shams' Brooklyn edge is evident. The Athletic was also the outlet that Kevin Durant felt comfortable talking to about his positive coronavirus test. As well, Shams reported on Spencer Dinwiddie's quest to tokenize his contract (similar to bitcoin).
- Adrian Wojnarowski has increased his percentage of tweets regarding the LA Clippers period-over-period, but so have the other two reporters.
- It's surprising that Dallas's numbers are so low, considering they're a good team with an international superstar.
- My hypothesis from my previous post is that Shams and Woj each have capable Mavericks deputies in the Tims (Cato and MacMahon, respectively) and decide to leave that market alone
- Shams does have the highest percentage of Mavericks tweets in all three seasons however.
Now, you'll notice that there's two teams missing from the above graph: the Golden State Warriors and the Los Angeles Lakers. Here's the graphs for those two teams
. As you can see, they would skew the previous graph far too much. During the 2019 NBA season, 27% of Chris Haynes's qualifying tweets could be possibly linked to the Warriors, and 14% of his qualifying tweets could be possibly linked to the Lakers.
Here's the top 10 agents in terms of number of potential tweets concerning their clients.
|Agent ||Haynes ||Shams ||Woj ||Total |
|Rich Paul ||15 ||28 ||24 ||67 |
|Mark Bartelstein ||4 ||16 ||30 ||50 |
|Jeff Schwartz ||3 ||10 ||25 ||38 |
|Bill Duffy ||2 ||13 ||14 ||29 |
|Leon Rose ||1 ||12 ||15 ||28 |
|Aaron Mintz ||2 ||9 ||15 ||26 |
|Juan Perez ||5 ||10 ||8 ||23 |
|Aaron Goodwin ||11 ||8 ||1 ||20 |
|Steven Heumann ||1 ||6 ||12 ||19 |
|Sam Permut ||1 ||13 ||5 ||19 |
Woj has the most tweets directly connected to agents by far. It wasn't uncommon to see "Player X signs deal with Team Y, Agent Z of Agency F tells ESPN." The agents that go to Woj (and some of their top clients):
- Mark Bartelstein of Priority Sports (Bradley Beal, Kyle Lowry, Gordon Hayward)
- Jeff Schwartz and Sam Goldfeder of Excel Sports (Khris Middleton, Nikola Jokic, CJ McCollum and Kevin Love)
- Steven Heumann and Austin Brown of Creative Artists Agency (Andrew Wiggins, Chris Paul, Donovan Mitchell and Zion Williamson)
One thing I found very intriguing: 15/16 of tweets concerning an Aaron Turner client were reported on by Shams. Turner is the head of Verus Basketball, whose clients include Terry Rozier, Victor Oladipo and Kevin Knox. Shams also reported more than 50% of news relating to clients of Sam Permut of Roc Nation. Permut is the current agent of Kyrie Irving, after Irving fired Jeff Wechsler near the beginning of the 2019 offseason. Permut also reps the Morris brothers and Trey Burke.
As for Chris Haynes, he doesn't really do much agent news (at least not at the level of Woj and Shams). However, he reported more than 50% of news relating to clients of Aaron Goodwin of Goodwin Sports Management, who reps Damian Lillard and DeMar DeRozan. Here are the top 10 free agents from Forbes
, along with their agent and who I predict will be the first/only one to break the news.
|Player ||Agent ||Most Likely Reporter |
|Anthony Davis ||Rich Paul ||Too close to call, leaning Shams |
|Brandon Ingram ||Jeff Schwartz ||Woj |
|DeMar DeRozan ||Aaron Goodwin ||Haynes |
|Fred VanVleet ||Brian Jungreis ||Limited data |
|Andre Drummond ||Jeff Schwartz ||Woj |
|Montrezl Harrell ||Rich Paul ||Too close to call, leaning Shams |
|Gordon Hayward ||Mark Bartelstein ||Woj |
|Danilo Gallinari ||Michael Tellem ||Woj |
|Bogdan Bogdanovic ||Alexander Raskovic, Jason Ranne ||Limited data, but part of Wasserman, whose players are predominantly reported on by Woj |
|Davis Bertans ||Arturs Kalnitis ||Limited data |
Thanks for reading! As always with this type of work, human error is not completely eliminated. If you think a tweet was mistakenly removed, feel free to drop me a line and I’ll try to explain my thought process on that specific tweet! Hope y’all enjoyed the research!
Bitcoin trading also happens on a short time scale, with the possibility of quickly erasing support levels. Spot buying and futures positions differ in their intentions, as BTC price discovery is a matter of different time frames and sentiments. Bitcoin trading volumes remain high above $43 billion per day, though the bulk of trading has switched to Tether (USDT), which is used as a safe haven ... As per the daily confluence detector, there is a lack of strong resistance levels on the upside. This should be encouraging news for the buyers as they aim to take BTC into the $13,000-zone. BTC ... At block height 633,024, reached on 12:30 UTC Thursday, the Bitcoin blockchain adjusted its mining difficulty to 13.7 Trillion with a 9.29% drop and reached the lowest level since January this year. Bitcoin price on Overbit. Head and Shoulders pattern per se is a corrective pattern, and rarely signals a trend-reversal. The pattern when formed signals the exhaustion of bulls, especially when ... Bitcoin Breaking Below These Two Levels May Suggest the Bull Trend Is Over. Nick Chong Aug 22, 2020 04:27 . Share. Tweet. Send . Share. Bitcoin has struggled over recent days, falling below the pivotal supports at $11,500 and $11,700. As of this article’s writing, the leading cryptocurrency trades for $11,450, posting a 3% loss in the past 24 hours. It is actually outperforming most ...
SUBSCRIBE FOR MORE HOW MUCH - http://shorturl.at/arBHL Nviddia GTX 1080 Ti - https://amzn.to/2Hiw5xp 6X GPU Mining Rig Case - https://bitcoinmerch.com/produc... The Long-awaited bitcoin mining calculator is now here don't read below or else If you turn on the 🔔 i will be very happy please im begging you ---[👌]-[socia... Bitcoin finding difficulty to break the key structure level. Price rejection at this point can be very fatal. Must watch key area for BTC at the moment. 🐦 FO... Bitcoin Android / iOS Gameplay HD By Ketchapp Do you like Bitcoin GamePlay? This Bitcoin Ketchapp Game Review is for ios,iphone,ipad also... Download Android... Free Bitcoin Mining get $2200 free Fast Bitcoin Mining,Calculation table,Level up - Duration: 5 ... Make Money $8000 Per Day With Bitcoin Without Investment Earn 1 BTC In 1 Day - Duration: 5 ...