Hang tight while we fetch the video data and transcripts. This only takes a moment.
Connecting to YouTube player…
Fetching transcript data…
We’ll display the transcript, summary, and all view options as soon as everything loads.
Next steps
Loading transcript tools…
Data Analytics in Soccer | Chalk Talk | 09.03.17 | New York City FC | YouTubeToText
YouTube Transcript: Data Analytics in Soccer | Chalk Talk | 09.03.17
Skip watching entire videos - get the full transcript, search for keywords, and copy with one click.
Share:
Video Transcript
and so let's get started because this is
what I want to talk to you about how we
use data analytics in soccer because
it's been a big boom over the last
stand' four or five years with data
analysis performance now is this
statistical analysis and specifically in
soccer I'm sure you've seen social media
Twitter everybody these logs are lots of
bloggers talking about how we use
statistics I will use data analytics but
I want to talk to you about how we use
it here at NYC FC and and basically I'm
gonna take you just talk to you a little
bit about the history of statistics and
data and where we're at now because
there's two types that we've used
physical data and technical data
I'm predominately predominantly on the
technical side but physical data is
where as a sport we feel like we've
mastered so a physical data I'm sure
you've heard of GPS so how far a play is
run what intensity is a player runner
how long does it play around for how
many accelerations how many
decelerations how many times that the
jump landed hit the floor
all this kind of physical data we feel
like we've mastered and that's been
through the use of GPS heart rate data
and I'll give you an example of how we
use it for instance if a player got
injured and had a muscular injury a
hamstring injury we could go back over
the period of time just before their
injury and we can look what did that
player actually do how long did he run
for what kind of physical law did he put
on his body that may have helped him
have that injury that injury occur we
can look at that time frame maybe six
seven eight weeks before we can see what
that player did and during our training
phases we can go back and we can see in
track okay we might be at a level where
this player could be at risk again so we
could manage display specific training
Lord during this week two week three
week period so from a physical
perspective we never learn we never hear
a Decatur chance of injury but we can
try and minimalize it and help this
player progress through a Pacific
training phase that he may have may come
have a Rico an injury try and prevent
that from the technical side of things
we have then
we're not would get into a stage where
we have access to lots of different data
has everybody heard of opto Sports yeah
opto no so for those who don't know opt
RA and
a statistical company who basically
watch our games and they called our
games just from a statistical
perspective they take information on
every single event on the field so a
pass a shot are throwing a tackle they
called everything and they produced the
data quantities quantitative
quantitative analysis and if everything
has happened on the field so that is the
type of data that I use and what we use
to produce a statistical analysis and
our data analysis as well and so for me
this is my definition of my role from a
dates perspective so it's bringing an
objectivity and a predictability to what
I believe is one of the most fluid
sports in the world I'm one of the most
opinionated sports in the world so I'm
sure everybody in the room today has an
opinion on how my safety played last
game where we good will be bad was there
somebody who was a player that was
particularly strong weak did we attack
well did we defend well everyone's got
an opinion and it's the same with fans
with everybody in the press in the media
with coaches with analysts with within
families people who support one team
another team everyone's got an opinion
so for me my role I see it's important
to bring an objectivity to kind of help
people's opinion support people's
opinions and see what has actually
happened from a different angle from a
very object see fungal so that's how I
see that's how I see my role what I can
produce and what I can bring from a data
perspective I'm from a video perspective
as well so statistics and analytics and
I want to talk about the differences is
there any statisticians are
mathematicians are who were anybody who
works in data analysis in the room yeah
can you tell me what you do can you
share what you do are can you share what
okay ask with know so for me I was from
a soccer perspective we have spent a lot
of time looking at statistics I'm not as
much time looking at analytics so you
may have seen social media at the end of
games there's lots of sort of passes
it's all successful passage how many
passes forward
how many shots how many and goals these
types of statistics for me we've not
looked at analytics and this is what the
main part of my talk my presentation
would be about the differences between
statistics and the differences between
analytics so from here this is from one
of our games this season and what in
this room what you think this data tells
us a below look at it look at it you
from yourselves what do your opinion
what what could have happened in this
game what could have gone on second
sorry very dominant CBO is very dominant
one quote would you agree with that yeah
is there anything else that you could
take from it apart from the overall
dominance of respective dominance of the
team say again so you can't tell who won
who did you think won the game seemed B
there's team a team a won the game yeah
so this is actually one of our games
this season and my next question is
which team do you think my CFC is Team B
seem be have you told him
that's gonna be my next question which
team is my CFC and if you're really good
which team is it that and we played
against and yes I London City when
played on my home first game of the
season okay so this was out this was
some of the basic data statistical
analysis that we did and by this game
the first game of the season that we
played and it's an example just a real
basic statistical overview of the game
so well done I'm really impressed didn't
think anybody will get that a far as
being really clever when I picked that
game okay so my next question is do we
need to question this data previously we
don't set a squad a game just an obvious
question do you think we need to
question this data I think we do because
we've lost the game but allegedly are
statistically we've been really really
dominant in every in every aspect of
being in possession out possession in
possession nearly 70 percent double the
passage apparently box entries nearly
double shot the shots on-target shots
inside the box 8 compared to this so we
still lost a game how did we lose the
game why did we lose a game and we're
going to question so I'm asking you your
opinion now can we make just looking at
this as a from an analyst perspective do
you think looking at this is enough to
make tactically help tactically improve
the team awesome help inform and make
need more historical data definitely any
No okay so yeah definitely something
didn't need more historical data just a
look back and see where we are like said
historical but for me just looking on a
very singular level these figures these
are variables just describe what has
happened there's no kind of underlying
theory there's no underlying explanation
we need to delve deeper into this we can
use it as a I was a good starting point
to see oK we've been dominant but how
what we use it as a gag as a GAD there's
no real kind of meaning meaning behind them
them
and that's how I see statistics and
that's why I feel as a sport soccer has
kind of we've been put back a couple of
years because we've been so focused on
statistics and they've been they've been
locked out so much in the media they've
been they've been built up so much I'm
really we needed to look at data
analytics and we needed to use these
variables to help do this like you
alluded to before create predictability
look it sit and get some more insight
into what is actually going on and
analytics for me is helping us do that
and we're scratching the surface we're
scratching the surface with what we're
doing at the minute and we're scratching
the surface but we're we're at a point
using the data that we have we're at a
point where we can now substantially
substantially say we can do this because
from an analytical perspective we have
the systems in place where every for
example every event that happens on the
field we can quantify and we can
quantify in relation to scoring and
conceding goals because ultimately
that's what we want to do as a team my
job is to try and help my team score
more goals than the opponent scores
against those when get three points so
every event on the field no matter if
it's a Shawn Johnson putting the ball
down for a goal kick and passing to the
right to Fred Breann or max you know our
our Joe Allen or whoever it might be
that has a quantitative and you can put
a quantitative value on that so what
Knights might happen going forward so if
you move into the attacking fear it
might be Maxie Morales beats to play a
1v1 and slides are through ball to WV
running into the box that might have a
bigger that will have a bigger value of
a pass
than the past from Shawn Johnson - it's
a fret so we're now starting to be able
to so quantify everything that we can do
and its effect and its influence on the
school play and subsequent possession
and this is the one value that we've
been looking at a lot this season has
anybody heard of this expected goals
you've heard of it what's your
you want to set the presentation sure
but that basically spot-on so expected
goals what is it he's hit the nail on
the head pretty much there so expected
goals look search shots and it looks at
up to coldest shots over a million
nearly a million a half shots in the
opto database and he looks at the
location on the field the type of shot
was it a header was it a volley was it
weak fought right foot left foot how did
it originate was it from a cross was it
from a free ball was it from a counter a
quick counter-attack a slow
counter-attack slow buildup play looks
at all these different variables and it
puts them together we get a regression
analysis so we look at the comparative
value of every the relate comparative
relationships between all these
variables and its effect and how do they
relate how do they compare and working
interact with each other through this
type of analysis to get an expected
goals so this means for every shot we
can assign a value how likely is that
shot or that attempt to a resort in a
goal our goal conceded why it doesn't do
though what it doesn't do and obviously
as analysts we need to look at the
limitations as well as everything that
we can do and we need to understand
these limitations and the limitations of
the expected goals is I'll give you an
example if Jack Harrison goes down the
right-hand side I'm puts across him with
his right foot although he's not really
likely to do it and and it goes right I
saw you jack it goes right across the
six yard box and there's nobody there in
front the goal to tap it into an empty
net because the goalkeepers dived than
the balls gone past him if nobody's
there and that shot doesn't happen
there's nobody there to take the shot we
can't we don't quantify that and qualify
that with an expected goals because a
shot didn't happen so that's one of the
limitations but for me it's not one of
the major ones as there's a reason
tactically behind there wasn't a player
in that position to have that shot okay
if that makes sense
well else it doesn't account for and
this is probably a major limitation for
me is the teammates defenders
goalkeepers the location of them who's
in front of the player when they've
taken the shot we're not looking at that
a player's their location at in in
position if I was to shoot and you were
to block it we're not taking that into
account okay which is probably a major
limitation but for what we have got on
what we do and how we Jets in how
reliable this and
and how substantial expected goals has
been we can't eradicate it but we're
taking it into account when were working
to make this a part of this okay so this
is the expected gulfs chart so I can
tell you in this blue zone here there's
a 33% chance of every shot to go in the
yellow zone 18% red zone 10% so what
does that mean everything outside of
that so every free-kick Andrea Pirlo was
scored in his career outside here you
probably should have passed
okay saying from from the from a run of
play it's gonna generally be a wasted
possession okay so you should look to
pass or create a further Avenue to
create an opportunity so cystic Leone
over 1.5 million shots okay
I'd let you find me let your own
opinions on that I believe it's quite
solid and I believe a lot of the chances
although and the emotion of the game it
kind of takes over you in these
situations okay was it the right
decision and when we get to a stage
which I'll talk about later we can
quantify or qualify and location and
look at that this I think will become
even more solid expensive goals and so
what do we get out of looking at
expected goals what do we get from this
from this distinct we get did we deserve
to win the game and I'll show you a
couple of graphs and a couple of tables
in a second
how is our team performing in attack
feed and in attack and in defense so how
are we for fine when we score how we
come performing from a defensive
perspective and this is a big one for me
because we can do all the work we want
from a statistical point which we can't
impact practice we can't impact training
we can't impact the 11 players that's it
and the 18 players in the squad that
take to the field every game day
then what's appointment okay so we need
to be able to influence and create
something where this analysis worked and
is it gone juices with how our coaches
work and can affect our players as well so
so
I'll go back to this little graph here
we decided that we can't really improve
performance directly just off looking
these metrics okay so from an expected
goals perspective from this same game we
were expected to count to have scored
one point one goal compared to our land
or cities not point nine one goal so
you're taking all their chances are
their shots everything we spoke about
here and they were only expected and
they were experts to score one goal we
were expected to constant one although
we've been dominant in every other
aspect of the game okay why is that so
this is a chance timeline so look at all
our attempts in blue and all our Landers
in yellow so this maps out in relation
to the time of the game where we've had
our chances and the value of each chance
that we have created our each shot we
have had the value of what's expected to
be to be a goal okay so you can see it's
chuckles to get to this point just over
fifty minutes to get to a similar level
of our land or chance the chance they've
created sixty percent chance of scoring
okay and over the course of the game it
took us over eighty minutes when you
look at how dominant we were they still
caused eighteen minutes to get to the
same level that Isle and those chances
were although we've had we control
possession from possessions that we've
been in there half half the amount of
time with the ball then we did without
it they stalk was 18 minutes to get
there so this provoked further further
and further questioning we take at this
we can use this as a good guide of the
periods where we've been quite dominant
but we create the chances that haven't
been that high in quality should we say
so this next one
now we're looking at where our shots
were from so relative if you remember
the map I showed you with the circles of
the blue red and yellow zones looking at
its actual chance we've had all inside
the box these are the eight shots
compared to all angles one chance and
how good this chance was for them to
score but as you can see we've had one
shot inside the Blue Zone where she's
fifty percent chance of scoring we've
had free shots inside the yellow zone
which is fifty four percent chance of
scoring saw them to them attempts
combined probably gives us that goal we
should have scored tatata game so we
created them similar chances there and
would think okay we can look at them
chances in a little bit more data how
did we create them chances why did we
not score outside of that we had four in
a red zone
but she's for any chance of scoring but
obviously we need to look at them in a
little bit more detail as well so now
looking at everything combined our
expected goals knowing what with
understanding my expected goals and how
we formulate and calculate is looking at
the chance timeline and the types of
chances we've conceded there can we
start to make a few tactical decisions
now I can we start to inform the process
yes no what do you think we're going
into a little bit more detail and we can
quantify a little bit better and I have
a better understanding of the types of
chances we've created in relation to our
dominance can we now start to formulate
that's that's working say that again sorry
yeah yes so now we have a better
understanding we have a better and we're
better educated in what we need to do
where we want to go how we want to do
that yeah okay definitely a hundred
percent but again for me tactically we
still can't really influence we still
can't really influence from this we have
a better deeper understanding and it's
pointing us in a direction and it's
pointed us in a direction where we can
do this start reviewing specific footage
so it's giving us a real form the
standing of where our chances have come
from the quality of them chances and
pointers in the direction of the areas
relative to the timeline where we need
to look how we need to look okay so we
go back and we review this and then this
would work on specific technical and
tactical content during training session
so that was pointers in that direction
to use the video to then see okay so
these chants would be created which had
first three percent chance of have been
converted what kind of chance was it and
how did we create that chance in the
context of that game so then we can look
at that chance was there ways that we
could create better chances or how did
we create that chance and then we look
at the other the chances that weren't
maybe so good and why was that why was
that so it's provoking us and looking
out in a different perspective so this
is where we are now in relation to
expected goals so expected goals in the
league we have the highest highest in
MLS so about of all the chances we've
created while the analysis that we've
looked up after all the chances during
the season we've gone through we've had
a lot will seem where I'm gonna tell you
out because I've given a lot of secrets
away but we've looked at how we create
our chances and now we're just above
Chicago in terms of goals up we have
expected to score which is for less than
what we have actually score
so we're outperforming what we are
predicted to be and we're supposed to be
this season but using this data so
pointers in the directions to guide us
and to inform the coaches and we've been
able to structure
training we've been able to use it as a
guiding principle we're able to use it
to to guide our focus on
goalscoring and and defendant and I
really on a really basic level and our
coaches take this information really
really well so come up with decisions
like this but we're using their
expertise from a soccer playing
perspective and coaching perspective
mixed with their data to come up with an
object if you point in a standpoint of
where to focus our attention because we
want to focus our attention on the plays
that are going to scores more goals and
stop us conceding some really obvious
level we want to focus on that and this
and the data is giving us that direction
to be able to focus as well as taking on
board the subjective feelings of the
coaches from a tactical perspective it's
really really important because they
have really good experience great
experience in the game they know the
game really really well their intuition
is not something that as data analyst
there's a farm science what we can
ignore notion no way because their
intuition and what they feel about the
game how they see the game is really
really important but we from an
objective standpoint need to use this to
help okay is this correct if it's not
how can we look at a way of saying okay
from their Jets if some point this is
what we need to do this is how we need
to do it okay then you come together so
it's a really really strong working
relationship one myself daily have with
the cultures of their of my CFC so just
to finish data analytics and we're next
so alluded to it before with location of
players and being able to map from a
technical standpoint where they are on
the field and how their relationships
integrate with each other and how my
movement five yards forward could
influence your movement five yards
backwards how does that in the middle in
and rating a goal scoring opportunity
and this is XY coordinate data location
data is what we're calling it this
company's core track up
trial in in the Premier League this type
of data so we're on the cusp of it to be
able to really solidify everything that
we do from data perspective and to be
having a little bit more predictive
analysis or what could actually happen
so try and map it out from the real
objective standpoint of what is going on
from a numbers perspective when we get
this XY data it's going to be really
really exciting for the sport and you'll
see a massive boom and a massive change
in how data analytics is perceived in
the outside wider community so we're
still we're looking for this and then
and last but not started on it yet but
as a group CFG from a research
perspective we're leading the way in
Europe and in England on trying to get
this and push this forward and try out
trial in different algorithms trial in
different software companies just to see
where we are at with this the location
data because once we get that from a
technical perspective and comparing the
interactions of every single player on
the field and have imagined an Excel
data sheet with millions and millions
and millions the roles being able to go
through that on how each players
movement has affected the next place
then we really start to see the game
from a tactical perspective and it
really starts to relate to the numbers
Click on any text or timestamp to jump to that moment in the video
Share:
Most transcripts ready in under 5 seconds
One-Click Copy125+ LanguagesSearch ContentJump to Timestamps
Paste YouTube URL
Enter any YouTube video link to get the full transcript
Transcript Extraction Form
Most transcripts ready in under 5 seconds
Get Our Chrome Extension
Get transcripts instantly without leaving YouTube. Install our Chrome extension for one-click access to any video's transcript directly on the watch page.
Chrome Extension