Algorithms & Deer Movement

Before cell phones, before the internet, even before trail cameras, we as hunters used to just get in to he woods and hunt. Theses days it's a completely different story as science and technology have infiltrated the way we hunt, predictive deer movement applications like Spartan Forge is one of these pieces of technology.

On this episode of the Hunting Gear Podcast, Dan talks with Bill Thompson of Spartan Forge, a mobile hunting map application designed to help hunters find the best days to go hunting. Bill talks about the advancements his company has made over he years, the science and technology behind the app, and some interesting facts about what he as seen from the data he collects every year.

[00:00:00] Dan Johnson: Go Wild is a free social community created for and by Hunters. This means that unlike mainstream social media, your trophy pictures won't be censored. They're encouraged. As you spend time on Go Wild. You will earn awesome rewards such as gift cards, free swag, and big discounts on brands like Garmin and Vortex.

You will even earn $10 just for signing up. Visit, download, go wild.com and sign up.

Ladies and gentlemen, welcome to another episode of the Hunting Gear podcast. I'm your host, Dan Johnson, and today we're going to be talking. With Bill Thompson of Spartan Forge. Now, if you're not familiar with Spartan Forge, it is a mobile hunting app that you can download to your phone. They also have [00:01:00] desktop availability.

But man we talk a lot about technology today. We talk about algorithms, we talk about collared deer studies, we talk about 3D map imagery, and just like it, it's crazy. Where hunting technology has gone in the last several years, cuz I can remember before, cell phones, before internet, before even trail cameras were out, I would just go into the woods and hunt.

And now it seems like if you're not taking some type of Some device with you that has, that, I would use the word technology. That is technology like a digital technology a computer program or something like that. Then you're lost, I use HuntStand right now, and I just imagine what you take those mobile apps away from me and having to go back to the old printed maps or the old.

Books and what that would be like in order, like [00:02:00] it would be absolutely crazy. And the industry has changed, the technology has changed and Some say it's for the better. Some people don't like it. But today we get into a really detailed conversation with with Bill about Spartan Forge how he came up with some of the stuff that he's come up with, the advancements that the app has made over the the last couple years.

How all of. All the data that he collects translates into information given to the end user on what days. Hunt. If you're not familiar with Spartan Forge, it is a predictive deer hunting, a predictive deer movement. Platform that tells you basically in your area what days are gonna be the best to hunt given, collared deer study data given weather how Im weather impacts deer movement, historical data like that.

And so it's a really [00:03:00] interesting episode and I'm sure if you guys are deer nuts, this isn't a hundred percent gear talk. Cause we also get into things like what drives. Certain aspects like humidity and precipitation and how that affects deer in different parts of the country. So it's a fun episode.

I always like talking with guys like Bill, who are extremely intelligent. And even at the beginning of the episode, we get a little background in, into Bill and his time spent in the military doing whatever it is he was doing in the military. So it's, it was, it's a pretty interesting episode. I know you guys will.

Before we get into today's episode, though, I do have to send a shout out to the people who make this podcast possible. If you guys are looking for a saddle. And I know that this year I'm gonna be taking I'm gonna be taking a couple more, what I would say, less spot in stock, more tree stand hunts. And so I'm getting ready to I'm getting ready and I'm excited to use a saddle more this year.

So if you're looking for a saddle hunting, accessories, platforms, climbing sticks, [00:04:00] you name it, go check out. If you are looking to document your hunts, then I strongly suggest you go check out Tcam. The Tcam has the new 6.0 version currently out. It has image stabilization. It has an L C D screen.

It records in four 4k. It can be mounted to your bow, it can be mounted to your gun, and that way you can record what out in the woods and then you. Come home, show the wife, show the kids, show your buddies even helps with shot placement. If you're unsure about where you shot a deer, let's say, you can take it home, throw it in your computer, review the footage, and then that'll tell you if you need to wait or go after the deer.

HuntStand is the next one. Let's see, T cam and then HuntStand. If you're looking for, today we talk a lot about mobile apps. So if you are looking for a mobile hunting if you're a hardcore deer hunter go check out. HuntStand has a lot of functionality behind it.

HuntStand has is the most, one of the most [00:05:00] popular apps on the market for a reason, and it's just because it gives you a lot of options. And while you're at huntstand.com, read up on all that functionality and then also read up on their pro whitetail platform that they've just. I guess introduced over the last year.

It's pretty intriguing. So go check that out, huntstand.com. And that's it. The other thing I wanna say real quick is we do a lot of talking about gear on this podcast, very little talk about con conservation. So if you are looking to give back in 2023, just, I'm not gonna sit here and try to pitch it.

Go look@fishandwildlife.org in, in the conserv. The conservation organization, 2% for conservation. Go check that out, see what they're all about. And if you want to get 2% for conservation certified, go check that, that website out. So we're done with the intro. Let's just get into today's episode with Bill Thompson of Thompson or Thomas.

No, I messed it up. [00:06:00] Either way. I apologize. Bill from Spartan. All right, from Spartan Forge, I have Bill Thompson on the show today. Bill, what's up man? Not much. Dan. How are you doing? Good. Hey, remind me again where you're located. I'm right now I'm out in 

[00:06:18] Bill Thompson: West Virginia in Bridgeport. I retired from the military last year in November, and my wife got a job out here.

So we've been out here in West Virginia. I can work from anywhere, but. I suppose it's been a couple years since I've been going back and forth from out here, but Yeah. 

[00:06:34] Dan Johnson: In West Virginia now. Gotcha. And so let's see, how many years in the military and what did, what was your kind of background in the 

[00:06:42] Bill Thompson: military?

Yeah, so I did almost 21 years. My background in the military is complex, but I started as an enlisted guy doing like signals intelligence, which is. , just think like radios, cell phones, that type of stuff. Signals, intelligence, exploitation and radars as well. And then I [00:07:00] pivoted from there to going into the human intelligence realm, which is sources and that type of stuff.

Deployed and doing that type of work. But from a technical point of view I became a warrant officer about 11 years ago, and the easiest way to think about a warrant officer is like a technic. In the military, just advising office general officers and colonels on what types of technologies are good, which ones are not good, how they can be integrated into the unit to be like a force multiplier or a a commander while they're, overseas or doing their mission in Garrison or whatever.

And they kind of warrant officers sit in between. We're not soldiers, we're not we're soldiers, but we're not enlisted. , and we're also not officers. We're like in between. So if you think about, the military as like a school, your NCOs are like your teachers, whereas your lower enlisted guys are your students.

And then you have your officers, which are like your principals, your vice principals. Administrators, that type of thing. Warrant officers would be the guys who are like recommending the curriculum and what books to use and what, that type of thing. I [00:08:00] gotcha. The technical experts of the 

[00:08:01] Dan Johnson: military unit.

Okay, I gotcha. And in order for you to recommend those things, you have to be somewhat well versed in those things. So what kind of training did you guys have to go through? Is it. Drone training or is it more hey here are a list of educational tactics that you can use to train your soldiers better.

[00:08:23] Bill Thompson: So for myself, the training that I went through was a lot of technical there's technical and tactical training, but a lot of it was computer programming. Network, like network diagnostics, understanding how, what networks are, what their vulnerabilities are, how they work. And then on the, and then there are other disciplines too where I learned other, you know how to go and meet a guy for a meeting and in the middle of maybe, in Pakistan somewhere.

And yeah, arrange a meeting where there's operational security and there's. Your physical security and you're thinking about [00:09:00] all of the third, second, and third order effects of those things. But I was, again, looking at those types of things from a technical perspective. So the training was vast.

There was lots of different, I, we could do a whole podcast on just the training and all of the stuff that go into that from the technical side to the operational side and the tactical side and all of that. But there was a lot of it there. And I guess if I had to summarize it all, it would.

Lots of learning, coding code bases network architecture, and then the tactical implementation of. The exploitation of those types of networks and so on and so 

[00:09:32] Dan Johnson: forth. Wow. Everybody talks about how big the United States military budget is. And we have the biggest, baddest military in the world for a reason. It's because we spend a lot of money on it. And were you able to get your hands on some pretty. , I dunno whether that's maybe software or just things like that are just like, holy cow, this is mind-blowingly awesome. 

[00:09:58] Bill Thompson: Yeah, so it[00:10:00] I'll say this about the, my experience of the US government doing on that side of the house.

Some of the stuff I did you could consider like ethical hacking. Yeah. And I think my last year I was an advisor. I was an advisor for a general. and it, and for a colonel that were doing development in what was called the offensive cyber realm, so that's ethical hacking. I believe that year, the US budget for the d o D was something like $750 billion

And I would say that 60% of that budget is judiciously executed, whereas the other 40% is wasted. Yeah. There's some reason and it's necessary for waste, but a lot of times it's just the government is very resistant to updating and resistant to cutting the fat. But so there are cool programs that I worked on and there are cool things that I built, but at the end of the day, it was the ability for me to do the kind of stuff that would be illegal in any other context.

I gotcha. So whether that's breaking systems, breaking [00:11:00] networks Think just, let your mind wander all of the things that you would want to do if you had, a free reign to break as much as you wanted to. And that's the role of the military is not to get too deep into it, but they have what's called, title 10 authority.

Yeah. And Title 10 authority just means you get to break shit. Yeah. And as long as it achieves the goals. So yes, there were cool tools and there were cool things that we got to use, but at the end of the day, it was more about the tactics and the operations that we got to engage in that I think really separates us from everyone else.

Cuz anybody else who were, who was doing the things that I was doing for my last 10 years, my military career would've been arrested and in the other context, yeah. So 

[00:11:36] Dan Johnson: yeah, that was the best part. That's absolutely crazy now. I've talked to some guys before who, I don't know if I would say they were directly in the same type of platform that you were in or doing exactly what you were doing, but they did some pretty cool things.

Where was the balance between like stress and, dude, this is awesome doing this [00:12:00] type like. . 

[00:12:02] Bill Thompson: Yeah. So the stress was more on like the family side? Yeah. Or not being around my kids when they were young. I missed the birth of my son deploying to Iraq, I think for my second tour to Iraq. The, for me at least I feel like I'm an outlier in the military because I just tr benefited Tremend.

For my military service in a way that I think is exceptional. And really a story that's only possible in the us. So I, I joined from a trailer home in the middle of nowhere, North Dakota where I really had no prospects for my future. And, got into the military and lucked into some things, which is again, a whole nother podcast that I could do about, one of the reasons I got into intelligence work is just because I simply forgot my driver's license.

So I was signing up to be an MP cause I was 16, maybe 17 at the time when I was signing up. And I wanted to be a military policeman and I just simply forgot my driver's license. And I was like, the recruiter's you can come back next month. Cuz we had [00:13:00] to go to Minneapolis from where I was from in North Dakota in order to sign up.

And I was like, look, I'm not going home for another month. Sign me up for something else and get me the hell outta here. and so to k to answer your question, my military service, I just benefited tremendously. Gotcha. I came out with degrees, I came out with certifications. I came out speaking three languages.

I came out with a wealth of knowledge and it's one of the reasons why I pushed so hard. When I started the company and I was bringing employees on, I was like, look, everything that we're gonna do, we're gonna try to benefit military people. in one way, or, we could talk about it later, but we do these veterans hunts and that type of thing where we're raising money for people that didn't benefit from my military service.

So for me, most of the suck involved in the military was being away from my kids, being away from family, missing things like death of grandparents and yeah, the type of stuff that you would want to be around for. But on the other hand, I did there's. I, again, [00:14:00] I could fill a whole podcast with what I thought were all of the awesome things I got to do in the military.

Breaking things, coming up with cool technology to assist the special operations command. Getting a latitude to do all of the kind of like cool stuff that you think for the first five or seven years, I really didn't get to do a lot of cool stuff. It was a lot of motor pools, it was a lot of checking vehicles.

It was being a grunt. Ruck sacking around and carrying stuff and PMSing vehicles, preventative maintenance and checks on vehicles, and just being a grunt. And then I fell into getting to do other stuff. Later, after, I guess I'd proved myself where I did get to do the stuff that I think most people dream about whenever they, a lot of guys when they joined the intelligence discipline in the military.

Dream about the stuff that I did for the last 13 years, and again, just super beneficial and super fortunate to get all of that , that, that's that's what it was to me. I didn't mind deploying, I didn't mind doing operations. I lived for doing operations. I lived for supporting the Battlefield Commander.

I lived for doing, technical [00:15:00] assistance for human intelligence operations, for signals intelligence operations. I used to fly around in airplanes and set up operations for ODA teams and special forces. To go in and do stuff. Like really got to get all of the benefits from it. And really the problem for me was just being away from my kids, but I'm making up for that now.

So there you 

[00:15:17] Dan Johnson: go. There, there's always time. There's always time. Yeah. Okay. 21 years in the military, you step away from it. You're retired now. Now how many years ago did you start Spartan Forge? 

[00:15:29] Bill Thompson: I actually started Spartan Forge while I was deployed to Afghanistan in 2015. Okay. So I was still active duty.

I, I didn't incorporate, I incorporated the business in 2017. The first name I had come up for Spartan Forge was called Open Season. And I started collecting collared g p s data and dealing with academics and talking to people. Through my experience. I was working at the time.

I had just gotten done advising on a development effort with darpa, which is like the mad scientists of the d o [00:16:00] d. They just get to think up all kinds of crazy stuff Hey, let's put a rocket on Mars. They do that type of stuff the, and so I was working with them and I had met some academics and made some inroads to people, and I had built or assisted and supervised the building of other neural network.

To benefit the government and the military. And it started occurring to me that I could probably do a lot of this stuff with GPS data from collar, our deer studies. So I started collecting that. I made my first phone calls to academics and talking to people just during downtime while I was in Afghanistan in 2015.

Between operations I would, just reach out and say, Hey, I'm here doing this type of stuff, and I think I could do the same thing. And a lot of academics were willing to share that data. And that's when I started doing that. I've been collecting it ever since. I just got some more deer data the other day that we're working on integrating to our neural network.

So pretty continuous since 2015. Yeah. 

[00:16:52] Dan Johnson: And When you first started Spartan Forge did or the company or what, that you had before Spartan Forge? [00:17:00] Before the name tag, was it always about forecasting deer movement, or did, had you planned into going into the high de the, the high definition, very detailed mapping and all the other stuff that comes with Spartan.

[00:17:15] Bill Thompson: Yeah. So no I wanted to focus on neural networks. I always saw Spartan Forge as a machine learning company. There are other things I want to do and I still am doing in my very little spare time in the neural network, machine learning, artificial intelligence realm that I do. And I always wanted to start the company as a machine learning company.

I, I still call it a machine learning company. I was originally supposed to be working. worked for about two years with another very prominent mapping company the most, probably the most prominent mapping company. And had basically gotten to the point where I was gonna sign a contract with them.

But then, things came up for me and I didn't, the contract wasn't favorable because it locked me down for many years. , [00:18:00] and I always thought, I'm want to focus on the machine learning side of stuff. I just need to integrate with an application that's already doing this, and then I can help them do it better.

Because a lot of these companies that are out there are just dealing in commodity data. They're, it's just, you can think of them as British Petroleum AMCO and another oil company, like all three of them are just selling. They're not doing anything crazy or innovative with the oil mapping data is out there.

None of these companies that launch these programs are collecting their own mapping data. They're paying other companies to go and do it or to get the information. So my focus was always what do we do with this mapping data? Or how do we improve the lot of hunters by capitalizing on this data and then integrating it with other pieces of.

That. And creating a comprehensive planning and execution picture for hunters. A lot like I did with the military. A lot of the programs I advised on the military were multi [00:19:00] intelligence, multi-source intelligence collection systems that presented information to commanders in what's called like a common operating picture.

This common operating picture is like from a military perspective, it's like a snapshot of what the battlefield looks like and all of the amplify. Information that's collected and updated so Commander can look at it at any time and have a pretty good picture of what's going on in their area.

Responsibility. My goal with anything I was gonna do in Spartan Forge was to integrate with another company and then show them what else was at the art of possible and put it all together in this, common operating picture. And Spartan Fords itself now is moving towards that.

I'd say we're about 40 or 50% of the way there. And yeah I, my goal was to integrate with someone else, but once I, I did a lot of investigation on some of these other companies and I don't wanna speak out of school on here. I encourage people to do their own look into these companies, but I didn't like some of the investors that they had.

I didn't like the core tenants or principles of the investors that they had, or the [00:20:00] leadership that they had. Whether it was because they were anti First Amendment or anti second Amendment, or. The, or the investors in some of these companies. And then, but I always said to myself, I'll still join with these companies because I'm, I don't want to do this all on my own.

But then the contracts were just so litigious and that, they wanna do a year-long contract with you then lock you down for not doing anything else with anyone else for five years. Yeah. And I just wasn't willing to do that. Then I basically was on the call one day with a guy who was in charge of one of these companies.

He was the president, I believe. And he just kept telling me like, look, just come with us. Do the mapping, do the neural network stuff. Let us worry about mapping. You don't wanna worry about mapping. Mapping's really difficult. And he said that three times to me on the call. And by the end of the call I was like, you know what I'm gonna do mapping anybody who says it , anybody who says it's that difficult, that.

Must be lying. Yeah. So I went and looked into it and I was like, this isn't difficult at all. The difficult part's, the machine learning part. So that kind of [00:21:00] led me to where I am today. Yeah. 

[00:21:01] Dan Johnson: All right. So you've used terms like machine learning and neural network. Yes. Yeah. So for a dumb old deer hunter like myself, explain what that is.

[00:21:12] Bill Thompson: So a neural net, the simplest way to think of a neural network is that it's a computer system. Recognizes patterns in data and then can make predictions on those patterns. Gotcha. So it, the ways that's called the neural network is because it's inspired by the structure and the function of the human brain.

So in a neural network there's like layers of neurons and each one of those neurons. Become weighted based on the data that you show it. So the simplest way to think about this as is, and I'm gonna kind of jack this up, but I'm trying to make the analogy work. If you have three ways that you travel to work from your home to work, if you have three routes that you take, you can think about that like a neuron, like a structure and say over the last year, your time in the [00:22:00] car is measured.

Every time you drive a certain route or a different. You could have a machine look at that and say over the past year, what was the most efficient route when it was raining? What was the most efficient route when it was sunny? What was the most efficient route during this time? Cuz there was traffic or this time when there wasn't traffic.

And then that neuron gets weighted. And then in the future, you can look at all of those conditions and then make an informed prediction on what route you should take to work based on past circumstances. So a neural network is just millions and millions of those neurons informing a decision based on patterns.

And it's the same way that like the human brain works, the human brain. We're just pattern, we're just really sophisticated pattern recognition machines. We recognize color. Like rgb and we recognize personalities and we recognize patterns of other animals and our environment and what things look like.

And those patterns are com comprises the structure of the brain. So in neural networks, you are just [00:23:00] making new neurons based on other data. So in this case, the first network that I made was, or, and my, myself and my founders made was based on caller GPS data to inform those neuron. So the, during the training, the neural network's given a set of input and corresponding desired outputs, or it, you ask it for what you're getting from it, and then you can measure it and it's success based on other data sets.

So say I get a bunch of collar deer data from Ohio Deer and I make a neural network for Ohio. Or for, I've made it for other, And then I'll get data from somewhere else. And then I'll wait and see how similar those Deere move or how un dissimilar those deer are and how they move. And then I'll make another network based on the new data, but I can test the old networks based on the new data.

So it's that's the I'm, I hope I explained that cuz it's 

[00:23:48] Dan Johnson: it's difficult. Is this, would this be an accurate statement? Basically what you have is just a ton of data and you're organizing. In a way that people can find, they can see it [00:24:00] and find patterns in it.

[00:24:01] Bill Thompson: Yes, exactly. Okay. The machines recognize the patterns and then they try to prescribe the patterns to people. Gotcha. The most difficult part of any neural network is the data collection. Yeah. Having a lot enough data to do it. Like I could teach you how to make a convolutional neural network in an afternoon.

The algorithms have been around for a long time. It doesn't take a ton of coding experience to actually do this part. Google's made a lot of it very easy through a program called TensorFlow. So you can do it on your own. It, the difficult part is getting the requisite amount of data to train it.

[00:24:34] Dan Johnson: Gotcha. All right. When we first talked, as a deer hunter I live in Iowa, you and just. Conversation purposes. You just mentioned Ohio, right? And so I live in Iowa. How does deer movement in Ohio or any other state, like what does that have to do with me here in Iowa and how can I look at that data from different [00:25:00] states and go, Hey, and this is talking about predictive deer movement at this point, but why does deer movement in Ohio matter to me in.

[00:25:09] Bill Thompson: Sure. So there, there's a few ways to answer that question. The first way to answer that question is they all are all the same species. So when you have the same species, there are evolutionary underpinnings that kind of are in every animal that are the same, at least from birth. And the only way that you can get rid of them or change them is through very harsh conditioning.

You can think about it. you like, what does Dan Johnson have in common with a guy from Africa? If I throw a baseball at either of them, even if the other guy's never seen a baseball, they will generally flinch and throw their hands in their face as the first reaction. And that comes from a very deep part of the lizard brain that every human has.

So reactions to stimulus are gonna be the same across a species [00:26:00] regardless of where the species lives. So if deer in the same. The reaction to weather events the re or to what weather events and how it influences deer and deer movement, or the responses to things like drought or different types of factors are gonna be the same at some level, or at least we know it's gonna influence the animal at some level and the same, but the difference has to become, and the reason why neural network's important to use in.

because the neural network can look at 50 pieces of input data and say a deer in Iowa and Ohio are pretty similar. But when I'm looking at this from North Carolina, you know what, the amount of time that adverse weather affects deer in North Carolina versus Ohio are not the same for different types of the year.

So for a three day rainstorm, it's gonna affect the deer. North Carolina differently than it would affect the deer from Ohio. But what the neural network knows is that it affects both. [00:27:00] So it's just the easiest way to answer that is the predictions differ and ba basically based on where you are, latitude and longitude wise in the US based on stressors in the environment differently, but it affects them.

Just differently. And that's the role, the proper role of a neural network. You would never be able to get some, a person to recognize all of those nuances. You'd have to hunt your whole life first, just to understand deer in an area accurately, but then to understand deer in different areas accurately in different places and what gets them going or what doesn't get them going or moving or all of those things would be an impossible data task for a human to carry out.

So that's the proper way for a neural network to. Separate the wheat from the chaff and find out what in a weather forecast is actually gonna impact deer movement and how, okay. So I hope that kind of answers that question. Yeah. Also, by the way, I have no short answers, , so I appreciate, I, I [00:28:00] apologize to everybody in the area, this stuff is very there, there's a lot going on here, and when I ask questions, I'm trying to be as accurate and as data science oriented as possible, but still presenting a palatable product for a podcast.

Yeah. So I, I hope I can manage that tension. Yeah. 

[00:28:16] Dan Johnson: Okay. And p positioning off that then, all right, we're still talking about predictive deer movement. How do I, the hunter use everything that you've just said. I download your app, I start to use it. How do I use it and how do I use it properly?

[00:28:34] Bill Thompson: So for there are a variety of different ways to use all of the features in the application, but just focusing on the neural network you should think about it as in general Deere movement. You shouldn't think about it in specific, it's not gonna help you. , it'll help you, but it's not gonna, it's not something I would use biblically against a five, six, or seven year old buck [00:29:00] because the one thing I've learned from looking at all of the GPS data, and I've looked at tons and more deer data than I think probably everyone anyone's ever looked at.

I, I'm not sure I've met somebody who's looked at more individual deer in their movements and what affects them. One thing I can say empirically about it is Bucks just none of them are the, and all of them are different for different reasons. So when it comes to using my neural network for a buck, especially a mature one, you're gonna have to allow it to or try to understand how it influences that deer.

And when it's correct and when it's not correct and then try to use it, but more use it for just general deer movement. So if you're just out trying to schwa a dough or you're just out trying to see deer, or you're taking a kid out with you, or you are trying to correlate movement with deer cameras and other things, then the neural network's super handy.

But there will never be a network and it's why I laugh to myself whenever I see these other predictive. Where it's like, Hey I'm Joe hunter. I've developed a neural network for a [00:30:00] deer that's gonna get you the biggest buck no matter where you are in the us. It's like none of them are the same.

And there's no way anybody, one person could develop a neural network that's gonna help you kill a buck, or that's gonna accurately say every deer in the woods is gonna stand up at, you look at some of these other apps and it's what? You need to be in the woods at 1:45 PM. . A. And so you think about that, and a lot of people get excited and they're like, all right, I'm running into the woods at 1 45.

But then you sit and think about the data, the proposition of the data, and it's so wait, this thing is saying that every deer in the woods is just gonna be moving at 1 45. That's not possible, right? So what the hunter needs to understand what the neural network like mine is that it's got thousands of years of deer data.

And what I mean by that is if one Deere wears one caller for eight years, that's eight years of deer data. If three deer wear three callers for eight years, that's 24 years of deer data. I've looked at, thousands of years of this dear data. And so the, what the network has sussed out from that is you basically have three types of move.

That categorized deer [00:31:00] movement. And that is during the daylight hours, they're gonna be staying in their bedding areas. They're, they may move out to transition areas or they could be anywhere in their range. And those are the three buckets. Anybody saying that they have a network that predicts any better than that is not being honest with you.

And the reason is because after all of this Deere data, that's about as good as it gets. So you can think about that as it, the value proposition for a hunter. When I'm looking at the neural network's gonna tell me, generally deer are gonna be staying close to bedding during day hours today, or they may move out into bedding in the staging areas or into a scrape line, or they may be just outside of bedding areas, or it could be their, what we call full range in the way that a neural network looks at full range.

If you are driving by a field every day to work where you never see deer, then all of a sudden there are 21 deer in that field, the neural network would call that a full range day. In other words, deer are out there moving much more aggressively than they normally would.[00:32:00] And there's a litany of factors that influence that.

And that's the best that you can hope to get from what would take you. You wouldn't have enough humans observing deer for a mon enough time. To get that any more clear on that data that, and that network that we've developed predicts accurately still only at 66, 60 7% of the time.

And I'm very clear about that because I'm not trying to oversell somebody or give somebody snake oil on something, but you're getting two-thirds of the time, you're getting an accurate prediction on what most of the deer population will be doing. Gotcha. When it comes down to patterning individual bucks or going after any individual bucks, you might find.

It might inform the neural networks, the neural network might inform me Someon saying, Hey, generally whenever it says core area day, my buck's not leaving Bedding. Yeah. And so that's something I would tell people to look at a one-on-one instance, because again, I'm not gonna sell 'em snake oil.

And tell 'em that this thing's far better than it is. It might not be great marketing, but I can sleep at night. 

[00:32:57] Dan Johnson: Yeah. And I've talked with [00:33:00] just about every predictive deer hunting model, app, whatever you wanna say that's currently out there. And for the most part, . There are some commonalities.

There are some similarities between yours and some of the other ones that are out there. But and you there just mentioned the, there's like a handful of things that are really th that go into the equation that predicts deer movement. The word algorithm pops up in, in some of this and.

How the out, how certain apps pop out, what days to go hunting and what times to go hunting and things like that. If I said to you, . I have 40 years of hunting experience and I go in and break down what my hunting experience has been in certain conditions in certain times a year, and I implement that into my algorithm.

Does that [00:34:00] benefit the, out the output of data at all? 

[00:34:05] Bill Thompson: I would say in that area, that person has learned and influenced and been able to control that. And set up, they're obviously very people that make those types of networks probably know a lot about their deer on their property.

But again, what moves deer on their property where they've learned or where they've looked at these things is not necessarily what's gonna get mo deer moving in Saskatchewan or Upper New York, or in Florida, which is why we collect this data from all over the us. The factors that get deer moving in Florida are not the factors.

I'll give you an example. In Mississippi, relative humidity and humidity in general is bears a factor on deer movement. It has almost no factor on North Dakota deer or Minnesota deer. It doesn't change the, it doesn't affect one way or the other really, that I can see in the data on what gets deer moving.

Another one that's, . Yeah. It's just, it depends on the state that you're in and [00:35:00] when I'm looking at the data, but then also when you start getting into areas, one of the, one of the things that affect your movement the most is rainfall. In Alabama, for instance I've been saying this for a couple of years now.

If there's a light rain, I would be hunting all the time. It seems to really get deer moving down in those. And I, I've talked to a lot of academics about this, and I think one of the reasons is because they have flash flood scenarios. And I think it, it helps the deer if they're dynamic during rainfall to make sure that they're not getting washed away or something else.

And it may be that they just have an evolutionary mechanism that kind of keeps them moving during the rain. Yeah. Because you never know what's gonna happen. Whereas again, That where in flatter areas or in the Midwest where flooding isn't the problem in a specific area outside of like the Red River Valley they, the deer don't react to rain like they do in someplace like Alabama.

But then again in a place like the mountain country in the No. Pa, North Carolina, the [00:36:00] Blue Ridge Mountains and out west. You'll see factors like cloud cover and thermal generation affect deer movement. Much more deer and the northeast need reliable thermal generation and in order to scent check areas and they need reliable cloud cover in order to understand how the wind is gonna shift or be different as a result of movement, especially in pressured areas.

Those things really change how. Nothing will get a deer moving in Pennsylvania, like a drastic wind shift. Drastic wind shift among, especially amongst mature deer. They might not move far, but they will move when wind shifts because they want to be in an area, especially when they're betting, where they can take advantage of the wind direction and thermal generation and the sun's placement in the sky so that they can smell everything around them.

If you have a consistent wind, that's the normalized wind for that year, you're not gonna see a lot of deer. Daytime deer move. In pre rut or, mid-October in Pennsylvania. Yeah. Whereas wind direction in really tight agricultural [00:37:00] country doesn't affect the ton. It doesn't, the deer don't have to move as far and they're in their shelter belt or their fence row.

And they're not gonna walk across that beat field or that soybean field and expose themselves just cuz the wind has changed. So again, that, that's what I'm getting at, what the neural network is. I'm able to take this data from all over the. And all of those factors I just talked about are fed into these algorithms and then the network can predict the most accurately based on where it is in the us.

Gotcha. 

[00:37:27] Dan Johnson: All right. And you got the predictive deer movement portion of this. And it sounds to me like that's just continuous, always learning, always improving model, that there's no end to it. It's just a continuous, trying to make everything better. Now, when it comes to mapping, right? . And you said the guy said at three times in a row, so you thought he's wrong.

I'm gonna, I'm gonna do it myself. And so you started working with a mapping and it sounds to me like, from what I understand, . Getting things like [00:38:00] landowner data or satellite imagery and things like that is fairly easy to do. What makes then Spartan forges maps stand apart from maybe the other apps that are currently on the market?

[00:38:16] Bill Thompson: Sure. So there's a couple of things there. The first I think is just the presentation and the user interface and the user experie. We are trying to allow hunters, we have five to 15 centimeter imagery for about 45% of the US and what that means is if you have a five me, the easiest way to think about it is if you have a five centimeter object on the ground and you fall into that coverage area, you should be able to see it through our mapping.

That's, hundreds of times better than. What is currently on the market with one to three meter imagery. And we just made a value proposition and investment from the beginning to invest in that type of stuff and work with these companies from the very beginning. Whereas it might be cost prohibitive for a very large company to enter into that agreement especially up front because of the amount of user base that they have is really what these [00:39:00] companies are interested.

But also we provide them data services back in other ways that are proprietary. But essentially we have these relationships with these companies, but then we also present different maps from different times of the year. And we also go historical. So on that data we go, we can go back 14 years on most of it.

Some of it is 10, and then a very small subset of it is seven years. So you can go back and look at how the landscape has changed over time. Again, in that high re. Form factor, but then we present other three other data layers that are one to three imagery through one to three meter imagery layers through throughout many times of the year.

So no matter where you are in the US, you're gonna have some good maps. And it's not just gonna be one map and you get what you get. There's gonna be these different types of ones, but then also the way that people can interact with them and set up their maps. And for instance, inside of our app mapping, applic.

if you want to look at if you wanna set your map up so that you have slope angle shading on Lidar [00:40:00] with property over it, you can make that custom map in what we call our lambda layer. But then if you wanna switch right back to an aerial, it's just one swipe of the thumb. Whereas another application that's gonna require, between five and seven clicks, and in some instances 15 clicks.

So you're spending a lot of time when you're in the. Changing things up and doing things, whereas in our application, it's a very quick movement. And that really comes from my military background in, in understanding that from a design perspective, I want people to have to physically interact with the application as little as possible in order to get what they need so that they can focus on what's around them and not having their nose in the phone the whole time they're in the.

So that usability and the way that people interact with maps and set them up but then also what we're gleaning and learning from the mapping. And for instance here shortly, we'll have out this one meter lidar imagery for most of the us. And it's very high quality lidar imagery where you can see in some, in instances you can see cow trails on the ground, [00:41:00] or you can see benches on the side of hills where you'd never be able to see it with top up or anything.

To really inform scouting or, even where there's like a seep on the side of a hill. This is stuff that people generally don't have access to or haven't seen in the past, and we've made a large investment in getting that out there. So again, it's not just the data or the access to data, it's how it's presented and then how we show the hunter that this, benefits them from a.

targeting and hunting perspective. 

[00:41:25] Dan Johnson: Gotcha. All right, so my next question then revolves around the data that you actually collect from the user that on how they're using it. So let's just compare me and you, for example you. may use an app and you may do all the predictive deer movement and you may check your weather and you may still drop pins and then look at the landowner data and do all that stuff.

Whereas I may go into it and use it and just go, Hey I'm only using this just to drop pins. And that's it. How do you use that information on Take the [00:42:00] next step into the best usability for the end user or to change or update the app based off that information. 

[00:42:11] Bill Thompson: So I, we don't take a lot of user data and look at that.

So what I'm looking at aggregated data to understand like where I might pay for more updated imagery because like the imagery that I talked about before is very so I'm, I'll look at user distribution and see where my largest user bases are, and then add imagery for those user bases first. That's the really high quality stuff.

I'll still give all of the other the rest of the United States, all the other imagery, repositories. But then I'm also looking at what pieces are being used the most or used the. and then interacting with profe, like the pro staff, the guys that we work with understand, hey, I don't, I see people aren't using journal, the journal in the application as often as they're using historical imagery or vice versa, and then try to pull the string on that and understand why.

But again, the way [00:43:00] that we architected this application is if you are just the guy that just wants to see where your property lines are and you just wanna drop. , like that's all there and the weather is there and you can just use it in that really simple fashion. But then if people go on like our YouTube and watch some of the videos, which I encourage people to do when they sign up for the application you can get really deep and really, into the details of the data where you can understand, what's the predominant win during the second week of November in Kansas.

In this part of Kansas, so you get really deep into it. But we try to build the application. So it can be as simple as you need it to be, or it can be as complicated as you need it to be, and you can dive as deep into the data as you need to be. But we're constantly looking at that to see what's being interactive with what's being used, and then understanding from the user where they.

I see improvement, and I think that's another thing that separates us from the, from other companies, is that if you message us on social media or ask a question about Deere data or about a feature or a feature you like or you [00:44:00] don't like or whatever my marketing guys will be the first ones to see it, but then they'll direct it to me and I'll be the ones who are answering those.

So I spend almost two hours a morning just interacting with users to pull the string on the second order of that data. People are interacting with me and I'm trying to understand exactly what it is they want to see for the user, whereas I don't think you see that in a lot of other companies where, you know, the ceo and the lead architect of the solution is interacting with guys at the tactical level to really understand what they need.

So that's just a kind of a cross section of how we do all of that. Yeah. 

[00:44:35] Dan Johnson: Awesome. All right. You guys let's just do this. Why don't you tell us what's new on the app maybe over the last six months what you guys have introduced, any changes or updates to the app? 

[00:44:48] Bill Thompson: Yeah, so we introduced a web app and that's gonna be updated here shortly.

I've shied away from doing dates now because there are things that I don't control, like the app store and approval and that type of, So [00:45:00] we're trying to release it here very soon, but we've released a web app. You can transfer your points on there from other applications if you want, but then you can also just have a different display to look at the mapping, to look at the areas to look at, to do all of the planning and stuff.

If you're at work or whatever and you got a lunch break and you want to hop on and look at the maps, you can do that and drop points, and that syncs up with your application. But then secondly, We've released some, slope angle shading, which again we were very meticulous and I just encourage people to look at it and like our competitors, the way that we built that and architected it, makes it extremely accurate in a way that I don't think other people have thought about.

And we've we've updated our five to 15 centimeter imagery. We have On our custom map layer, we've updated the way that people interact with that and can set up custom maps and then be able to, as I said before reference those more simplified maps. So it's like you can either click the maps button or you can swipe through, but they're basically four maps or four layers.

There's a top of layer, a lighter, or a [00:46:00] aerial layer, a hybrid layer, and then your custom layer, and you can switch between those very quickly. We have our intel tab that we're constantly updating that gives you, pal palatability for forage that deer feed on the area the buck to do ratios for county by county.

The state draw odds, the state. The state hunter population, the distribution popular tracks of land. That's all stuff that we've added in the last year. And then going into this off season, we're adding that lidar data that I talked about before, which is for about 65% of the US has one meter resolution.

We added we have a functionality called Blue Force Tracker that will also be up this. Which is essentially, say you and your son are hunting the same land together. You can hi him and yourself, download the Spartan for jab. And then you click your property and then you say add a Blue Force Tracker team.

You put your son's email in there, and now anytime you're on the property together, your autos sharing location and your autos sharing points. , and you can do that with other buddies if you're scouting at [00:47:00] large tracks of public land together. Or if you're granting access to somebody else to hunt your property, you can tell 'em, Hey, when you're on the property, I'm gonna invite you to this Blue Force Tracker team.

So I know when you're in my backyard hunting, so I don't send my kids back there. Yeah. That will be out this spring. And then we've also partnered with Eastmans and their tag hub data. It's gonna be coming out in the application here very. Where you'll be able to look at state by state, draw odds and count down to the county level and understand your likelihood of where and when you'll, you can draw a tag.

And then we are gonna be coming out with some neural networks and algorithms in the future that will help optimize for people that are trying to do tag draws throughout the us. And then we're finally trying to put out some of our core wear. We were trying to get those out late last year, but we just realized some problems with the data that we've been fixing over the spring and the summer.

So essentially what that'll be is just new ways to look at topography and things to enter that influence deer movement. It starts with this [00:48:00] lidar layer that I'm talking about, but then it'll eventually get to the point where you can just highlight a piece of ground and the neural network will recommend places that you should scout.

That those things are all coming out here. And we're hoping to have, or no, we will have, we will button the product down by probably August. So the application that you have going into the hunting season will be the application that you have throughout the year. We've only been out for a year, so last year we had to do in season updates, but we're gonna stay away from that this year.

And to just make sure people have the final product once the season. And we've got a couple other trucks up our sleeve for stuff that we're gonna get out before now and then, but that's the brunt of the offering. Okay. And we're also, at the $39 level right now. Yeah. We'll probably end up having to raise it five or $10 after we put some of these other services that I've talked about out.

But for people who sign up now it's, 39 bucks with a 20% off discount, it's 32 bucks for the year. And, but then we also have a free application that gives you free property data and. So if you just want to use the free version of the application, you'll get free [00:49:00] property data on there.

You shouldn't have to pay for that. You've already paid for it when you paid your taxes. So that's in the free app, and you can drop pins in there and use it that way if you want. 

[00:49:07] Dan Johnson: Gotcha. Cool, man. And so based off of everything you've said in this interview, it sounds to me like you're, you are trying to make Spartan Forge almost a one stop shop app for everything.

Yeah, absolutely. 

[00:49:26] Bill Thompson: And we don't want out, we don't wanna overwhelm the user by putting too much in there. But essentially, when I was planning, when I was a white pill hunter, which I don't have time for anymore, I'm hoping to get back to it here one day, but I would be consulting four or five, six apps while I was doing my planning.

My goal is to put that all into a one-stop shop and put all that data in the same place and push the technological edge and capacity. Of what data can do for hunters and put it all into place that looks nice, is easily usable and highly customizable. 

[00:49:56] Dan Johnson: Awesome. I tell you what, bill, man, I really appreciate you taking [00:50:00] time outta your day to hop on and bs with us today and give us an update on Spartan Forge.

It sounds like not only are you busy, but you got a good thing going over there, so thanks again. 

[00:50:10] Bill Thompson: Yeah. Thank you for your time, Dan. Have a good one.

‍