- Both teams pressed high looking to turn the opposition over in midfield and break.
- Toronto had vastly more opportunities converted into shots but struggled to convert those shots into shots on target.
- We expected the unique 3-5-2 from Toronto. Atlanta’s tactics were interesting, both in the 1st half and after an adjustment in the 2nd.
(as always, the green maps from MLSSoccer.com, the charcoal ones from whoscored.com)
Both teams pressed
But holy hell did Toronto press early. And for the second time in two weeks, I was surprised to see a very strong team press the visiting Atlanta side. Yes, Friday’s post was dead wrong again. Perhaps, I should start to expect this. It seems as though the scouting report on ATLUTD around the league is “stop their playing out of the back and create some cheap chances of your own – taste of their own medicine.” Atlanta also pressed, especially early, but as you can see below (left, Toronto tackles, interceptions, recoveries; right, Atlanta; bottom, possessions lost), the Reds had the better of it.
You can see the balance of play leaning in Toronto’s favor in terms of defensive aggression. For instance I count 11 turnovers in Atlanta’s own half, to Toronto’s 6. Here are the unsuccessful passes in each team’s half (Toronto left, Atlanta right):
That’s Atlanta dropping 29 unsuccessful passes in their own half (!) compared to Toronto’s 14 (+ a few dribbles). Wow, yea Atlanta got pressed. And I’d venture to guess they should expect to see more of that on this road trip.
For the 2nd week in a row, I must mention that purely from a practical perspective, I believe at this point in the season, Atlanta can optimize the volume and quality of chances created/conceded by playing more direct — playing long, and if the pass doesn’t come off, applying pressure immediately. This is because Atlanta has the perfect skill triangle in the front four to create chances from more direct pressing and counter play: 1) aggressive and energetic pressers 2) pace 3) eye for through balls. The team did in fact play direct more often in this game, partly by necessity, but there were definitely spells where Atlanta stubbornly played the ball out of the back to frequently bad outcomes. That said, the team continues to pick up points on the road in difficult situations, and this surely builds confidence in Martino’s vision, so to be clear, I’m not second guessing the long term value here (and I love watching our team pass the ball) — just pointing out that the more direct and risk-off path towards points in the short term might be to hit the ball long more and immediately press the second ball – and like I said, to be fair, ATL did this a few times and created goals doing it. Don’t run a grammar check on that “sentence.”
Chances (left Toronto, right Atlanta)
Toronto out-shot Atlanta 20-6 (9-5 on target). The expected goals models all seemed to agree that Toronto had the upper hand (2.28-1.1, 2.25-1.21, 1.98-1.21) and looking at the xG race graphic from @11tegen11’s model, Toronto and Atlanta each had 2 big chances. Toronto converted both of theirs (both of which were very nice) and Atlanta converted 1 of 2 (the long ball over the top to Tito in the 2nd half). The other big chance for Atlanta was Villalba’s shot from the penalty spot which was smothered at point blank on the rebound off Almiron’s solid effort in the 48th minute.
Also, in case it wasn’t clear enough, here’s the amount of “emergency defending” (clearances + blocks) for each team. Toronto left, Atlanta right:
Toronto had many, many more “half-chances” than Atlanta, and they squandered most of them shooting off target. Some were on frame and Atlanta’s Kann did well to save, but others sailed over the cross bar as I winced with fear. In this sense, it is fair to say Atlanta were “lucky” to come away with a draw since they largely cannot control whether the opposition’s aim is true or not. But importantly, I would not say Toronto were unlucky. If you want to take advantage of bossing your opponents for large portions of the game and creating loads of chances, you must put the shots on frame or prepare to draw matches at home. Toronto had 5 unblocked shots in the penalty area which did not test the keeper. This brings us to goal keeping – another element which is traditionally thrown into the “luck” discussion. Kann did wonderfully with the shots he faced – and shouldn’t be faulted with either Toronto goal. Conversely, Bono made an amazing save on a swerving Almiron drive late, but perhaps he would want that first Atlanta goal back with Villalba through but shooting from a wide angle (personally I thought it was a decent enough chance that I stood up as he sprinted into the box, but I suppose the historical conversion stats from that spot disagree with me). Through 5 games, Kann rates second best in the league at goals allowed minus expected goals. He’s letting in 0.57 goals per game less than the models would predict given the sort of shots the opposition are taking. This is by no means a perfect way to measure goal keepers, but it’s something.
In the first half, Atlanta’s formation was fascinating. The official @ATLUTD tweet showed a 4-3-3 with Villalba leading the line and Almiron left. However the FoxSportsSouth broadcast showed a diamond 4-4-2 in the pre-game with Asad and Tito as forwards and Almiron between the lines. To my eye, it was something in between, with Fox getting points for being the closest. Perhaps @ATLUTD was intentionally misleading its opponents with the tweet (sounds a little silly though). It’s difficult to differentiate between a diamond 4-4-2 and a false 9 (4-3-3), and I’m not sure we need to, really. I think the idea here was to maybe do some of the false-9ish stuff we discussed briefly last week with Almiron creating a real headache for Michael Bradley and letting Villalba and Asad attack the flanks of the back 3/5.
To my eye, Villalba played wide right (and furthest forward) occasionally taking on primary striker responsibilities while Asad played wide left (and also forward-ish) but drifting around a lot and more-than-infrequently swapping roles with Almiron, who played mostly central as a play-maker/counter-starter, just a bit deeper than Asad and Villalba. Again, it was very fluid as these three guys would shift around all over that upper band of 3, often depending on where their press took them. I’m not sure, but it seemed at times like part of this rotation was intentional on the part of the players to conserve energy. I even remember one occasion where Asad was wide left, Almiron central, and Tito moved up to the #9 spot with Mears bombing forward on the right to replace him. Matt Doyle’s description of a “lop-sided 4-3-1-2” is probably a good one. One consequence of this tactic was that the idea I’d described last week about “turning the back 3 into a back 5 by attacking with width” did not come to fruitition as the 3 centre backs were mostly left alone, seemingly sufficient to cover against Asad/Villalba and at times Almiron. So while, Atlanta wasn’t successful in the opening 45 in pushing Toronto’s wingbacks back into their own half, they were successful at exposing the centre backs (like the opening Villalba goal where Almiron threads the ball between two of them). I’m curious if Mavinga was included in the starting 11 specifically to match Villalba’s pace. That would be ironic in hindsight.
The second half gave us a fun tactical surprise. It’s hard to tell exactly when Jeff Larentowicz shifted back between Parkhurst and Pirez to form a 3-man backline, but it’s as early as the 49′ restart after Villalba’s goal to open the half and no later than 59′. Watching the broadcast in real time and having moved into third gear with a Redbreast Irish whiskey (Tropicalia and Glendalough being the first two), I didn’t fully notice the shift, but rewatching on DVR and looking at some of the player maps from the 2nd half, it is hard to ignore its impact. Mears and Garza were able to venture further forward and Parkhurst and Pirez were also able to shift out and up just a bit.
In general, Atlanta did a better job managing Altidore/Giovinco after the adjustment. Of the two, Altidore, who actually plays deeper than Giovinco for large stretches, seemed to disappear as the game wore on. One thing that didn’t change much was the general approach of Atlanta’s front 3. Almiron remained relatively free but predominately in the central “between the lines” role, and when Toronto had possession in Atlanta’s half, Almiron would drift free and upwards instead of actively defending so he could start counters at pace. Of course when Toronto had the ball in their own half, Almiron pressed them relentlessly.
Toronto’s press had slowed to a halt by the time Asad was sent off. Meanwhile, Atlanta was flying. It’s wild to watch the speed and aggression of Almiron and Asad and Villalba without the ball. Even late in games, they’re just so quick and aggressive. But the red…
One thing I wanted to mention because it struck me watching the game back again is that in the 55th minute, Asad appears to get decked with an elbow or forearm by Vazquez and Asad goes down just in the corner of the screen. Vazquz stands over him aggressively and is right in his face as if to say “get up, you diver” (or something worse). Shortly after, we hear the broadcast team mention that Gressel was having some words with Vazquez. I can’t tell from the video if it was a violent play that did connect with Asad, or if it was Asad trying to get Vazquez sent off. Either way, I can’t help but think that this played a part in Asad’s frustrated “swim move” later in the 75th. My two cents is that the red wasn’t a perfect call but it’s justified, if that makes sense. More importantly, I really enjoyed Tata’s tactics after the card and Atlanta’s spirit. They organized into a 5-3-1 sort of thing and kept countering TFC — great stuff from a tired but fighting Garza late on. And Kenwyne was near perfect when he came on.
Postscriptum: On Finishing, again
It has been mentioned repeatedly that ATL are running a bit hot in terms of conversion percentages and goals over expected goals. Atlanta’s 46% conversion rate is second only to Portland’s 52% (vs a league average 33%), suggesting we should expect some regression to the mean – and I think we should… I will say however that Atlanta is also running hot in terms of shot accuracy with 47% of their shots testing the keeper (league average is 36%). This is historically high as Kevin Minkus point out and I suppose it should regress some. But at the same time, I’m encouraged. If finishing truly is a crapshoot, because it’s one part “shooting is hard” and one part “you can’t control if the keeper makes a great play,” then Atlanta seems to be taking care of business on part 1 – either because its attackers are better at shooting (? they are DPs/TAM players I suppose), or 2 – because Atlanta are creating the kind of chances which make it easier for a shooter to be accurate (e.g. one-on-one with keeper), or 3 – yea, just luck I guess. Point is, if shot accuracy persists, we may see some high conversion persist as well. That might be analytics heresy, but happy to listen if anyone has the hard evidence against it.
Postscriptum: Player Performances
Toronto had a number of good individual performances. Jozy (left) and Giovinco (right) were a handful, 12 shots (5 on target) and 6 key passes between the two of them.
Cooper (left): 2 shots, 3 key passes, 92% pass accuracy, and 4 dribbles
Vazquez (right): 2 shots on target, 2 key passes, and generally looked dangerous.
For Atlanta, Almiron (left) was critical and somehow managed 86% pass accuracy despite being the attack starter. Gressel (middle) had his first poor-looking game on the maps (54% passing from midfield), although I enjoyed his committed work-rate especially after Asad was sent off. And Asad (right) was sort of everywhere as usual, but was frustrated all game in terms of final output, and I wonder if this was a factor in the red.
Happy to hear your thoughts on the game, and/or feedback on the write-ups. Starting to get a few of these under my belt. Let me know what you hate before it becomes “the thing I always do.”