WEBVTT
Kind: captions
Language: en

00:00:00.480 --> 00:00:06.480
That's his third 20-yard reception of the game,
and this one is gonna go for a touchdown! You got

00:00:06.480 --> 00:00:10.622
number 88 on your fantasy football team, you've
got to be loving this game right now.

00:00:11.417 --> 00:00:16.720
Chris Anthony uses math to predict the future, albeit
in a different way. He writes a column for the

00:00:16.720 --> 00:00:21.280
National Football League's website.
He is obsessed with fantasy football.

00:00:22.400 --> 00:00:25.760
I have a natural love of football.
I was born in Green Bay, Wisconsin,

00:00:25.760 --> 00:00:31.360
so it just kind of is in my blood, and also as
a professional, I work with data all the time,

00:00:31.360 --> 00:00:36.880
and building FantasyOmatic was my ability to
converge the two together. About 10 years ago,

00:00:36.880 --> 00:00:42.000
I wanted to figure out a way that I could beat
my friends and family at fantasy football and I

00:00:42.000 --> 00:00:48.080
decided to build an app in FileMaker, and I was
just kind of using it for my own hobby purposes,

00:00:48.080 --> 00:00:53.280
and all of a sudden, I had an idea about a way that
I might be able to kind of beat the system. So,

00:00:53.280 --> 00:00:58.400
I decided to build an application to do that.
The fantasy football industry today is about

00:00:58.400 --> 00:01:04.880
18 billion dollar industry, so it is a large
industry, right, and we're having an impact on

00:01:04.880 --> 00:01:10.800
that industry using the tools that we've got for
machine learning in FileMaker. We have been able,

00:01:10.800 --> 00:01:17.120
over time, to do a better job predicting fantasy
football outcomes and FileMaker has grown with us

00:01:17.120 --> 00:01:23.167
and we've been able to take advantage of some of
the newer tools in FileMaker, ah, to help us do that.

00:01:29.440 --> 00:01:33.440
Since the application has been around for
10 years, and FileMaker keeps updating and

00:01:33.440 --> 00:01:36.960
giving us new technologies to work with,
we're able to modernize the approach.

00:01:36.960 --> 00:01:41.760
For example, we used to use an API connection to
a system that would run our machine learning model

00:01:41.760 --> 00:01:46.400
and create the ratings for us. Now, we integrate
that natively into the platform using CoreML.

00:01:46.400 --> 00:01:51.920
Once I got on with FantasyOmatic, a lot of my
time was spent on developing and working on the

00:01:51.920 --> 00:01:56.720
predictive game models. Those game models
ended up actually predicting three of the

00:01:56.720 --> 00:02:02.800
first four Super Bowl winners. We actually were
able to use that system to outperform a lot of the

00:02:03.440 --> 00:02:08.720
big data predictive industries on their
predictions, and their models. At FantasyOmatic,

00:02:08.720 --> 00:02:16.080
we we actually use a calculation field, uh, to
write a weekly preview article for players to

00:02:16.080 --> 00:02:20.880
look out for each week, and the readers of
course don't know that this is actually an

00:02:20.880 --> 00:02:26.160
AI that's writing the article. Here we are, a
really small team, being able to compete with

00:02:26.160 --> 00:02:31.120
big data and large organizations that are throwing
millions at these applications, and we're able to

00:02:31.120 --> 00:02:35.840
do it more efficiently and with more agility
because we're using the FileMaker platform.