And the walls came tumbling down. (Nate Silver, Day 18)

Mexico had the Oranje right where they wanted them. They had a well earned one-nil lead, and were in full defense mode. Unfortunately, even the wall of Ochoa isn't impenetrable. Wesley Sneijder drilled a laser home in the 88th minute and Mexico was broken. The Netherlands dominated the next few minutes and Arjen Robben drew an exceedingly soft, but technically legitimate penalty. The comeback is actually sightly more impressive than it appears as it was in heat that felt like 40 Celsius on the field. I can see why Mexico wore down. They expended a lot of energy in a first half they mostly controlled. When Huntelaar knocked the penalty home, Mexico's collapse was complete. In the end, they didn't deserve better as it is just too hard to "turtle" against this Dutch squad. MC_Mex_Ned Now the Oranje will face Costa Rica, who managed to outlast a Greek squad that simply couldn't put the ball in the net. Over the course of qualifying they were about three goals short of their xG, a fear they replicated in only four games at the World Cup. Interestingly enough, Costa Rica only had two shots on goal against Greece, and the one that went in was a soft roller that the defense and the goalie both would have had a shot at HAD THEY TRIED TO BLOCK IT! Seriously, Sokratis Papastathopoulos just stood there and watched it roll by. As his namesake said: "Let he that would move the world first move himself." -Socrates Greek goalie Orestis Karnezis didn't exactly cover himself in glory on the goal, but I'll get to him shortly. Oscar Duarte picked up yellow cards in the 42nd and 66th minutes and was sent off. That left Greece with 24 minutes to generate a goal. While their offense wasn't quite as aggressive as one would expect, they did pick up a goal when Papastathopoulos knocked in a rebound in the 91st minute. That left Greece with a 30 minute overtime to take advantage of a tired and undermanned squad. Alas, despite a few good opportunities, they couldn't get a second goal and went into penalty kicks. There Costa Rica managed to get all five shots past Karnezis. I know penalty kicks are hard to stop, but in the end, six of the seven Costa Rican shots on goal went in, and now Greece is going home. That they dominated the game according to xG is going to be of little consolation: MC_GRE_CR There are extenuating circumstances, but Costa Rica has only managed five shots on goal over their last two games. Yes, the game vs. England wasn't a must-win, but they still wanted a point to win the group and avoid Colombia. After scoring and then going down a man, it was unlikely that they'd get many opportunities. Even so, you cannot trust your defense to hold against the Netherlands for 90 minutes, and if they can't find the offense that was able to light up Uruguay and Italy, they will go home in the quarterfinals. Nate Silver continued his unbeaten streak in the knockout rounds and ran his tab to $933,329. I don't expect his streak to continue, as he is showing some faith in the large underdogs: Game 53: Nigeria to advance, 6.73%: Risking $62,813 to win $276,377 Game 54: Algeria to advance, 5.60%: Risking $48,749 to win $341,242 If either of those wins, Nate will finally finish a day with a million dollar bankroll, but let's not kid ourselves: Germany is facing France in the quarterfinals. France has been dominant this tournament and Nigeria doesn't have the tools to stop them. Germany has Algeria seriously outclassed. If I had access to 538's goal distribution, I suspect I'd be backing Algeria +1.5 goals, but unfortunately my only option is to back them to advance. So be it. Earlier today people were commenting on how many players various club teams had sent to the World Cup. You know how many West Ham sent? Zero. Perhaps this is the reason: london_falling That is a police procedural where a special squad is formed to track down the "Witch of West Ham". I highly recommend it. I suppose I should be happy West Ham didn't flirt with relegation last season. OK, back to the World Cup. Germany and France should cruise, but upsets can happen. Brazil wasn't expecting to have to go to penalty kicks to defeat Chile.  I'll enjoy the games tomorrow, but the real treat is the afternoon game on Tuesday when the United States attempts to defeat Belgium for the first time since 1930. Seriously. Until then... Seth Burn @sethburn on twitter

Brazil survives. James Rodriguez is Gifted. (Nate Silver, Day 17)

The Brazil-Chile game was far more interesting than the result suggests. Brazil scored early on a beautiful set piece off of a corner kick. It was technically an own goal, but it would have been a goal had the defender not touched the ball. Brazil would have had a couple of other opportunities with Neymar, had he been willing to shoot with his left foot. Chile equalized off of a terrible miscue from Hulk. In the second half Hulk scored, but it was correctly ruled that it was a handball. In fact, not only was it a handball, but Hulk barely got his leg on the ball and was lucky it rolled in. That was the best scoring chance of the second half. Chile played quite defensively and were able to make it to extra time. For most of extra time they played for a shootout, but managed an excellent scoring chance in the 120th minute. They rocketed a shot off the crossbar and barely missed a goal. In the shootout, Júlio César was simply better than Claudio Bravo. I don't want to suggest Chile played Brazil fairly evenly. They didn't: MC_Bra_1 However, they were able to prevent Brazil from scoring and that's somewhat disconcerting for Brazil, particularly given their lackluster play in the group stage. Chile and Mexico played Brazil to a draw over a period of 210 minutes. They need to be able to trust someone other than Neymar. That brings me back to Hulk. Before penalty kicks I tweeted: Lo and behold, Hulk tried to smash one down the center and the goalie blocked it. Hulk has brought Brazil nothing but pain. That's only slightly better than Fred, who has brought Brazil pretty much nothing at all. Neymar is great, but it is hard to win a World Cu with only one elite player. In fact he might not even be the best player on the field in the quarterfinals: Oh my. He's the most gifted player I've seen since the Sentanta sports version of Wayne Rooney. [youtube id="kj9nb8fgMVI" width="633" height="356"] You know what the worst part about all of this is from Uruguay? Colombia had a player who could put the ball into the net, while they didn't. That led to a two-nil scoreline with this shot chart: MC_Col_1 I won't miss Uruguay. I will say that If I were Brazil, I'd be very concerned about facing Colombia. I understand that Colombia hasn't faced the toughest path to get here, but they've been handling their opposition with impressive ease. From here on in Brazil doesn't have any more easy games. Speaking of easy games, Nate Silver ran his tab to $744,887. Not bad. He'll be investing a portion of that sum tomorrow: Game 51: Netherlands to advance, 25.52%: Risking $190,095 to win $114,515 Game 52: Costa Rica to advance, 9.70%: Risking $83,362 to win $73,927 (Updated) Both games feature clashes between offensive teams (Costa Rica, the Netherlands), and defensive teams (Mexico, Greece). In both cases Nate is backing the offensive team, and in both cases the offensive team is favored. I definitely agree with the Costa Rica play as Greece have been exceedingly feeble on offense. Mexico are much more dangerous, but then again, so are the Dutch. I'm looking forward to both games and another great day of futbol from the World Cup. Seth Burn @sethburn on twitter

Gifolution: The Graceful Aging of #NED Star Arjen Robben

Arjen Robben is amazing. He's amazing because he only has one real dribbling trick in his bag, yet no one can stop it. He's amazing because he went bald around age 20. He is amazing because he bebopped around most of Europe's largest clubs until finally finding a home at Bayern Munich. More than anything, however, the dude is an amazing footballer. Look at the gifolution below and the consistency of outstanding production even crossing the age 30 barrier. It is rare for speed players to maintain those gifts without making major adjustments to their game, but Robben seems the same, only maybe a little more relaxed and a little more lethal. His 2010-11 radar covers so much space, it looks like an umbrella with a broken spoke. Very few players in the world have that type of production, and their names are Messi, Suarez, and Aguero. His rates that year were that good. Here's one more amazing thing about Robben before I leave you today: In NONE of the last five seasons has he played 20 90s or more. Always a little bit injured. Always a lot brilliant. Arjen Robben. Arjen-Robben-0914

Gifolution: Chilean Dynamo, Arturo Vidal

Vidal is one of the most important players for both Chile and three-times Serie A champions Juventus. He's a force of nature on the pitch, posting one of the highest tackle rates per game of any player in the world,  but has he always been this good?

As you can see from the gif below, not quite. 2010-11 is when we first see that he's both a scorer and a strong ball winner in midfield, while 2012-13 was what really got him recognition as one of the most complete midfielders in the world. This past season, you can see a small drop in most of his stats. Whether that's a result of the knee injury he picked up that needed surgery before the World Cup, or simply a difference in tactics is unknown.

Whatever the case, Chile need him to be his usual dynamic self to stand a chance of getting past Brazil today.



Ian Darke is an idiot (Nate Silver, Day 16)

Imagine for a second that you are sitting in a Chinese restaurant eating "spicy red snapper"* and drinking whiskey. It is lunchtime and you have a stress headache and are already a bit buzzed. You are watching two screens at once, one showing USA-Germany, the other showing Portugal-Ghana. You are desperately rooting for the United States to advance. They are down one-nil to Germany, but Ghana is also down one-nil. It would take at least two "bad goals" to knock you out. Suddenly, in the 57th minute, Ghana scores! Rage and fear fills you. And then, Ian Darke with his infinite lack of wisdom says, "That's good for the Americans as a draw between Portugal and Ghana puts them through!"

In that scenario, you might very shortly learn the Chinese expressions for "Sit the fuck down" 坐在他妈的下来 and "Shut the fuck up" 他妈的闭嘴

Thankfully, after the longest 23 minutes ever recorded, Ronaldo scored and I was able to stop doing butt kegels. Ze Germans win the group, and we get to prepare for Belgium.

Speaking of Belgium, they were able to defeat South Korea despite being down to 10-men. I don't want to focus on the poor performance of the Asian contingent in the World Cup, but it was abominable. Belgium won their group, advancing along with Algeria. I greatly enjoyed the Russia-Algeria game. Russia needed a win and played like it, scoring on a Kokorin header in the 6th minute. They continued attacking, but Algeria was able to match them and avoided falling further behind. In the 60th minute, Islam Slimani scored on a header of his own, and Algeria were back in command. Russia made a game effort of it after that, but weren't able to break through and Algeria advanced to the knockout round to take their crack at Ze Germans. All of this left Nate Silver sitting with a comfortable $647,555.

The knockout rounds open with the hosts attempting to crack the dwarf fortress. I think this is a pretty bad matchup for Chile, as they've been taking advantage of home field advantage in their earlier games. They've been able to be fairly aggressive, while picking up fewer fouls and cards than one might expect. That won't be the case tomorrow, particularly if they pick on Neymar. On the other hand, they could probably kill Hulk and only get a yellow, so long as Brazil still had a substitution left. This line has moved steadily towards Brazil, so Nate is only ending up with a moderate stake on them:

Game 49: Brazil to advance, 19.00%: Risking $123,055 to win $35,153

Originally it appeared Uruguay was going to be a fairly competitive foe against Colombia, but FIFA felt compelled to act against Suarez, who is now gone from the tournament. It seems likely that his team will join him. Nate agrees:

Game 50: Colombia to advance, 16.3%: Risking $111,281 to win $62,179 (Updated)

Reminder: If Brazil advances, I will update the Colombia stake using the market prices available at the time.

I am very much looking forward to the next round of futbol. Time for the elimination games to begin. 🙂

*No way the "red snapper" stands up to a DNA test, but it's very tasty so I don't care.

Gifolution: Bastian Schweinsteiger 2009-2014

An ever-present fixture in both the German and the Bayern Munich midfields, Bastian Schweinsteiger has a resume like few other players in football. Still just 29, he has 103 caps for the German squad and 322 first team appearances for Bayern.  He also has won 7 Bundesliga titles, 7 DFB -Pokals, and a Champions League medal plus another finals appearance. The one thing he has never done is appear in the Final of a World Cup. Will 2014 be that year? Schweinsteiger-09-14

Eine Hand wäscht die andere? (Nate Silver, Day 15)

Lionel Messi is very good at football. That's nice, because Argentina's back line is surprisingly awful. It probably won't be enough to sink them vs. Switzerland, but it is something to keep in mind when the competition gets tougher. Despite losing to Argentina, Nigeria advanced anyway. They'll get get a crack against a tough French team that denied Ecuador's bid to advance. France and Ecuador played to a scoreless draw, which opened up the door for Switzerland, who demolished Honduras three-nil, and get their crack at stopping Messi. I wish them best of luck. Iran had an outside shot of stealing Nigeria's bid, but a 3-1 loss to Bosnia-Herzegovina did them in. Despite going 3-1, Nate finished the day slightly down, dropping to $707,574. He'll take another crack at the final day of group play: Game 45: USA, 7.53%: Risking $40,088 to win $456,203 Game 46: Ghana +0.5 goals, 24.6%: Risking $162,003 to win $148,627 Game 47: Belgium, 21.96%: Risking $86,250 to win $62,100 Game 48: Algeria +0.5 goals, 22.3%: Risking $87,969 to win $79,972 Assuming no line moves, an 0-4 day would leave Nate at $331,264. A 4-0 day would raise the tab to $1,681,270.  Exciting stuff, but not nearly as exciting as the drama of USA-Germany. Both teams control their own destiny, and both advance with a draw. However, a draw would leave the USA with the tough take of Belgium (most likely) in the round of 16, followed by Argentina (most likely) in the quarterfinals. A win would get them Russia or Algeria (more on that shortly), followed by France (most likely). Don't get me wrong, if Germany wanted to play for a draw I'd acquiesce, but there is something to be gained in victory. However, in pure futbol terms, this is a mismatch. Not only are the Germans much better than the Yanks, but they're better rested and healthier as well. Germany could simply go all out with reasonably little fear of losing, and even if they did, they have excellent tiebreaks. If Germany lost one-nil, Ghana would need a four-nil victory to advance. Portugal would need to win by 7. Germany is going to advance, but they'd like to win the group. My guess is if the game is tied late, both teams park the bus. However, the first half and early in the second should be highly competitive and I don't expect it to go well for the Americans. That means they'll be rooting for... Portugal. Portugal is in a very tough spot. They need the Germans to blow team USA out, while they crush Ghana. It's possible, but Portugal are probably going home. Ghana needs to win, but if they do they're actually in decent shape. They'd need Germany to win, and at that point it would come down to tiebreaks. If America loses by one in a high scoring game, Ghana would need to win by two (or one in a very high scoring game). If both teams end up with identical tiebreaks, Ghana will advance because there is no way the USA will win in a drawing of lots. It's just not happening. You know what scenario would amuse me the most? Germany crushing, while Ghana draws in excruciating fashion. A Portugal goal in the final minute would be poetic, don't you think? 🙂 Group H is interesting as well. Belgium is highly likely to win the group. South Korea has an outside shot advancing, but they'd need Algeria to fail to win, and then beat either Algeria or Russia on tiebreaks. Oh, they'd also have to beat Belgium. The real drama is Russia vs. Algeria. Russia needs a win, Algeria needs a draw or better. Algeria is a slight favorite to get the draw, but it's close. Whomever gets the result will likely get a game vs. Germany as a reward. Good luck. Gameday. Either team USA advances or I'll be drunk by the time Group H kicks off. Then again, if we advance, I'll be drunk by the time Cleveland screws up the first pick in the draft. Wish my liver luck and enjoy the futbol!

Au Revior Republic of Côte d'Ivoire (Nate Silver, Day 14)

Iran feels your pain. The United States thinks you could taste an advance to the knockout round. I've mentioned futbol could be cruel, but this was an absolute disaster. Yes, it was a penalty. The Greek offense that hadn't scored a goal in the World Cup managed to get two. It would have been poetic had Greece managed to advance to the knockout round scoring only one goal, but alas. It's also a tremendous upset that Greece played in the only truly enjoyable game of the day, but there you go.

If you wanted to defend Colombia-Japan, that's reasonable. However, Japan were awful this World Cup and Colombia's 4-1 score line was a fair reflection of the quality of the two squads. Colombia deserves to win the group, while Japan deserves to go home.



If there is any justice, Colombia will face a Uruguay squad sans Luis Suarez. In fact, Suarez should have been given a red and should now be banned for the rest of the tournament. I can't say Italy deserved a draw, especially since they wasted time at every opportunity (and should have earned multiple yellows in the process). Just another ugly game in a day full of them.

The ugliest of all was England-Costa Rica. No offense to Costa Rica, who played hard and won the group with a draw. However, I expected England to put together a more dynamic effort than that. The two teams combined for six shots on goal. Just a lousy day of futbol all around.

Lousy, except perhaps for Nate Silver. Despite the late loss by the Ivory Coast, Nate managed to run his tab to $721,317. Not bad. Today's slate:

Game 41: Argentina, 8.4%: Risking $50,970 to win $27,524

Game 42: Bosnia +0.5 goals, 19.15%: Risking $126,059 to win $44,231

Game 43: Honduras +0.5 goals, 24.03%: Risking $172,890 to win $354,883 (Updated)

Game 44: Ecuador +0.5 goals, 9.28%: Risking $55,912 to win $87,392 (Updated)

Nigeria could win the group with a dominant result over Argentina. They could also be eliminated if they lose and Iran beats Bosnia. In other words, they are in the same situation that the USA is against Germany. Unlike Germany, Argentina has clinched their spot in the next round, but they want to win the group to avoid having to face Brazil before the final.

Iran is a slight favorite to make the knockout round if they win. That's a pretty big if.

Group E is actually fairly interesting. While France hasn't technically clinched, they're almost certainly advancing to the next round. While Honduras hasn't technically been eliminated, they're very likely out. The real question is whether or not Ecuador can stay ahead of Switzerland. They currently have a 2-goal lead in goal differential, but now they have to face France. Switzerland is likely to win vs. Honduras, so there is a lot of pressure on Ecuador. Best of luck to Los Amarillos.

Back to Nate. After 40 games Nate Silver has increased his bankroll by an average of 5.0638% per-game. If that continues, he'll finish with over 2.36 million dollars. Thank goodness I don't believe in jinxes.

I do want to take a moment to crowd-source a question: How would you communicate the concept of the huge swings from full-kelly betting in an image? So far I have the idea of a pirate ship on the seas cresting a wave while other ships are getting tossed and swamped, but I don't think that quite captures the concept. If anyone has any thoughts on the subject, please comment here or tweet @sethburn with your ideas.

Enjoy today's games. Hopefully we're just a day away from the most boring draw in history, but I'll get to that tomorrow.

Introducing Possession-Adjusted Player Stats

We’ve known for a few months now that defensive rate stats at the team level are fairly useless. Tackles and interceptions – the base defensive action metrics -  have almost no correlation between shots allowed or goals scored. However... once you adjust these numbers at the team level for amount of possession, the r-squared when compared to shots conceded, goals allowed, etc generally shoots up in the .4 range, which is about the same as you get for possession itself. This isn’t perfect, but it’s not meaningless, which is a positive.

The reason we see a much higher correlation when we adjust defensive stats for possession is simple: opportunity. If your team has possession of the ball, you can’t rack up defensive rate stats.  Teams that have a ton of possession don’t give their opponent the ball very often, and thus can’t accumulate defensive stats.

What do you do when you know the basic rate stats are meaningless? You adjust them. Hopefully in a way that isn’t completely terrible, but we’ll wait and see on that.

But what about at the player level? Shouldn’t a guy who plays on a team that always has the ball (think of Sergio Busquets) get more credit than a guy who plays defense for a team that never has the ball (think anyone who plays for Tony Pulis)?

Of course! Math and logic both say this is correct.

So, those of us who are working on the player stats side of things need to adjust the numbers to compensate for this. As long as you have the data, this shouldn’t be that hard, right?

*sound of crickets chirping*

The initial adjustment method I used had some issues with extreme values. Barcelona’s int + tack went from 36.8 up to 97 or so. When you distribute that at the player level, what you end up with is Dani Alves making 13.4 tackles and 10.6 interceptions. I like Dani Alves a lot, but... that’s a tough pill to swallow. This is especially true once I started thinking about changing the radars to account for this. One of the things I like about the radars themselves is that they are a snapshot of reality. These are actual numbers produced by a player in a season. By not using abstraction (like percentages), you keep them approachable to your average fan.

24 interceptions and tackles a match for Dani Alves is no longer approachable.

Thankfully, I work with really smart people like Marek Kwiatkoski, and he helped me find a better fit for the adjustment that remained at least somewhat attached to reality and maintained the solid correlation between goals conceded/shots conceded and possession adjustments. If you want to know more, skip to the methodology section at the bottom.

So What Now?

The point of all this was to see what we end up with at the player level if we adjust for a team’s possession. Tackles and interceptions as they currently exist are mostly just noise because they don’t account for opportunity – can we do better? Whatever we end up with, do you see the best players on high possession teams rise to the top of the rankings? If so, we’re probably on the right track in developing defensive metrics that also correlate to things that help clubs win matches.

Well, let’s see what happens to our tackle and interception rankings when we adjust them at the player level. We are working on English Premier League data for 2014. I’ve listed the rank of top 15 tacklers here by the sigmoid adjustment, then the simple adjustment (see the methodology section), and finally by their base numbers.




As you can see, the two forms of adjustment produce fairly similar ranks to each other, but can produce dramatically different rankings than the base numbers. And this is how the interception rankings change with the adjustments.




And finally the combined defensive output of interceptions plus tackles.




And if you want straight player names next to each other, the column on the left here is the sig adjusted rankings, the column on the right is the base rankings. Notice all the guys who play for top teams on the left hand side? That’s pretty cool.




Additionally, here’s a look at how much the numbers change from the base vs the sigmoid adjustment for the top 15 adjusted Int + Tackle guys in the league.




Those numbers are no longer “real” numbers. You could think of them as tackle and interception points or "defensive things" instead, if that makes it easier to swallow. They are, however, much more closely related to outputs that help teams win matches, which is exactly what we were looking for at the start.

This could probably be another 5000 words long looking at different leagues and such (people will obviously want to look at Barcelona stats with La Liga as well as what Pep is doing at Bayern Munich – we’ll get to it), but for now I’m going to leave it here.

Any surprises listed there? Problems you see? Complaints? This is totally new research, so I expect there will be additional fixes/adjustments/hiccups before we get to a final product. If you’re really interested, check out the methodology discussion below.



@mixedknuts on Twitter

Methodology Info - In Steps Marek

First of all, the sample used at the team level was 4 full seasons from 2009-2013 across Ligue 1, Bundesliga, Serie A, EPL, and La Liga. When I did my initial research on this a few months ago, I was looking at the defensive rate metrics at the team level compared to shots conceded, goals allowed, goal difference, etc.

I was talking about this to my sometimes partner in crime Marek Kwiatkowski, and he suggested using a slightly different approach for the player info than I was initially using. Instead of taking the full difference in possession, Marek’s first new approach (which I called the “simple approach” above) was to attach everything to a base of 50, which is the game by game mean. So a 66.7% possession rate would go from 2:1 to 1.33:1.

I liked this because it delivers numbers more attached to reality, especially for extreme possession numbers. However, at the team level, it lowered the r-squared from shots conceded to .24 from .4. This is more correct than just using tackles and interceptions. It is possibly less correct than the full base team adjustment I was working with initially.

The next step Marek suggested was testing out a sigmoid function, which I had never heard of.  Swapping this in pegged the r-squared for both shots conceded and goal difference at .39, meaning we’re explaining about 39% of the variation in both of those outputs with adjustments to two simple defensive rate stats. When I first looked at this I was a bit disappointed, but given how complex football is, I actually think that’s pretty good.

Here’s the equation for those who are interested. The bit at the end is the possession adjustment.

Tackles * 2/(1 + e^(-0.1*(x-50))), x/50 for x in [0,100]

As noted above, the basic adjustment and the sigmoid version produce very similar ranks. It’s just that the sigmoid one gets more extreme in adjustment as you get further away from 50% possession.

All of this assumes that I have also done the basic queries correctly in pulling game by game possession stats for each game the players were involved in. I don’t have that info on a minute-by-minute basis right now, so it’s not game-state adjusted or anything.

It also assumes that adjusting defensive stats by possession to increase correlations makes sense and doesn’t simply fall on its face from a methodology standpoint. The logic behind it makes sense to me, but I’m just some guy who works in gambling, not a Ph. D. in stats or math.

Again, we’re probably imperfect at a number of levels.

But it’s a start.

Mexico mauls, Croatia croaks, Uruguay needs a win (Nate Silver, Day 13)

day 13
day 13
Heading into yesterday we knew that the Netherlands and Chile were both going to advance, the question was who was going to win the group. We also knew Brazil was likely to win group A, while Croatia needed to beat Mexico. Spain faced Australia in what was effectively an international friendly, while Cameroon were fed to the hosts. Australia played hard, but the Socceroos were clearly outclassed and even a somewhat lackadaisical Spanish side was able to win convincingly three-nil. Chile and the Netherlands played for the right to avoid the hosts. Chile needed a win, while the Netherlands only needed a draw. That meant when the game was tied nil-nil late, Chile needed to be the aggressor. Unfortunately for them, they wore down and a Leroy Fer header put them behind in the 77th minute. Shortly after a Memphis Depay tap in ended all the drama. At that point Chile expected to face Brazil, while the Netherlands were awaiting the Mexico-Croatia winner. That might not have worked out for the Netherlands, but Neymar scored twice in the first half, while Fred scored early in the second. It turns out, that would have been enough to win the group for Brazil (ahead of Mexico on tiebreaks), but a late goal from Fernandinho sealed it. However, the real story is just how badly Hulk has continued to play. Brazil has not played well this World Cup and I'd be concerned about whether or not they have a higher gear for the elimination rounds. Chile might not have the talent to take advantage, but Colombia might in the quarterfinals (Or Italy, or Uruguay). Miguel-Herrera Mexico just needed a draw to advance vs. Croatia, and played like it for most of the game. They let Croatia have most of the possession, but denied them good scoring opportunities. Mexico was able to break through in the 72nd minute, and after that Croatia collapsed, giving up goals in the 75th and 82nd minutes. That was the point when Mexico was a goal away from winning the group. It didn't work out as the wall of Ochoa was finally breached in the 87th minute. The result sets up an interesting matchup between the Netherlands and Mexico. If styles make fights, this should be a good one as the Netherlands were built to attack, while Mexico was built to defend. However, that game isn't until Sunday, and we have four more games today. As the image at the start of this piece shows, the good ship Full-Kelly hit some rough waves yesterday, dropping to $446,092. Today's games: Game 37: Uruguay, 9.95%: Risking $35,607 to win $66,940 Game 38: Costa Rica +0.5 goals, 19.78%: Risking $79,457 to win $74,960 Game 39: Columbia, 28.54%: Risking $155,059 to win $174,502 (Updated) Game 40: Ivory Coast +0.5 goals, 9.8%: Risking $41,178 to win $14,199 (Updated) Nate and Uruguay see eye-to-eye: just win baby. Italy advances with a win or a draw, while Uruguay needs a victory. Costa Rica wins the group with a win or a draw, and might win on tiebreaks with a loss. England are going home no matter what happens. Columbia are guaranteed to advance, and clinch the group with a win or a draw. Japan needs a win, plus some tiebreak luck to advance. Greece is probably in with a win, although they could lose out to Japan on tiebreaks. Ivory Coast just needs a draw to advance, and has a very slim shot of winning the group with a win. A bad day could leave Nate at $197,298. A sweep would leave him with at least $725,484, and possibly more depending on the where the lines are on the afternoon games when the first games finish. Very exciting stuff. Uruguay and Greece playing for wins, Nate playing for cash. Enjoy day 13 of the World Cup! 🙂  

The Top Young Prospects Across Europe - A YAPSS Data Dump

Raheem Sterling In the two weeks since I started writing about this project, I have made a new discovery: YAPSS plus a number of leagues is big. I am left with a dilemma. I could spend the rest of the summer just writing about attacking players you might not know, and while that might be interesting for some, it would preclude me from doing all sorts of other things, like applying the model to other positions, developing new defensive metrics, fullback radars, or digging into all the locational data we recently received. All of the scouting you see below was done with Opta data. So what I am going to do instead is to give all of you the players the model likes. That way the information is out there, and if you guys want to write about them, just tell me when it’s done on Twitter, and I will RT it for you. The names below are not in any particular order, but all of them are 23 years or younger, and all of them trigger the scouting model as potential future stars. I have done very little league adjusting here, mostly because I don’t have enough historic data in “lesser” leagues to build new baselines, so keep that in mind when you are evaluating the predictions. Also remember, this is the model for ATTACKING players. The one for midfielders and fullbacks is something else entirely. EPL YAPSS Prospects EPL_YAPSS_2014 Bundesliga YAPSS Prospects Bundesliga_YAPSS_2014 For Bundesliga in particular, there are a number of guys that are just outside the model triggers that I would look at, mostly because they are quite young vs the age cutoff. These include Julian Brandt, Leon Goretzka, and Jose Campana. There's also the weird factor that a guy who plays fullback (Ricardo Rodriguez) is listed in the attacking player model. That has happened one other time in the five-year data set that I have. That player's name was Gareth Bale. La Liga YAPSS Prospects LaLiga_YAPSS_2014 Serie A YAPSS Prospects SerieA_YAPSS_2014 Ligue 1 YAPSS Prospects Ligue1_YAPSS_2014 Eredivisie YAPSS Prospects Erediv_YAPSS_2014 Does the Eredivisie have more good young players, or do they simply give more young players playing time? An excellent question, and one I don't have nearly enough data to answer right now. My instinct is that it's a bit of both. English Championship YAPSS Prospects Eng Champ_YAPSS_2014 Jordan Ibe is right on the edge of triggering the model. Given his extremely young age (only 18), he’s another guy I would add into the pot. Championship teams should be scrambling in hopes of getting him on loan for a half or full season from Liverpool. Opta_200px

Soccer Can Be Cruel (Nate Silver, Day 12)

A draw wouldn't have been exceptional for the Russians, but they could have lived with taking two points into their final game against Algeria. Nate Silver definitely could have lived with it, but alas, Divock Origi scored in the 88th minute. The goal was deserved, as Belgium had been dominating the play from the 80th minute onward. Now Belgium has advanced, while Russia doesn't entirely control their own destiny. True, a win vs. Algeria almost certainly sees them through, but it is possible for South Korea to advance with a win if they demolish Belgium and have better tiebreaks than Russia.

A win vs. Algeria won't necessarily be easy. The Desert Foxes played Belgium very tough, and thoroughly demolished South Korea. The 4-2 scoreline doesn't tell the whole story, as Les Fennecs mauled South Korea early, taking a three-nil lead, as well as the first 10 shots of the game. That South Korea was able to rally a bit is commendable, but they got their butts kicked and are now in dire straights. Algeria is very likely in with a draw, and has an outside shot of winning the group outright with a win. If they can maintain their level of play from the first two games against Russia, I like their chances of advancing to the next round.

Speaking of advancing to the next round, the Americans could almost taste it. They held a 2-1 lead in the 95th minute. Portugal had time for one more attack. Unfortunately for the Americans, it was a brilliant ball, leading to a perfect header to tie the game. It had been a back-and-forth game, with Portugal carrying the play early. Their first goal was the result of a defensive misplay, but they had been on the attack from the opening kickoff. After that goal the Americans regrouped and carried most of the play. They equalized on an rocket from Jermaine Jones to which Beto never reacted. The Yanks then took a lead on a deflection off of Clint Dempsey's abs. That lead held until Varela's header. Sigh. In a perfect world team USA would shake hands with the Germans, but that seems unlikely. Sure, if we can play them to a draw through 75 minutes they might ease up, but right now we have to assume America needs to earn at least a point. Of course, a Portugal-Ghana draw would eliminate both teams, and even a Portugal win puts America in pretty good shape to advance. The truth is, a draw was an excellent result for team USA, it just tastes like ash after leading in the 95th minute.

That result was good enough to give Nate Silver another positive day, finishing with $493,771. He was in real trouble, having lost the first two games, and trailing at halftime in the third. After that sweat, he now gets to deal with multiple event Kelly-staking, as we get our first taste of simultaneous futbol,  starting with group B:

Game 33: Australia +0.5 goals, 1.51%: Risking $6,845 to win $14,563

Game 34: Chile, 8.2%: Risking $39,878 to win $74,173 Followed by group A:

Game 35: Brazil, 31.18%: Risking $129,019 to win $21,933

Game 36: Croatia -0 goals (same as Draw no bet), 7.44%: Risking 22,890 to win 21,594

I'll update the stakes on group B if Australia or Chile manage to come through for Nate.