Our most recent take on James was last summer, after the Heat won their second consecutive NBA championship. We looked at the odds of James catching up to Michael Jordan’s six championship wins, the most in the salary-cap era (since 1984-85). We put the odds of the Heat winning the 2014 championship at about 33 percent, with the chances getting lower each subsequent year. So given this year’s loss, James might be making a smart move with his current opt-out. But given James’s win shares, we put his odds at winning a championship — regardless of the team — at about 20 or 25 percent. We were partially right — Wade and Bosh did show solidarity in a move to Miami. But we were wrong in predicting James’s decision. In a follow-up piece, we further evaluated the merits of James joining the Heat. There was a major hit to his popularity because of the way he announced the decision and because of Miami’s relatively small fan base. We argued this could have reduced his earning potential by $150 million. We conceded that “James may able to redeem himself through athletic success,” which turned out to be the case: He made $42 million from endorsements in 2013, which dwarfs the $28 million he made in 2009 when he was with Cleveland. The world is abuzz with the news that LeBron James is opting out of his contract with the Miami Heat. There’s plenty of speculation about what this means and where he might decide to go, if anywhere, but we thought it might be interesting to take a look at our own history of guessing James’s future. Here’s a roundup of predictions that we’ve made, and how accurate they turned out to be:The last time James opted out of his contract was in 2010, when he left the Cleveland Cavaliers and signed with the Miami Heat, joining Dwyane Wade and Chris Bosh there. Before he announced that decision, we analyzed his value as a player, as well as the different options available to him, and tried to determine the best route for him to take with regard to his personal brand. We decided he’d be best off trying to both win a championship and do so in a way that would be perceived as overcoming a challenge. We doubted that “forming a ‘dream team’ with Wade and Bosh in Miami would do that much for him,” so it was most likely that he would stay with Cleveland or join the Knicks, while Wade and Bosh would stick together in either Chicago or Miami.
Kevin Durant held back LeBron James.Miami’s LeBron James was not bad. He was good most of the game, in fact. But Kevin Durant was better, much better, all game long. The difference between the two superstar players helped Oklahoma City overcome a 13-point deficit in Game 1 of the NBA Finals to earn a 105-94 victory.The Thunder have a 1-0 series lead with Game 2 Thursday back in Oklahoma City. And it was Durant most responsible for OKC’s enviable position. He scored 17 of his 36 points in the final period on an dizzying array of shots. James had seven of his 30 when the game was decided. To wit: He had only one basket over the first 8:15 of the fourth.James averaged just three points in the fourth quarters of the Heat’s six-game loss to Dallas last year, taking almost all the heat for Miami’s Finals failure. He was good in this one, Durant was just better.And when fans chanted “MVP! MVP!” late in the game, they weren’t talking about James, the guy who won the regular-season award.They meant Durant, who is in a race with James for his first ring — and maybe the title of best player in the world is on the line, too.But first thing’s first. The game matters most, and James needs more help. Russell Westbrook shot just 10 of 24, but after halftime he and Durant outsored the entire Miami team 41-40. He finished with 27 points, eight rebounds 11 assists and just two turnovers. Meanwhile, Dwyane Wade had 19 points but shot just 7-of-19 for the Heat, while Shane Battier provided some rare offense by scoring 17 points, his high this postseason.Turning to a small lineup late in the third quarter, the Thunder improved to 9-0 at home in the postseason. Defensive ace Thabo Sefolosha helped defend James during the Thunder’s comeback, relieving Durant of the burden so he could focus on his scoring.A couple of key stats were significant. One, OKC outscored the Heat 24-4 on fastbreak points. Not good, especially considering how effective Miami has been on the break. The second stat is this: The Thunder scored 56 points in the paint. Really not good.
Jeremy Lin, the point guard who slept on his brother’s couch as he crafted one of the most dynamic 35-game stretches in NBA history, was allowed to walk from the New York Knicks to the Houston Rockets Tuesday night.The team that gave Lin the opportunity to emerge from a nobody on the bench into an international somebody – even gracing the cover of Time magazine — elected to not match a three-year, $25.1 million offer sheet he signed with the Rockets. So in essence, the player that charged “Linsanity” and the millions of off-the-court marketing and sponsorship dollars walked from the Knicks without the team receiving any compensation.Knicks sources and coach Mike Woodson said for a week that the team would match any offer to retain Lin. And considering Knicks owner James Dolan had doled out enormous money to the likes of Eddy Curry, Larry Brown and Isiah Thomas, it seemed reasonable to assume he would do so to keep Lin, a worldwide phenomenon. But, in the end, Dolan considered Lin is too expensive.Lin took to Twitter when it became official. “Much love and thankfulness to the Knicks and New York for your support this past year…easily the best year of my life#ForeverGrateful.”“Extremely excited and honored to be a Houston Rocket again!!#RedNation,” Lin added in another tweet.A team source told ESPNNewYork.com earlier this week that the third year of the Rockets’ offer — worth $14.8 million — caused the Knicks to consider letting Lin go. If the Knicks had matched the offer, they would have been subject to a luxury tax in the third year, potentially bringing their total out-of-pocket cost for the team’s salary to about $43 million in 2014-15.The Rockets’ offer to Lin would pay him $5 million in the first year, $5.225 million in the second and $14.8 million in the third, according to sources.The Knicks, realizing a $30-million penalty in Year 3 of the contract that would come with matching the offer, made a sign-and-trade deal to acquire another point guard, Raymond Felton. Felton played one season with the Knicks, but it was nowhere near as exciting as Lin’s New York stay.Lin, 23, was a revelation, performing brilliantly for a stretch that included 38 points in a win over the Los Angeles Lakers on national television and hitting the game-winning shot at Toronto. In between, he was mostly steady, although he did have a turnover problem. He averaged 14.6 points, 6.2 assists and 3.1 rebounds in 35 games with 25 starts before his season was cut short because of surgery to repair torn cartilage in his knee.An undrafted guard out of Harvard, Lin was cut twice in the preseason, once by Houston, and played in the D-League, before leading the Knicks to seven consecutive wins that electrified the city and the Asian community. His rise earned him a significant raise, from $788,000 last season to $5 million this season.His departure has sparked much debate among media and fans in New York, a debate that will rage for quite some time.
Dwight Howard‘s became NBA bad guy No. 1 after last season’s incessant complaining about his situation, where he would be traded to, his coach and just about anything else. Out of it came the nickname “Dwightmare.”Looking back on it now as a Los Angeles Laker, Howard said the whole protracted, tried saga made him stronger.“I think there’s a reason why everything happened the way it happened,” said Howard, who ended up traded to the Los Angeles Lakers. “So far it’s been an unbelievable experience for me. It’s like a dream come true.”Howard endured a tumultuous final season with the Orlando Magic in 2011-12. The three-time defensive player of the year demanded a trade to the Brooklyn Nets at the beginning of the season, but the two sides couldn’t reach a deal.I did want to go to Brooklyn. That’s a place where I told the Magic that I really wanted to go,” to Stephen A. Smith and Ryan Ruocco on ESPN NewYork 98.7 FM. “[But] I was traded to the Lakers, and I think it was a blessing in disguise.“I thought I was going to get traded at the beginning of the year, actually; that’s when I asked for it. But everything happened for a reason. I had to go through last year to get to where I’m at today. It’s made me a stronger and better person for it. I had to go through the hell and the stormy forecast to come out to a place like this . . . and I’m thankful for it.”Going through it, however, was tumultuous. The perception was that Howard changed his mind about his loyalties every other day, fostering an impression of unsteadiness.“The whole year a lot of people were making up a lot of stories about ‘This deal is getting close, that deal is getting close to being done,’ or whatever, but none of those deals were ever close,” Howard said.Eventually, despite myriad trade rumors all the way up to the trade deadline, Howard stunningly agreed to waive his early-termination option and committed to stay with Orlando through the 2012-13 season.“I think a lot of it was people just felt like I was going back and forth with the whole thing,” Howard said. “But the business side, people don’t understand, when you’re doing business you have to be a shark. You have to demand things. If you don’t, people will run over you, and that was a lesson that I learned.“At the end of the day, you can’t please everybody. There’s gonna be people happy about me staying, there’s gonna be people happy about me leaving. I’m over that now. I can’t control the way how people feel about me.”
Floyd Mayweather Jr. dominates Robert Guerrero to win title fight with lopsided decision.A 36th birthday, a year off — two months of which was spent in jail last summer on a domestic abuse conviction — would normally be the recipe for disaster for any fighter taking on a quality opponent. But Floyd Mayweather Jr., the pound-for-pound king, is not just any fighter. Instead, he looked the way he always looks: dominant.Mayweather easily retained the world welterweight championship with a masterful one-sided beatdown of interim titlist Robert Guerrero on Saturday night before a crowd of 15,880 at the MGM Grand Garden Arena.Guerrero had called out for Mayweather to fight him for the past couple of years, and you know what they say: Be careful what you wish for.Mayweather appeared vulnerable in his previous fight, when he was tagged repeatedly in a unanimous decision win against Miguel Cotto last May, leading some to suggest that perhaps Mayweather’s years of dominance were coming to a close. But after he took apart a man who was six years younger, physically bigger and unafraid to try to rough him up, forget about it.It was all Mayweather. All three judges, Jerry Roth, Duane Ford and Julie Lederman, scored it 117-111 for Mayweather. ESPN.com had it 119-109 for Mayweather.The fight was the first for Mayweather under a 30-month deal he signed with Showtime/CBS — after dumping longtime TV home HBO — that could see him fight up to six times and earn $200 million-plus. His next date is Sept. 14, also at the MGM Grand, and if Mayweather continues to perform as he did against Guerrero, it could be a deal well worth the investment, because “Money” dazzled — and earned a minimum of $32 million, tying his own record for biggest purse in boxing history.Mayweather rarely got hit cleanly, a testament to his father, Floyd Mayweather Sr., a defensive-minded trainer who returned to head his son’s corner for the first time since a junior lightweight title defense in 2001. They have been estranged on and off for years while Floyd Jr.’s uncle Roger Mayweather served as his trainer. But with Roger increasingly feeling the effects of diabetes and Floyd Jr. realizing he needed to go back to his defensive basics, the match worked.“I was really happy to be back with my father,” Mayweather said. “I knew after the Cotto fight, I was getting hit too much and my dad would help me get hit less. My defense was on point, and he told me just stick to your defense — the less you get hit, the better — and to box smart.”Read more: ESPN
Florida State’s handling of the sexual assault allegations of star quarterback Jameis Winston–and potential Title IX violations– are under investigation by the U.S. Department of Education’s Office of Civil Rights.The investigation stems from a complaint filed with the Office for Civil Rights by the accuser in the Winston case, according to USA Today.“Our client is particularly gratified by the OCR’s decision to investigate and look for discrimination and find remedies to it,” the accuser’s lawyer, Baine Kerr, said, “because her primary goal, from the beginning, has been affecting change that will make women at Florida State safer on campus.”The accuser, a Florida State student, said she was raped by Winston in December 2012. Deadspin.com reported Thursday that school officials met with Winston in January 2013 to discuss the matter. But the school might have violated federal law by delaying its investigation and meeting with Winston, alone, despite legal advice to the contrary.No charges were filed by State Attorney Willie Meggs against Winston, the reigning Heisman Trophy winner. Winston’s accuser filed her complaint with the Office of Civil Rights in March, 16 months after she first reported the incident.The OCR tells schools that “conduct may constitute unlawful sexual harassment under Title IX even if the police do not have sufficient evidence of a criminal violation. In addition, a criminal investigation into allegations of sexual violence does not relieve the school of its duty under Title IX to resolve complaints promptly and equitably.”Failure to comply with federal Title IX laws, which cover sexual harassment and violence, can result in a school losing its federal funding.“We can confirm that we have been notified of the OCR investigation; however, due to federal and state privacy laws the university cannot comment,” Florida State spokesman Browning Brooks said in a statement.The OCR will investigate all sexual assault complaints at the university during the last three years.Winston, who won the Heisman in the week after learning he would not face charges, is currently playing for the Seminoles baseball team. He will be a sophomore for the coming football season.It is unclear if the accuser will pursue a civil case against Winston or the university. Her family accused the Tallahassee Police Department of delaying the investigation and discouraging her from going forward with the case because of the public attention it would receive. Tallahassee police have defended their handling of the case.A civil case can still be filed after the OCR investigation. Kerr confirmed a Deadspin.com report that football players Chris Casher and Ronald Darby were recently charged with violations of the school’s Code of Conduct.Both Casher and Darby told police they witnessed the sex between Winston and the accuser. Casher told police that he attempted to join, but was told to leave by the accuser. Punishments under the code include expulsion from the school. Darby is a starting cornerback on the team.
(Wikimedia Commons)Kareem Abdul-Jabbar has heard about the Cleveland Cavaliers star Tristan Thompson cheating on Khloé Kardashian. He’s also heard people say that Kardashian should have seen it coming due to his athlete status and he wants folks to know that’s the wrong way to think.The retired NBA star explains the issue with boxing athletes as cheaters in a Monday, April 16 essay for Cosmopolitan, saying it’s still bad even if it doesn’t compare to making a sexist joke or buying into racist stereotypes.“This is a pervasive opinion people have of pro athletes — and not even the worst one,” he wrote of unfaithful pro-players. “But that doesn’t make it true or acceptable. Imagine how different the response would be if the comment were, ‘He’s Black.’ But athletes are expected to accept the insulting stereotypes, shut up, and dribble.“This offensive characterization is in no way on par with the much worse stereotyping of women, people of color, Muslims, Jews, immigrants, and the LGBTQ community,” he continued. “But when we allow any prejudice to pass unchallenged, we endorse all prejudice. That smarmy joke about dumb blondes that a colleague tells at lunch demeans all women. That ‘observation’ about smart Asian students or gay flair or rapist immigrants endorses an environment of lazy thinking that carries over into other decision-making that is ultimately detrimental to society. When we hear casual stereotyping and say nothing, we collude in the detriment.”After acknowledging the role confirmation bias plays into pushing forward the notion that athletes are all cheaters, Abdul-Jabbar also admitted that sports stars from all kinds of games have had their share of indiscretions.“What makes the Khloé Kardashian situation more poignant — and more ripe for tabloid fodder — is that this is not the first time she’s been publicly cheated on, not even the first pro athlete to cheat on her,” he said, possibly referring to Rashad McCants, who claimed the cheating story was made up for TV.“Certainly, wealth, power, and fame give athletes, politicians, musicians, actors, and filmmakers more opportunities to be unfaithful to their spouses,” Abdul-Jabbar continued. “But availability is no excuse for availing. Nor is it okay to use the bad behavior of some to characterize all. I’m not here to defend or condemn Tristan Thompson’s actions. If people feel the need to judge him, let them do so based on his behavior, not his profession or gender.”Meanwhile, folks on Twitter have been giving their take on the 71-year-old baller’s essay.“Right, so it’s our fault this stereotype persists, we make them cheat by our micro-aggressions,” someone said. “Got it. It’s a form of stereotyping, therefore a form of #racsim and so on, blah, blah, blah.”“Well they need to stop cheating so they aren’t stereotyped… if the shoe fits, buy a matching shirt,” someone else said.“Lol, but why isn’t this stereotype pushed on actors, musicians and politicians too?” someone else wondered. “They cheat just as often? Why not criticize the home wreckers, groupie, and gossip sites that take advantage of often false rumors to defame celebrities?”
PA PassLB Wasted yards * Fewer than 20 observationsSource: NFL Next Gen Stats 1st7.1– Distance traveled by a defender while biting on a play-action fake is a fairly precise way to quantify just how fooled a defender was on a play. Continuing to move toward the line of scrimmage when the offense is passing is a problem; defenders want to “get depth” as soon as they can if they identify pass. Any movement toward the line of scrimmage is usually wasted.After summing up the total distance traveled for each of the plays, I calculated that on the average play-action pass play, the middle linebacker covers 7.5 yards of wasted ground. In seven instances in our sample, teams ran 15 or more play-action plays in a single game. Those games would have offered the middle linebacker the most opportunities to figure out the play-action, but the average distance traveled was 8.2 yards — even higher than the overall average.I broke out the average wasted distance traveled by linebackers by the number of times a play-action pass was called in a game to see how teams reacted. It turns out that the wasted distance traveled was remarkably stable. 3rd7.8– 15th*11.1– The play-action pass is one of the most effective calls in all of football. The three teams that use the play-action the most — the Rams, the Patriots and the Chiefs, according to data from Sports Info Solutions — each locked down a first-round bye in the playoffs. Across the league in 2018, quarterbacks with at least 100 pass attempts average 1.39 yards per attempt more out of play-action than they do on all other plays.1And it’s not just guys like San Francisco’s Nick Mullens and his +4.2 yards per attempt play-action differential who are bringing up the group average. Of 40 qualifying quarterbacks, 77.5 percent have a higher yards per attempt on play-action passes than on other plays. This pattern of play-action success holds true for every year that we have data.22005 through 2018. Yet despite this success, the league average share of plays that are play-action passes is just barely above 20 percent.Why is play-action so effective? When defenders bite on a play-action fake, they move out of position for defending the pass and create clear lanes for the QB to throw to the intermediate and deep parts of the field.But NFL coaches tend to run them only a handful of times per game because they appear to believe that overuse of play-action will cause linebackers to stop biting on the fake. Diminishing returns will set in, defenders will stop respecting the run, and the superiority of play-action will vanish. But is this actually the case? Do linebackers start to ignore the fake handoff if they see it many times in a single game?Until very recently, we had a hard time answering this question with the data that was available. But in the past couple of weeks, the NFL released a tranche of Next Gen tracking data for 91 games from 2017 via its inaugural Big Data Bowl. Michael Lopez, the NFL’s director of analytics, spearheaded the effort to allow analysts to dig into the tracking data and mine it for insights. I was able to use this data to quantify the effect of play-action on the movement of middle linebackers — and to see if a high number of play-action plays had any effect on the outcome of the plays.I took each of the 1,235 play-action plays in the sample and isolated just the middle linebacker’s movement from snap to throw.3Both outside linebackers and box safeties are also influenced by play-action, but their run fit responsibilities are sometimes less clear, so for this study, I focused just on the middle linebacker. I measured the distance traveled by the defender while moving forward toward the line of scrimmage at any angle, and I stopped counting the distance as soon as he turned and retreated into coverage. If two linebackers were playing on the inside, I included only the player who moved the most toward the line of scrimmage during the play. Below are three animations that help illustrate the process.4Animation code courtesy of the NFL.The first shows the entire play with all players involved: 5th7.5– 6th7.3– More play-action passes do not mean fewer wasted yardsAverage yards wasted by the middle linebacker on each play-action pass in a game 10th7.9– There is a lot of good research showing that teams don’t run enough play-action. Most of the arguments for limiting its use are unsupported by the evidence. Now, thanks to the NFL’s Next Gen data, we can add evidence that middle linebackers won’t stop biting on the play-action, even if it’s used more than NFL coaches have been comfortable running it. 9th8.4– The third shows when I stopped counting the linebacker’s movement as “wasted” for the purposes of the study: 16th*4.7– 4th7.1– 11th7.2– 14th*6.4– Linebackers bite just about the same amount the 11th time a play-action pass is called in a game as the first time it’s called. It’s only after we get to 12 play-action passes in one game that things start to get wonky — but that may be because of the small sample sizes of those instances.Across the entire sample of 91 games and 1,235 plays, I found no correlation at all between the number of times a team ran the play-action and total yards of wasted ground by middle linebackers.5R-squared of 0.001775, p-value of 0.5766. We’d love more data to examine, to look closer at what happens when more play-actions are run. But given what we know about the effectiveness of the play, the self-imposed threshold set by play-callers of roughly six to nine play-action fakes per game is likely too low.Stopping the run is a major focus at every level of football, and the NFL especially makes it a high priority to effectively defend the run. Teams do this by coaching their linebackers and box safeties to play the run first in nonobvious passing situations. This emphasis on run stopping comes at a cost, however. Defenders must read their “run keys” — movements by the offense that indicate a run is coming — and react quickly to fill their gaps and prepare to make a physical play. It could be the case that defenders simply don’t think about how often the team is faking the run but instead just read and react to their run keys.To play fast in the NFL, it’s often said, you can’t think but instead must react based on instinct and training. Perhaps that instinctual reaction explains why play-action continues to be effective no matter how often it’s used. It’s also probably the case that certain teams and players are more susceptible to play-action than others, and smart NFL teams will identify and exploit their opponents’ tendencies.Those smart NFL teams should also pay attention to exactly how they use the play-action. According to the Sports Info Solutions data, passes thrown 7 yards deep or less are caught less frequently on play-action than on other passes. This could be because defenders have moved toward the line of scrimmage and are in better position to make a play on the ball. Play-action is only more effective than other passes when the ball travels at least 8 yards in the air — over the head of the linebackers who’ve been fooled. 12th*8.5– The second shows the entire play with the middle linebacker and quarterback isolated: 8th7.2– 2nd7.7– 13th*10.5– 7th7.2– Check out our latest NFL predictions.
2San Francisco114201317Jacksonville141999 Well, that didn’t take long. We’re just one week into the 2017 NFL season, and there’s already a new team atop FiveThirtyEight’s Elo ratings. The Kansas City Chiefs rose to the top slot after beating the preseason No. 1 New England Patriots 42-27 in last Thursday’s season opener, and now Elo gives them a league-best 14 percent chance of winning the Super Bowl.OK, so maybe you shouldn’t bet the farm on that Chiefs championship just yet. (The Vegas books still list K.C. in a tie for the seventh-best Super Bowl odds of any team, alongside the Atlanta Falcons.) But Kansas City occupying first place in an NFL power ranking is a pretty rare sight nonetheless. Since the 1970 NFL-AFL merger, the Chiefs have only held the No. 1 slot for eight total weeks,1Including this week. the same as the New York Jets and Tampa Bay Buccaneers. Compare that with the team the Chiefs beat — the Pats, who lead all post-merger franchises with 137 total weeks at No. 1 — and you start to get a sense for just how unusual a top-ranked K.C. squad is.The last time the Chiefs ranked first in Elo? Week 11 of the 2003 season, when Kansas City had just crushed the Cleveland Browns to move to 9-0 on the year. It was K.C.’s fourth consecutive week at No. 1, but the celebration wouldn’t last: Dick Vermeil’s Chiefs were upset by Cincinnati on the road the following week, part of a 4-3 stretch to close the regular season. And in the playoffs, they lost an epic shootout with Peyton Manning and the Colts in which neither team punted. K.C. fell to 7-9 the next season, and the offensive firepower briefly captured with Trent Green and Priest Holmes was never quite achieved again. CAR65CAR71CAR+1.4– 7Washington51199221Kansas City82017 13Seattle30201629Atlanta0— FiveThirtyEight vs. The CrowdLast week, we rolled out a new game in which we invite you, the readers, to make picks against our Elo algorithm — as well as each other. (The more confident you are in your choices, the more points you win.) After each week, we’ll tally up everyone’s scores, and you can see where you stand relative to Elo and your fellow players.As a side benefit of this exercise, we can also use the results of your picks to figure out which games and teams the crowd most disagreed with Elo about — and who was right. For instance, in Week 1, readers were all over the woefully high 57 percent win probability Elo gave the Andrew Luck-less Colts on the road against the L.A. Rams. (In Elo’s defense, it doesn’t factor in key injuries like Luck’s.) Los Angeles ended up crushing Indy 46-9, costing our algorithm a lot of points in the process.Here are all the games from the opening week of the season, in order of how much readers outsmarted Elo:3The Chiefs-Patriots matchup isn’t included here because our game launched last Friday, after the season opener was played. 10Minnesota42201625Buffalo41991 MIN59MIN54MIN-6.7– 6Indianapolis52200921N.Y. Jets82010 our PICKWIN PROB.READERS’ PICKWIN PROB.ACTUAL WINNERREADERS’ NET PTS 8Miami50198524New Orleans62010 9Oakland43200225Cincinnati41982 HOU74HOU75JAX-3.1– IND57%LAR53%LAR+8.2– 14L.A. Rams27200229Detroit0— 4Green Bay67201120Baltimore92009 CIN63CIN58BAL+4.6– 15Philadelphia22200429Houston0— Kansas City is far from being the franchise with the least time spent in first place. Among teams that have reached the top slot, five — New Orleans, Cincinnati, Buffalo, Carolina and Arizona — have spent fewer weeks in first than K.C. Four other teams — Detroit, Houston, Cleveland and (amazingly2Given that they’ve made two Super Bowls.) Atlanta — have never made it to No. 1 in the post-merger era.Maybe the Chiefs will have the staying power of the Patriots, whose 11-week reign at the top K.C. snapped this week. Or maybe they’ll be like last year’s Minnesota Vikings, who spent one solitary week at No. 1 (Week 6, if you’re keeping track) before fading away. DAL65DAL60DAL-6.7– Elo’s dumbest (and smartest) picks of Week 1Average difference between FiveThirtyEight reader points won and Elo points won per matchup Source: Pro-Football-Reference.com 11Chicago40198827Carolina22015 ATL72ATL80ATL+1.9– GB61GB56GB-6.6– 16Tennessee17200829Cleveland0— WSH58WSH52PHI+4.6– OAK51OAK54OAK+0.5– 12Denver37201628Arizona12015 PIT76PIT86PIT+1.9– DET54ARI52DET-8.3– The Week 1 winner is …Congratulations to Dante Sblendorio of Livermore, California, who absolutely humiliated Elo with this week’s high score of 225 points. Dante, a 25-year-old physicist, went for an extremely aggressive approach that paid off: He picked 12 of 14 games correctly, all at 100 percent confidence. Too bad he wasn’t in Vegas.Remember: You can start playing the prediction game this week, even if you didn’t get your picks in last week.Check out our latest NFL predictions. 5Pittsburgh59200921Tampa Bay82003 3Dallas108199619N.Y. Giants112008 Most weeks spent at No. 1 in NFL Elo ratings, 1970-2017 BUF64BUF71BUF+2.6– 1New England137201717L.A. Chargers142009 DEN74DEN66DEN-8.5– RKTEAMWEEKSLAST YR AT NO. 1RKTEAMWEEKSLAST YR AT NO. 1
Chess is an antique, about 1,500 years old, according to most historians. As a result, its evolution seems essentially complete, a hoary game now largely trudging along. That’s not to say that there haven’t been milestones. In medieval Europe, for example, they made the squares on the board alternate black and white. In the 15th century, the queen got her modern powers.1Long ago, the queen could move only one square diagonally at a time.And in the 20th century came the computer. Chess was simple enough (not many rules, smallish board) and complicated enough (many possible games) to make a fruitful test bed for artificial intelligence programs. This attracted engineering brains and corporate money. In 1997, they broke through: IBM’s Deep Blue supercomputer defeated the world champion, Garry Kasparov. Humans don’t hold a candle to supercomputers, or even smartphones, in competition anymore. Top human players do, however, lean on computers in training, relying on them for guidance, analysis and insight. Computer engines now mold the way the game is played at its highest human levels: calculating, stodgy, defensive, careful. Or at least that’s how it has been. But if you read headlines from the chess world last month, you’d think the game was jolted forward again by an unexpected quantum leap. But to where?The revolutionary is known as AlphaZero. It’s a new neural network, reinforcement learning algorithm developed by DeepMind, Google’s secretive artificial intelligence subsidiary. Unlike other top programs, which receive extensive input and fine-tuning from programmers and chess masters, drawing on the wealth of accumulated human chess knowledge, AlphaZero is exclusively self-taught. It learned to play solely by playing against itself, over and over and over — 44 million games. It kept track of what strategies led to a win, favoring those, and which didn’t, casting those aside. After just four hours of this tabula rasa training, it clobbered the top chess program, an engine called Stockfish, winning 28 games, drawing 72 and losing zero. These results were described last month in a paper posted on arXiv, a repository of scientific research.Within hours, the chess world descended, like the faithful to freshly chiseled tablets of stone, on the sample of 10 computer-versus-computer games published in the paper’s appendix. Two broad themes emerged: First, AlphaZero adopted an all-out attacking style, making many bold material sacrifices to set up positional advantages. Second, elite chess may therefore not be as prone to dull draws as we thought. It will still be calculating, yes, but not stodgy, defensive and careful. Chess may yet have some evolution to go. For a taste of AlphaZero’s prowess, consider the following play from one of the published games. It’s worth emphasizing here just how good Stockfish, which is open source and was developed by a small team of programmers, is. It won the 2016 Top Chess Engine Championship, the premier computer tournament, and no human player who has ever lived would stand a chance against it in a match.It was AlphaZero’s turn to move, armed with the white pieces, against Stockfish with the black, in the position below: In the middle of the AlphaZero paper is a diagram called Table 2. It shows the 12 most popular chess openings played by humans, along with how frequently AlphaZero “discovered” and played those openings during its intense tabula rasa training. These openings are the result of extensive human study and trial — blood, sweat and tears — spread across the centuries and around the globe. AlphaZero taught itself them one by one: the English opening, the French, the Sicilian, the Queen’s gambit, the Caro-Kann.The diagram is a haunting image, as if a superfast algorithm had taught itself English in an afternoon and then re-created, almost by accident, full stanzas of Keats. But it’s also reassuring. That we even have a theory of the opening moves in chess is an artifact of our status as imperfect beings. There is a single right and best way to begin a chess game. Mathematical theory tells us so. We just don’t know what it is. Neither does AlphaZero.Yet.DeepMind was also responsible for the program AlphaGo, which has bested the top humans in Go, that other, much more complex ancient board game, to much anguish and consternation. An early version of AlphaGo was trained, in part, by human experts’ games — tabula inscripta. Later versions, including AlphaZero, stripped out all traces of our history.“For a while, for like two months, we could say to ourselves, ‘Well, the Go AI contains thousands of years of accumulated human thinking, all the rolled up knowledge of heuristics and proverbs and famous games,’” Frank Lantz, the director of NYU’s Game Center, told me. “We can’t tell that story anymore. If you don’t find this terrifying, at least a little, you are made of stronger stuff than me. I find it terrifying, but I also find it beautiful. Everything surprising is beautiful in a way.” AlphaZero is already behind by two pawns, and its bishop is, in theory, less powerful than one of Stockfish’s rooks. It’s losing badly on paper. AlphaZero moved its pawn up a square, to g4 — innocuous enough. But now consider Stockfish’s black position. Any move it makes leaves it worse off than if it hadn’t moved at all! It can’t move its king, or its queen, without disaster. It can’t move its rooks because its f7 pawn would die and its king would be in mortal danger. It can’t move any of its other pawns without them being captured. It can’t do anything. But that’s the thing about chess: You have to move. This situation is known as zugzwang, German for “forced move.” AlphaZero watches while Stockfish walks off its own plank. Stockfish chose to move its pawn forward to d5; it was immediately captured by the white bishop as the attack closed further in.You could make an argument that that game, and the other games between the two computers, were some of the strongest contests of chess, over hundreds of years and billions of games, ever played.But were they fair? After the AlphaZero research paper was published, some wondered if the scales were tipped in AlphaZero’s favor. Chess.com received a lengthy comment from Tord Romstad, one of Stockfish’s creators. “The match results by themselves are not particularly meaningful,” Romstad said. He cited the fact that the games were played giving each program one minute per move — a rather odd decision, given that games get much more complicated as they go on and that Stockfish was programmed to be able to allocate its time wisely. Players are typically allowed to distribute their allotted time across their moves as they see fit, rather than being hemmed in to a specific amount of time per turn. Romstad also noted that an old version of Stockfish was used, with settings that hadn’t been properly tested and data structures insufficient for those settings.Romstad called the comparison of Stockfish to AlphaZero “apples to orangutans.” A computer analysis of the zugzwang game, for example, reveals that Stockfish, according to Stockfish, made four inaccuracies, four mistakes and three blunders. Not all iterations of Stockfishes are created equal.DeepMind declined to comment for this article, citing the fact that its AlphaZero research is under peer review.Strong human players want to see more, ideally with the playing field more level. “I saw some amazing chess, but I also know we did not get the best possible,” Robert Hess, an American grandmaster, told me. “This holds true for human competition as well: If you gave Magnus [Carlsen] and Fabiano [Caruana] 24 hours per move, would there be any wins? How few mistakes? In being practical, we sacrifice perfection for efficiency.”Chess.com surveyed a number of top grandmasters, who were assembled this month for a tournament in London (the home of DeepMind), about what AlphaZero means for their profession. Sergey Karjakin, the Russian world championship runner-up, said he’d pay “maybe $100,000” for access to the program. One chess commentator joked that Russian president Vladimir Putin might help Karjakin access the program to prepare for next year’s Candidates Tournament. Maxime Vachier-Lagrave, the top French player, said it was “worth easily seven figures.” Wesley So, the U.S. national champion, joked that he’d call Rex Sinquefield, the wealthy financier and chess philanthropist, to see how much he’d pony up.“I don’t think this changes the landscape of human chess much at all for the time being,” the grandmaster Hess told me. “We don’t have the ability to memorize everything, and the games themselves were more or less perfect models of mostly known concepts.”In some aesthetic ways, though, AlphaZero represents a computer shift toward the human approach to chess. Stockfish evaluated 70 million positions per second, a brute-force number suitable to hardware, while AlphaZero evaluated only 80,000, relying on its “intuition,” like a human grandmaster would. Moreover, AlphaZero’s style of play — relentless aggression — was thought to be “refuted” by stodgy engines like Stockfish, leading to the careful and draw-prone style that currently dominates the top ranks of competitive chess.But maybe it’s more illustrative to say that AlphaZero played like neither a human nor a computer, but like an alien — some sort of chess intelligence which we can barely fathom. “I find it very positive!” David Chalmers, a philosopher at NYU who studies AI and the singularity, told me. “Just because it’s alien to us now doesn’t mean it’s something that humans could never have gotten to.”