We need a finisher

[umpiredice]

41 minutes ago, Dub Narcotic said:

Dwayne de Rosario: .27 goals/appearance

Dale Mitchell: .35 goals/appearance

Cyle Larin: .26 goals/appearance

Larin already has a very reasonable strike rate and is not even near his peak years yet.

At first I didn't see the decimal points and I was like WHAT???

[gardencity]

I can't help but note the irony of this thread: We need a finisher. Every 6 months for the past 20 years someone has lamented this glaring weakness. We can't score, our defense is lacking and disorganized at the most inopportune times, and the brass are in a constant state of damage control and reorganization. We are a G 7 country whose Men's National Team program looks like something out of the Third World....without the corruption.

[hodgkiss]

i find this topic very funny. Here's why:

Pretty much everyone identifies with the fact that we have trouble scoring. It's been an issue in the past and continues to be. De Vos, who I appreciate has a voice in Canadian soccer has made it quite clear that Canada need to concentrate on tactics and technical ability over scoring which is why the grass roots teams are not supposed to keep score. I have had a few battles over Twitter with him and others about this issue and now that he is the Director of Development, I fear this is going to get worse. I feel that we will field teams that are all about flash that can't score goals. 

Our players need to be hungry, learn that participant ribbons for technical ability don't win World Cups and a desire to win at all costs is what it takes.

Canada, stop being so nice. Grow some balls, score goals and win the damn game!

[Free kick]

5 hours ago, dsqpr said:

For comparison, here are the numbers from the top scoring "relatively modern era" England strikers and midfielders:

Rooney 53/116 = 0.46
Lineker 48/80 = 0.60
Owen 40/89 = 0.45
Shearer 30/63 = 0.48
Lampard 29/106 = 0.27 (M)
Platt 27/62 = 0.44 (M)
Crouch 22/42 = 0.52
Keegan 21/63 = 0.33
Defoe 19/55 = 0.35
Welbeck 14/34 = 0.41
Chivers 13/24 = 0.54

Source: http://www.englandfootballonline.com/teamgoals/Goals_10-49.html

As far as I am concerned, anything less than 0.33 is simply not good enough for an out and out striker. 0.26 is the kind of number I'd expect from an attacking midfielder like Frank Lampard in the list above.

But those stats dont account for service and quality of service that forwards receive.  More specifically, scoring chances created.  I know we often lament the missed chances in games.   But  If we actually look closely at all of our matches over the long stretch,  its not often that we out chance our opposition when the opposition is not one of the Concacaf minnows.  

The quality passes and crosses are seldom there in the final third.   Looking back over the last four to five cycles,  we have often had prominent fwds who have had scoring success at the club level.  There has been Radzinski, Dero, Bunbury, Larin, and even Peschisolido. Many of these guys were stars at their clubs.   But have we ever had a number 10 in midfield who can pull the strings? And deliver the the decisive quality balls?  Most of our prominent players in MF have really been defensive mids who were never counted on for that at the club level.    That includes Deguz and Hutch.  

We have been able to sustain possession if going left to right and back. But that quality cross or through ball? Thats the big question.  

 

[Guest]

17 hours ago, dsqpr said:

A few interesting numbers from our much more successful women's team (which, so far as I can recall, never had that wonderful number 10 setting up goals galore for everybody either):

Sinclair 165/250 = 0.66
Hooper 71/128 = 0.55
Lang 34/92 = 0.37
Tancredi 22/107 = 0.21
Beckie 12/22 = 0.55

Particularly interesting to compare numbers for our two young talents: Larin and Beckie.

And somewhat randomly, our former host of Soccer Saturday, Graham Leggat, RIP.

Leggat (Scotland) 8/18 = 0.44
 

Source for all the above: wikipedia

 

Its not really a great idea to compare a Men's and Women's team.

[Obinna]

17 hours ago, dsqpr said:

A few interesting numbers from our much more successful women's team (which, so far as I can recall, never had that wonderful number 10 setting up goals galore for everybody either):

Sinclair 165/250 = 0.66
Hooper 71/128 = 0.55
Lang 34/92 = 0.37
Tancredi 22/107 = 0.21
Beckie 12/22 = 0.55

Particularly interesting to compare numbers for our two young talents: Larin and Beckie.

And somewhat randomly, our former host of Soccer Saturday, Graham Leggat, RIP.

Leggat (Scotland) 8/18 = 0.44
 

Source for all the above: wikipedia

What are you attempting to draw from this? 

[Dub Narcotic]

On 9/8/2016 at 10:16 PM, dsqpr said:

There is obviously an element of truth in what you say but you are basically making excuses for our poor forwards. Just taking Larin for example, he has had lots of chances: including a memorable ballooned shot over the top of an open goal, a one-on-one with the El Salvador keeper the other night which he scuffed wide, and again against ES, De Jong's cross which he side footed over the bar from 3 yards out; and he has wasted them. If he had taken even half of the great chances he has had he would be at 0.5. It is a good thing we play against minnows regularly because that is where our forwards pad their averages UP to 0.25! Not good enough. And it isn't just down to lack of service.

Do you not understand the expected goals statistic? There is an objective way to measure how well chances were taken and it complete refutes your argument. Our players actually scored more than expected relative to the chances they had.

[Guest]

2 minutes ago, dsqpr said:

My point is that good strikers average more than 0.33 goals per match. That's it!

Larin does average more than 0.33 gpm for club when he actually has support.

[jonovision]

On 9/7/2016 at 7:38 PM, Dub Narcotic said:

If someone did an xG map of that game I bet it would be pretty close to three. As mentioned, two of the goals should not have counted and there weren't that many clear-cut opportunities.

Wrong. Edgar was offside, but on Larin's goal, despite him being in an offside position, since the ball was merely played off Hoilett by the defender, rather than actively played forward by Hoilett, it's not offside. Source: a tweet during the game by someone who knows the laws much better than I do.

[kohanz]

2 hours ago, dsqpr said:

My point is that good strikers average more than 0.33 goals per match. That's it!

This is the danger of statistics in the hands of people who don't usually deal with numbers. You've chosen an arbitrary cutoff of 0.33 gpa, fine. The sample size for Larin (5G 19A) at the MNT level is just way too small to be drawing any conclusions from in that respect. If he put away just ONE more of his chances, he'd be at almost 0.32. If he put away 2 more chances, he'd be well over the mark at 0.37. It's ludicrous that one missed goal would separate him from being "good enough" to be not "good enough", especially at this stage in his career. Your intuition may be right about Larin, but basing that conclusion on "stats" is way too early. Look at his MLS numbers, with a much larger sample size, he's at 0.57. I think most people would agree he's underperforming at the MNT level, but I would expect him to "regress to the mean", in a good way, rather than write him off at such a young age.

[Dub Narcotic]

Your point is taken but we've seen that no one scores much with Canada. This team does not generate many chances unless playing absolute minnows.

[kohanz]

On September 10, 2016 at 4:00 PM, dsqpr said:

I agree that the sample size is relatively small, but I still think it is big enough to be meaningful.

I didn't want to get too technical with you, but the answer to "I still think it is big enough to be meaningful" is "no, it's not". There are statistical tests to determine what kind of samples sizes you need to make claims as to whether a comparison (e.g. is Larin's strike rate actually less than 33%) is meaningful ("significant" is the term used) and in this case it would certainly not be. Statistically speaking, Larin's true strike rate is unknown (and probably changing over time as he develops). The 6G 19A = 0.27 is just an estimate of that true strike rate based on what he has done so far. The question is, is Larin's true strike rate less than or greater than 0.33 (I will leave the fact that this is a somewhat arbitrary number alone for now)? The fact that a single goal could change the answer from a yes to a maybe/no means that we cannot say, with any great degree of confidence, that Larin's true strike rate is actually below 0.33. This is a fairly simplified explanation, but I don't want to get super technical.

Anyway, the point is, your intuition or "feel" about Larin may well be true, but no, the numbers don't back you up beyond coincidence. They are mostly "inconclusive" at this point. If Larin had scored 1 or 2 goals in 19 apps, then we could have some confidence in saying that he would score at a rate of less than 0.33 going forward. If Larin scored a brace in his next game, would it completely change your opinion of him? I would think not, but yet the "numbers" wouldn't back you up at that point. This just goes to show that, at this point, his numbers at the international level are not enough to judge him on alone.

Sadly, I'm not even a big Larin backer. I merely have some experience in sports analytics and this kind of stuff riles up my inner nerd.

[rob.notenboom]

On September 10, 2016 at 2:00 PM, dsqpr said:

I agree that the sample size is relatively small, but I still think it is big enough to be meaningful. True, if he had scored a couple more goals he would be over 0.33, and if ifs and buts were candy and nuts we'd all have a merry Christmas.

My 0.33 mark is based on many years of watching football. Good strikers are typically between 0.33 (one goal every 3 matches) and 0.5 (one goal every 2 matches). Great ones are over 0.5. If you look at the numbers I've posted above you will see they support this opinion.

My opinion of Larin is based on watching him play for Canada (I don't watch MLS). His stats currently support what I have seen. I hope he gets a hat trick in his next match and goes on to prove me completely wrong. But players who start out as 0.25 strikers (one goal every 4 matches) typically stay there.

Edit: I find it quite touching that so many people are stubbornly sticking up for Larin in spite of the evidence that points quite clearly to mediocrity (at the International level). I know you all have high hopes that the future will be better than the past. May you all be proven 100% correct.

I've actually thought a bit about Larin while watching Harry Kane both for club and country. In the last 3 premiership seasons, Kane's strike rate was above 0.5 but with the England squad it is below 0.3. I think there is an argument to be made in both Kane's and Larin's cases that this is somewhat reflective of the number of chances that their respective teams create and how they use them as strikers. With the defensive style that Canada has been playing, Larin often ends up at the top on an island and against better teams we create very few chances (this is not to say he hasn't had gilt-edged opportunities that he's failed to take). With Orlando this is rarely the case, but also with Orlando, Larin seems to be far less wasteful. In fact, there are a few games that come to mind where he was clutch.

I'm not really disagreeing with you though. Larin has been frustrating at times in international play and I suspect that there's a combination of style of play/tactics, cohesiveness of the team, and psychology that helps explain why he seems to be a poorer finisher statistically internationally and much better in MLS. You should watch him in MLS. I think you'd be impressed with how quick and precise his finishing sometimes is. He seems to be much more in the groove in MLS, but I think Orlando also use him better.

 

**Admittedly in the Kane example above, he has not played enough games for his international strike rate to be statistically significant either. I just thought it might be an instructive comparison.

[kohanz]

Well at least we don't have Larin taking corners!

[kohanz]

9 hours ago, dsqpr said:

When you are sampling to determine, for example, what percentage of people will vote Tory at the next election, you are sampling a population that exists. There are millions of people who will vote and if you asked all of them you would get the true mean. Since this is not practical you instead ask a small number of them and then try to estimate the true mean. You do this by using statistical analysis to calculate, for your sample size and the assumed population distribution (e.g. "normal"), a 95% confidence interval which is "plus or minus 'x' from your sample mean" (I.e. you are 95% sure that your sample mean deviates no more than "x" from the true mean -- which is plus or minus 1.96 standard deviations from the true mean in a normal distribution). Hence you get poll results such as "accurate to within plus or minus 2%, 19 times out of 20". A result (sample) is considered to be "not statistically significant" when there is less than a 95% chance that it differs from the true mean.

But none of this applies to calculating a player's scoring percentage because you can ALWAYS calculate a player's TRUE strike rate EXACTLY! In Larin's case it is 5/19 = 0.26 for Canada. The future is unknown. There is no existing population of thousands of matches played for Canada that we are sampling from. At the present time there are exactly 19 matches and we have included ALL OF THEM! He may never play for Canada again. That may be it for him. Or he may play another hundred times, and each time he plays we will be able to calculate his strike rate PRECISELY! No need for a 95% confidence interval! Of course we could choose to sample only 5 of his matches and maybe by chance he would have scored 3 goals. In this case we would use 0.6 as an estimate of his true strike rate, which we know is 0.26, and we could calculate our 95% confidence interval based on our sample size of 5.

To summarize, it is quite wrong and shows a complete lack of understanding of statistics to say that the difference between Larin's "true" strike rate for Canada and 0.33 is not statistically significant, or that his "true" strike rate for Canada is not known. It IS MOST CERTAINLY KNOWN - IT IS 5 GOALS IN 19 MATCHES = 0.26. It may change in the future, but today that is what it is - PRECISELY!! 

You are basically saying that if Larin had 1 G in 2 A to start out his national team career, you would be instantly sold on him. It would be "statistically significant" because this is his "true strike rate". Who cares how many games he has played, right? To you, he would be a 0.5 strike rate guy, end of story. The fact that your opinion on him could literally change based on the next game he plays for us (he bags a brace), shows how tentative your grasp of the numbers is.

What you said doesn't make any sense. Barring a disaster, Larin *will* play more games for Canada. Probably many more. Yet you are saying that the question of what his long-term strike rate will be has already been solved. To you, it will be exactly 0.27 for the rest of his career, or at the very least, never ever come above 0.33. Those kind of claims are basically gambling, they are not based on numbers. Again, I am not saying you are wrong. In the long run, Larin's strike rate may indeed fall below 0.33. Based on current numbers, it is more likely than not, but not by a large margin. The very next game could realistically change that assessment. Your opinion of Larin could be right, but the numbers don't back you up with any conviction in the way that you're trying to use them.

Answer this question: Larin scores 2 goals in his next game, bringing him to 0.35, which is greater than 0.33 - suddenly you're sold on him?!

[rob.notenboom]

11 hours ago, dsqpr said:

I agree with you that Kane has not been very good for England Rob. If he wants to be remembered as a great England striker, he is going to need to up his scoring percentage.

I'll address your "statistically significant" comment in the following more technical post.

Just to add a bit to this, I think something you said above is very important but got lost in all of this - international level. Regardless of whether or not one thinks that the level of these international matches is higher than MLS (I'm not eager to spark that debate), it is one more important factor in a whole series that make the contexts of these situations very different, and clearly to Larin, context matters. 

I realize I'm kind of restating, but I truly feel that the context has an effect on Larin's psychology/confidence. In the Mexico game, he was in alone on the keeper on the left, and instead of taking the shot with his left he elected to try an awkward toe-poke with his right and shot well wide. Then this summer with Orlando he had a similar opportunity, took the shot with his left and scored it in the far side. Fast forward to the El Salvador game and he took a shot that was almost a carbon copy of the Mexico one. 

I may be making too much of it, but to me this speaks to Larin's confidence level in the two different contexts. 

[kohanz]

@dsqpr, there is no sudden change from my side. My point from the beginning was that your opinion on Larin was not bolstered by referencing his strike rate being lower than the arbitrary marker of 0.33. Since 2 goals in his next game, not likely, but not out of the realm of possibility, would have the numbers go from "backing you up" to "contradicting you" ( I put those in quotes because IMHO the numbers are at this point too inconclusive to do either). Anyway I think this has run its course.

[Dub Narcotic]

Isn't this a regression problem rather than a sampling problem? I will say it's frustrating that we can only really find goals/appearance rather than goals per 90/94.

[Bertuzzi44]

Yes we need goals but Canada has almost always tried to win games with defence. That is one reason why historically Canadian strikers have low strikes rates. That and quality, but as others have mentioned, Radzinski, Bunbury, DeRo, Larin, etc, aren't chop liver.

When is the last time Canada consistently used 2 strikers? And I don't mean calling up 2 strikers for a 23-man roster.....

[kungfucious]

1.  thanks for starting this thread!  

2. That and quality, but as others have mentioned, Radzinski, Bunbury, DeRo, Larin, etc, aren't chop liver.. 

pls pls pls pls dont ever mention larin in the same comparison with any of that group again.  

3. nothing official, but i think about 2/3 of u are larin fans. im sorry, but i have to be more critical. at 21, either u have the touch or u dont. this qualifying round, mr larin proved he absolutely does not and is the biggest guilty party for the team not progressing. im sorry. i want to believe, but was chicharito, perralta, donovan, or dempsey EVER this wasteful?  

4. for those of u pointing out that he doesnt get good service, he gets the ball 20yds from goal, the chances he gets are near impossible... do u think inzaghi, raul, torres, solskjaer, owen, et'o, trezguet, shevchenko, larsson, van nistelrooy, would screw up 90% of the half chances this guy does?  perhaps this is the answer to my own question: a poacher only needs a half chance and larin is definitely not one. that .26 figure u stats guys are always repeating on larin: it is more accurate the percentage im going to eat my words.

[jpg75]

If anyone has lots of time on their hands how about charting Larin's goal rate from Game 1 to 19 and see how it evolved, to see if he has consistenly scored at a .26 (or approx.) rate?

[kohanz]

Just to throw another high-variance stat into the mix, if you only count meaningful games (e.g. not friendlies), his strike rate improves slightly (4G 13A) to 0.31. 

That's over 919 minutes, about 70mpg. You can convert it to 0.39G per 90 minutes, which would be good for an attacking midfielder, but not for a striker (see Ozil, Mata, Gotze: http://statsbomb.com/2013/08/an-introduction-to-the-per-90-metric/). 

In his young MLS career, he scores 0.69G every 90 minutes for a sample size over 4x greater (minutes-wise), which is excellent for a striker.

I don't think anyone denies he needs to up his production for the CanMNT. IMHO, it's too early to make a ruling on his international career though.

[Bertuzzi44]

15 hours ago, kungfucious said:

2. That and quality, but as others have mentioned, Radzinski, Bunbury, DeRo, Larin, etc, aren't chop liver.. 

pls pls pls pls dont ever mention larin in the same comparison with any of that group again.  

I think when it's all said and done, he will definitely be among those names. Going by his club performances, he already is...

[GimliJetsMan]

Surprised to see the "Advanced" stat numbers. Shows that Canada had it's traditionally awful offense. I really started this thread as a way to lament some missed opportunities we've seen at the national level.

Stats don't really change my opinion about finishing, given the quality of chances missed. We've seen Canadians waste too many A+ chances for my taste, not just in this cycle, but over years and years.

[dyslexic nam]

11 hours ago, Bertuzzi44 said:

I think when it's all said and done, he will definitely be among those names. Going by his club performances, he already is...

God I hope this is true.  It is far too early to write Larin off in the red and white - or even to try and predict a relatively low ceiling for him.  Having said that, there has definitely been a wasteful aspect of his play for us that I hope we can solve.  We have a guy who obviously has a nose for goal.  Given our lack of effective alternatives up front, our next coach would be well advised to implement a system where we can exploit Larin's ability.  Bunkering down and defending as a team can be a good strategy, but it can't be the only strategy if one ever expects to accumulate meaningful wins.