The Project Management Problem Nobody Talks About... | Gary Sutton | DSH #2013

Summary

Digital Social Hour – Episode Summary

Episode: The Project Management Problem Nobody Talks About... | Gary Sutton | DSH #2013
Host: Sean Kelly
Guest: Gary Sutton
Date: June 11, 2026

---

Overview

This episode takes a deep dive into project management failures, misunderstood sports analytics, and the evolution and socioeconomic impact of artificial intelligence. Gary Sutton—business intelligence leader, author, and data science expert—joins host Sean Kelly for his first in-person podcast, engaging in an unfiltered dialogue about hidden problems in project management, myths in NBA analytics, and the future of AI in work and society. Sutton challenges conventional wisdom and offers thought-provoking insights rooted in data, while Sean keeps the conversation punchy and relatable.

---

Guest Introduction & Background

[00:42–02:27]

• Gary’s Bio: Business intelligence and analytics leader, author of two books ("Statistics Slam Dunk" and a second focused on data science in Python), and working on a third about classic algorithms behind modern AI.
• Unique Approach: Gary’s books use real NBA data, testing hypotheses in each chapter (rather than the typical “teaching one method per chapter” structure), making each chapter a standalone project with findings that, although based on NBA data, are cross-functional.
• Quote:

  “Each chapter really is a standalone project... In the course of teaching these different quantitative methods, I reveal some interesting and maybe some controversial insights about the NBA.” (Gary, 01:35)

---

The Realities of Sports Analytics

[02:27–09:57]

• Role of Analytics in Pro Sports: Every big league team now staffs data experts; analytics play a significant role, though Gary cautions that they shouldn’t be the sole decision driver.
• NBA Draft Analysis:
  • First-round draft picks fall into three distinct clusters:
    1. Picks 1–5: Much higher average win shares, and the only reliable source of future superstars.
    2. Picks 6–10: Significantly less productive than top 5, but better than the lower first round.
    3. Picks 11–30: Rarely produce stars—standouts like Curry (#7) and Parker (late 1st round) are exceptions.
  • Quote:

    “Most future superstars are drafted in the top five. That doesn’t mean every top five pick is a future superstar, but future superstars come from the very top.” (Gary, 07:10)

• Tanking & Incentives: Data confirms teams are incentivized to lose to get top draft picks—“tanking” does statistically improve chances, even after lottery tweaks.
• Spending & Success: Higher payrolls correlate with more wins and more championships, with some exceptions. Salary caps were intended to create parity, but according to Gary’s data, they didn’t work.
  • Quote:

    “Teams that do spend more on player salaries do win more… and I think it’s becoming even more pronounced in recent years.” (Gary, 12:24)

---

Debunking Beliefs: Defense, Hot Hand, and Home Court

[17:54–31:56]

Defense Wins Championships – Or Does It?

• Gary dismisses the idea that “defense wins championships,” arguing the data shows winning is about balance between offense and defense.
  • Quote:

    “It’s a cognitive illusion… The most successful teams score more points than they give up. You can be not very good on defense, but exceptional on offense, and win a lot of games.” (Gary, 18:05)

The Hot Hand Myth

• Traditional wisdom suggests players on a streak are “hot.” Gary tested this using free throws (comparable, controlled attempts) and found streaks are more like coin flips (independent events), not dependent.
  • Quote:

    “To me, they look more like coin flips… It looks more random than purposeful.” (Gary, 26:41 / 27:50)

Home Court Advantage & Referee Bias

• Teams at home are called for fewer fouls and attempt more free throws, which is statistically significant and not random. Gary attributes it partly to subconscious referee bias influenced by the crowd—even during pandemic games at neutral sites, “home teams” got the advantage.
  • Quote:

    “Home teams aren’t called for as many fouls… and they get to shoot more free throws. The probability that those variances are random is really, really low.” (Gary, 29:25)

• The effect is likely due to atmosphere, with referees subconsciously affected by home crowds.

---

Wisdom of the Crowds: Sports Betting

[32:38–36:32]

• Gary researched whether oddsmakers or bettors (“the crowd”) are better at predicting outcomes. Collectively, the closing betting odds (reflecting aggregated bets) outperform the opening lines set by bookmakers, especially earlier in the season.
  • Quote:

    “The gamblers collectively are smarter than the oddsmakers… The closing odds, more often than not, are closer to the actual results.” (Gary, 34:28)

• He likens this to “Who Wants to Be a Millionaire?”—the audience’s aggregate answers are usually right.

---

LeBron vs. Jordan: What Does the Data Say?

[36:32–38:47]

• While longevity and stats say LeBron has likely had the better career, Gary and Sean agree that stats alone aren’t enough; qualities like leadership and making teammates better are not fully captured by current analytics.
  • Quote:

    “Win shares is nice, but it doesn’t measure—are you making players better? Analytics aren’t necessarily going to capture that... That’s why the data by itself can be misleading.” (Gary, 37:56)

---

AI, Algorithms, and the Future of Work

[38:47–54:45]

Gary’s Books and AI Insights

• Gary’s third book (in progress) highlights how current AI uses “timeless” algorithms—many invented decades or even centuries ago—but achieves today’s performance only through engineering breakthroughs (processing power and data scale).
  • Quote:

    “AI is empowered by algorithms developed many, many years ago… The difference now is the scale, thanks to more processing power.” (Gary, 39:50)

• OpenAI’s key insight: bigger datasets + more computing = better results.

Choosing Between LLMs and AI Outputs

• Gary uses multiple LLMs (e.g., GPT, Claude) and often cross-checks responses, noting each has different strengths/weaknesses because of differing training data.
• Sean shares a story where five AI models gave him five different diagnoses for his dog’s X-ray, underscoring the need for human judgment.
  • Quote:

    “You can’t just blindly follow AI… there still has to be some human element.” (Sean, 45:17 / Gary, 47:29)

• Gary frames it as “knowledge versus wisdom”—AI supplies information, but human understanding and judgment are still essential, especially for troubleshooting models in production.

Will AI Take All Our Jobs?

• Gary references ongoing debates:
  • Marx predicted permanent job loss from automation.
  • Schumpeter (“creative destruction”) suggests economies recover and create new types of jobs after disruptive innovations.
• Gary is cautiously optimistic: there will be job displacement and a “gap” period, but history suggests economies eventually adapt and create new roles combining old skills and new technology.
  • Quote:

    “We’re in one of these creative destruction phases… It will eventually end, and companies will figure out the new jobs we need.” (Gary, 54:02)

---

Project Management: The Real Problem Nobody Talks About

[55:03–60:40]

The Hard Truth

• Project Management Misunderstood:
  • Many organizations use project management as a “fallback role,” assigning unqualified staff just because they’re neither great analysts nor programmers.
  • There is an overemphasis on soft skills; the industry underestimates the need for quantitative, analytical project management skills.
  • Result: Most complex projects fail (“complete clusterfucks”) due to poor project management rather than execution.
  • Quote:

    “Projects go off the rails more often than not because of bad project management… Organizations underestimate the skill set required.” (Gary, 56:26 / 58:18)

• Solution:
  • Project management should use quantitative techniques (Monte Carlo simulations, Markov chains, probabilistic modeling) to properly plan, sequence, and execute projects—much like other data science methods that manage uncertainty.

---

Notable Quotes & Memorable Moments

• On NBA Draft/Tanking:
  “If you want any reasonable shot at drafting a potential superstar, you need to draft in the first five… The data says [tanking] makes sense.” (Gary, 06:34 / 08:16)

• On Sports Spending:
  “There is benefit in spending money… On average, championship teams spend more on player salaries than other teams that qualify for playoffs.” (Gary, 10:12 / 11:38)

• On AI's Foundations:
  “Modern AI is not brain power—it’s processing power, running old brainpower at a much larger scale.” (Gary, 43:02)

• On Human vs. AI Judgment:
  “Knowledge versus wisdom. Language models provide the knowledge, but you still need human wisdom to diagnose.” (Gary, 45:32)

• On Economic ‘Creative Destruction’:
  “Technology allows for creation of new jobs, sometimes new industries—but you end up having this gap.” (Gary, 52:11)

• On Project Management:
  “They focus on the soft skills… but project management also requires hard skills—just like any other role.” (Gary, 56:26)

---

Key Segment Timestamps

• [00:42] – Gary’s intro and approach to teaching statistics/data science
• [05:44] – NBA draft clusters, superstars, tanking
• [10:59] – Payroll, salary cap, and parity
• [17:54] – Defense, offense, and NBA success
• [21:08] – Project management roots in quantitative methods
• [24:14] – The "hot hand" fallacy
• [29:25] – Home court advantage and referee bias
• [34:28] – “Wisdom of crowds” in sports betting
• [36:32] – LeBron vs. Jordan: the analytic divide
• [39:50] – “Timeless Algorithms” powering today’s AI
• [47:29] – AI knowledge vs. human wisdom
• [51:43] – Creative destruction in economic cycles
• [56:26] – The true skills gap in project management

---

Takeaways

1. Analytics aren’t absolute: Data can guide, but not fully define, winners—whether in sports or in business.
2. Project management is often the weakest (and least appreciated) link: Most projects fail not due to technical tasks, but incompetent or insufficiently skilled project managers.
3. Beware of accepted wisdom: Classic phrases (“defense wins championships,” “the hot hand”) are more myth than fact, as shown by Gary’s data-centric approach.
4. AI will disrupt—but probably not destroy—jobs: While short-term displacement is likely, economic history suggests new roles will arise.
5. Human insight remains crucial: In both evaluating players and interpreting AI output, wisdom and context matter as much as statistics and predictions.

---

Links and References

• Gary Sutton: Author page, LinkedIn (linked in show notes)
• Books: “Statistics Slam Dunk”; second data science book (Python); upcoming “Timeless Algorithms”
• Upcoming Event: R programming/data science conference in Poland

---

For listeners who want contrarian perspectives on sports, analytics, and AI—and especially for those in project management—Gary and Sean’s conversation is a must-hear (or, with this summary, a must-read).

Transcript

[A]

James Hart. If you looked strictly at his stats or looked at it from an analytical standpoint, he would stand out, I'm sure. But I'm not convinced that he's really a winner. Even seasoned project managers don't know what they are. And the companies that are hiring project managers and assigning them to projects, they don't know them either. And so you've got projects that are, like I said, they're complete cluster.

[B]

Okay, guys, we got Gary on the show today. I believe it's his first in person podcast, right?

[A]

It is.

[B]

Wow.

[A]

Appreciate it.

[B]

Thanks for choosing. Yeah, thanks for choosing my show. I been looking at your stuff and you talk about some interesting things, man. So thanks.

[A]

No, thank you.

[B]

We're gonna dive into it. For those that don't, that don't know who you are, could you give a quick bio on you, what you've done, what you're working on?

[A]

Yeah, I've been for many years a business intelligence and analytics leader and I'm also the author of a pair of data science and statistics books. I'm working on my third book right now. My third book is about the older algorithms that are now powering modern AI. So that will be released later this year. But I've also written a couple of other books, again in the data science space. My first one is titled Statistics Slam Dunk. So it's very different from from other manuals, at least in a couple of meaningful ways. Number one, most books tend to dedicate a single chapter toward teaching a single method. So there's a one to one relationship between chapters and techniques or methods that are being taught. My book instead takes a hypothesis and it tests that hypothesis and any method or technique that's needed to then confirm or deny that hypothesis is taught within that chapter. So each chapter really is a standalone project. So that's number one. Number two, as you and I were discussing in the green room, I work with MBA data sets throughout, whereas most books work with stale or synthetic data. So when you consider, okay, I'm testing a hypothesis in each chapter, I'm using real MBA data. In the course of teaching these different quantitative methods, I reveal some interesting and maybe some controversial insights about the mba. The methods that I teach are cross functional. They don't apply just to MBA data. You can use those in any domain. And for most data use case. But the findings are NBA specific.

[B]

Yeah, yeah, we're going to dive into some of those. Some of those are very interesting. And now you're seeing a lot of NBA teams have job roles just for Data guys.

[A]

Oh, yeah.

[B]

Each team has like a statistics expert. They're telling the players how many threes to shoot, like, what's the best shot to take. It's a big part of professional sports now, right?

[A]

Oh, yeah, yeah. NBA, the NFL. Absolutely. Huge. Major League Baseball. I think it's starting to make a bigger impact in the other sports that aren't the big three. Right.

[B]

It's crazy. Like, they're using nerdy guys to win professional sports.

[A]

Yeah, it's usually guys younger than me. They're more like your age, but the nerds. But yeah, you're absolutely right. And I'm a big analytics guy, obviously, but I don't believe in using analytics alone to assess players or other situations.

[B]

It's part of it, but not everything.

[A]

Correct. Because now it's my view.

[B]

Yeah, no, that's a good.

[A]

That's my view.

[B]

It's a valid view because now in the NBA playoffs, what's working in the regular season with shooting a bunch of threes isn't working in the playoffs.

[A]

I think there are some players. Okay. I'm going to diss on James Harden.

[B]

Okay.

[A]

Okay. I don't know what you think about him. I don't care for.

[B]

I'm not the biggest fan.

[A]

Okay. If. If you looked strictly at his stats or looked at it from an analytical standpoint, he would stand out, I'm sure. Yeah. But I don't. I personally don't care for his game. We'll see if I'm proven wrong this year. I don't think he's ever going to win a title. I know he's in the Eastern Conference finals, but I'm not convinced that he's really a winner.

[B]

I don't think he can win a title as the main guy, as the number one option.

[A]

That's a fair point. That's a fair point. He's got some good players around him now in Cleveland.

[B]

Donovan Mitchell.

[A]

Yeah, yeah. And they've got a. They've got an absolutely legitimate chance of beating the Knicks. Of course. I think. I think either of those teams will. Would have a challenging situation in the Finals against whoever comes out of the West.

[B]

But I think the west matchup right now is pretty much the Finals, in my opinion.

[A]

Yeah, I agree.

[B]

That game last night was crazy. Were you watching?

[A]

I did not get a chance to double last night's game. Yeah. Okay.

[B]

But anyways, going back to the stats of stuff, so what actually matters with NBA statistics?

[A]

Yeah, I looked at it more from. I'm not so much into the player analyses. I'M more interested in the, I guess the big picture stuff like I looked at, at the draft and, and specifically first round draft picks where they're drafted or selected in the first round and then how do their careers map out once they're picked? So does it really matter? Like, I think if you looked at the NFL, Sean, there's a lot of variation in terms of what you might expect a first round pick to have. As far as a productive career in the NBA, the problem is much more acute. You've got picks that. Well, let me take a step back. When I did my analysis, I would divide the first round draft picks into three very distinct groups or clusters. You've got draft picks or players selected in the top five. And so I'm sure you're familiar with the metric of win shares.

[B]

Yeah.

[A]

Okay. So when you average out the number of win shares for players selected between one and five, that number is much, much greater than any other first round draft pick. Oh really? Oh, yeah. Like two to three times at least higher.

[B]

Wow.

[A]

Than what you otherwise get in the first round or should otherwise expect. So that's one group or cluster. The second group or cluster would be draft picks six through 10. So those players on average have careers that are much less than what you would expect from a player selected between one and five, but otherwise much greater than any other first round pick. So when you think about draft picks 11 through 30, that's your third group or cluster.

[B]

Got it.

[A]

So that's fascinating in the sense of, okay, if you want a best chance at drafting a superstar, if you're a team, Sean, that's trying to rebuild your roster through the draft, you need to get to the top of the draft because you need to draft in the first five if you want any reasonable shot of drafting a potential superstar. So very few players who end up having outstanding careers are selected outside of the top five. Very few. Steph Curry was a seven pick and Tony Parker was drafted. He had a great career with San Antonio. He was drafted fairly late in the first round if I remember correctly. But those are the exceptions, not the rule. So most future superstars are drafted in the top five. That doesn't mean that every top five pick is a future superstar. But future superstars come from the very top of the draft.

[B]

Yeah. Because there's been a lot of bust too.

[A]

Oh yeah. Oh, absolutely, absolutely.

[B]

Probably more of them are bust compared to superstars, I'd imagine, right? Statistically?

[A]

Oh yeah, absolutely, absolutely. So if you're a team that wants to rebuild your roster through the draft, you need to get to the top of the draft. So how are you going to do that in today's system? Even with the lottery? And the lottery has been out there for many years now, it's been tweaked a couple times. You need to have a woeful one loss regular season record. Okay. So it's not a truly competitive system where you're rewarded for success and vice versa. You're rewarded by having a poor one loss record.

[B]

Tanking. Yeah.

[A]

Yeah. So is, is there an incentive to tank and does it make sense to tank and, and the data says that it does.

[B]

Oh really? Even with the new changes? I know they change it a couple times.

[A]

Absolutely.

[B]

Because I think the lottery this year, the team that tanked was the Kings, I believe, and they didn't get the first pick, they got like the seventh pick or something.

[A]

But you would not otherwise have any opportunity or any chance unless you had a poor record to begin with. You're right. With the lottery, you're not guaranteed if you have the worst record. Back in the day, you were guaranteed to have the very first pick. Right. With the lottery, you're not guaranteed to have the first pick, but you otherwise don't have any other chance.

[B]

It's better than nothing, right?

[A]

It's better than nothing.

[B]

The chance of getting that pick is better than no chance.

[A]

Absolutely. And even Mark Cuban was recently defending tanking and saying that, well, it's really not that bad for the league. You know, tanking has been. I don't know if it's a problem or it's just been an issue, but it's been an issue or a problem, however you want to define it for many, many years. What suggests the NBA hasn't resolved it. But yeah, there's. Based on the data that I looked at and evaluated, there's an incredible incentive to tank. If you want to build your roster through the draft, you can go the free agency route. There's. There's other ways to build up your team. But if, if the draft is the way you want to go, you need to get to the top of the draft. And the only way to do that is to have a losing record, a really, really bad one loss record.

[B]

Right.

[A]

So you need to lose games, whether that's purposely or not. So purposely. But that's your only chance to get to the top of the draft.

[B]

It also seems like big market cities seem to be better. I don't know if you've seen any stats with paying players more versus their win record. Do you notice that because, like, LA seems to always be good. New York, big cities with big money.

[A]

I. I did. I did some analysis on that. There's. There's. So. So there is. There is benefit in spending money.

[B]

You know, I've never promoted a casino ever. But I'm in Vegas and I finally met a team I trust. Legends. Legends with a Z. Established brand, real reputation. They pay out, and they pay out quick. Sports slots, original games, live tables, all in one spot. And you get 100% match on your first purchase. They're live in 39 states in the U.S. texas, Georgia, Florida. @legends.com. rest of the world, you could play at Legends IO. Click the link below. Use code DSH. Check out the link below.

[A]

When I analyzed team payrolls and bumped that up against how well they performed, whether they qualified for the postseason, whether they actually won a championship, there's a little bit of a difference between teams that won a title versus other teams that qualified for postseason play. So on average, championship teams do spend more on player salaries than other teams that otherwise qualified for the playoffs. But there's a huge difference between teams that qualify for postseason play and teams that didn't. So it definitely pays to pay. Right.

[B]

Like the Yankees. I feel like they spend a lot, basically.

[A]

And there are some exceptions. I mean, look at the Knicks. We were talking about the Knicks just a moment ago. Okay, so they're an exception. I mean, for many years, they spent a lot of money, player salaries, and they haven't won a title since.71. Yeah, okay, so that might change this year. We'll see.

[B]

We'll see.

[A]

But. But they're an exception. Or you look at a team like San Antonio on the other. On the other side, they've been relatively frugal, but under Popovic, they were, year in and year out, one of the top teams in the league, and they won several titles under him. So again, there are exceptions to every rule. Those are the exceptions here. But I'll go back to what I said a moment ago. Teams that do spend more on player salaries do win more.

[B]

Okay, so overall, there's a trend.

[A]

Oh, absolutely. And I think it's becoming even more pronounced in more recent years than it was.

[B]

Right.

[A]

Maybe 10 years ago.

[B]

That's why they had to introduce the salary cap. Right. Because teams were outspending other teams.

[A]

Yeah. Okay, so fair question. So the NBA instituted a salary cap just before the 84, 85 season. So it's been in play for what, 40 years now. Right. And so the league had to go to the Players association to get their agreement to put a cap on player salaries. And the argument that the league made back then was, yeah, we're going to suppress wages to some extent, of course, but we think that by capping salaries, it will produce parity across the league and that's better for the long term benefit of the league. So take a hit here. But longer term, the league will be much better off. And so of course the players association agreed to that. The problem is that that's complete bullshit. Okay, so when you look at the data, there really wasn't a parity problem to begin with. Okay? So in my first book, I demonstrate different ways to go about measuring dispersion in data. So think of it, Sean. Think of a data set as the number of regular season wins for each NBA team by season. Okay, got it. And if you have a lot of teams that are, let's say right around 41 wins, right? So they're at 500 or just above or below 500, you have very low variance in your data and that equates to high parity, at least from an intra season parity perspective. And conversely, if you have a few teams that finish the season with, let's say, around 60 wins, and then conversely, you have other teams that finish the same season with around 20, 25 win, you have high variance and low parity. So from an interest season standpoint, when you look at the data and plot it out by year, before the 84, 85 season, when the NBA put the salary cap in place, you really don't see an interest season parity problem. Parity did drop for the last, I would say three seasons before the salary cap. So maybe in defense of the league, they were looking at those three data points and came to the conclusion that, oh, we have a potential problem here that we need to solve, but three data points don't make a trend. Three data points is more noise than anything else. So when you pan out and look at the data more holistically, when you look at parity, interest season parity by year leading up to the salary cap, and you look at the same data after the salary cap, interest season parity actually got noticeably worse after the salary cap was implemented, regardless of how you want to measure it. So the data otherwise looks the same. So I don't think the NBA, at least from an intra season standpoint, intra season parity, I don't think the NBA really had a problem to begin with. And if they did, they definitely didn't solve it by implementing a salary.

[B]

Interesting. They didn't have all this data back then, I'd imagine to think about, or

[A]

like I said a moment ago, they might have been looking at those last few data points and trying to head off a problem, which again, that's not how I would have approached it. And if I were part of the players association, I wouldn't have gone along with it. If that was your argument and you've got just a small number of data points and you're thinking we have a problem here. No, take a step back and look at the data more holistically and see what the real trend is. But that's from an intra season parity perspective. You also have intra season, right. Where if you have high inter season parity, you have churn at the top of the league year over year. Right. So let's say you've got your top eight teams for one season. Well, if you really have inter season parity the next year, you should see a lot of churn at the top. So the same eight teams that were at the top of the league this year was. Would not be the same eight teams at the top of the NBA the next season. Right, but that didn't fundamentally change. There was always some churn.

[B]

Yeah, it's in a top few.

[A]

Yeah, a top few. But even before the cap was implemented there was some churn. After the salary cap was implemented, there was the same amount of year over year churn. So that didn't really change either. So I'm not exactly sure what the NBA was looking at. I'm not exactly sure why the players association would have reported prove the cap given the data that was then available. But I don't think there was a problem to begin with. I don't really think the NBA had a parity problem. I know we think that, okay, the Celtics or the Lakers are winning most of the titles, but. But in the whole scheme of things, I don't really think there was a problem.

[B]

Well, back when they won a lot of them, there were less teams also.

[A]

There were. There were. And when you add more teams, maybe temporarily there's an opportunity for less parity because newer teams might struggle at first. Yeah, I tried to account for that in my analysis, but it's hard to

[B]

even come up with a number for that. Right.

[A]

Yeah. I still don't think that that was material in any way.

[B]

Now, according to just stats and numbers, we don't have to piss anyone off. But does defense actually win championships? Because that's like a classic thing you hear all the time. Does it actually win championships though?

[A]

No, I think It's a cognitive illusion, like the hot hand and like the idea that games are won in the fourth quarter. Those are also, in my opinion, cognitive illusions. So you tell me what you think, but I think basketball is basically a 50, 50 game between offense and defense. It might not be exactly 50, 50, but it's close.

[B]

Yeah.

[A]

And I think the idea, well, the idea that defense wins championships actually goes back. Think Bear Bryant, the legendary football coach at Alabama. He was the one who, who, who made that case, and it transcended from football to basketball, maybe even other sports. But I think where it comes from is when you look at, when you look at postseason games versus regular season games, postseason games historically and typically are lower scoring games than what you see in the regular season. So teams that win and lose are both. They're scoring more points and they're giving up fewer points. So if you're a winning team in the post season, you're. You're advancing to the next round of the playoffs. You're, you're, you're looking at the scores and you're comparing those scores to what, a typical score, what might have been of a regular season game, and you're thinking, wow, we're scoring fewer points, but obviously we're giving up even fewer points and we're winning. So therefore, defense wins championships. I think that's maybe where it comes from or where people get the idea that defense wins championships. But when you do the analysis, it really goes back to the point I made a moment ago. It's really a 50, 50 game between offense and defense. And the most successful teams, I mean, this goes without saying, but the most successful teams score more points than they give up. And a team can be not very good on defense, but exceptional on offense. Right. So they're scoring more points per game than they're giving up. They can win a lot of games.

[B]

Yeah.

[A]

Right.

[B]

So the podcast receives a lot of sponsorship inquiries and people coming on the show. I needed to find a program that helps systematize and organize all these applications and inquiries. That's where Zapier comes into play. Zapier is how you actually deliver on your AI strategy, not just talk about it. With Zapier's AI orchestration platform, you can bring the power of AI to any workflow, so you could do more of what matters. You could connect top AI models like ChatGPT and Claude to the tools your team already uses, so you can add AI exactly where you need it, whether that's AI powered workflows, an autonomous agent, customer, Chatbot, or Something else. You can orchestrate it with Zapier. Zapier is for everyone, tech expert or not. Teams have already automated over 300 million AI tasks using Zapier. Join the millions of businesses transforming how they work with Zapier and AI. Get started for free by business using zapier.com SocialHour that's Z-A P-I-E-R.com SocialHour so

[A]

the recent Golden State warriors teams might fall into that bucket, for instance, and you can be very good defensively, but you're not good offensively. Right, Right. And so therefore, your defense isn't really helping you out. It's the lack of offense in that case. So I think it's. I think the best teams are balanced in some way between offense and defense.

[B]

Yeah. The Spurs, I feel like, are a good example of that.

[A]

Yep. Yep.

[B]

Victor's getting three, four blocks a game. Isn't that crazy?

[A]

That is.

[B]

That is nuts. That's like unheard of. But they also can score. They got so many options. I feel like they might win it this year.

[A]

Well, yeah, they've got a great opportunity. I mean, you and I were saying the winner, this year's champion, will probably come out of the west, and they've got a tremendous shot.

[B]

Yeah, we'll see what happens. Yeah. You said the hot hand is a myth. So, like, that's just an illusion or.

[A]

I think it is. This is very controversial. You know, when I. When I wrote my first book, it's funny, I was not going to assess anything on the hot hand because I didn't know what I could contribute. That was new. Right. So others have evaluated the hot hand, and some have said, like I'm saying, it's a cognitive illusion, others say otherwise. Players will tell you otherwise. Right. So I wasn't going to include anything around the hot hand in my first book. Again, just for the very simple reason that. Okay, well, it's been evaluated before. What can I do that's new? But then as I was writing the book, I got to thinking that many readers are going to expect to see something. Right. I mean, so, okay, so you're analyzing the NBA, but you're not addressing the hot hand. And so I found a way to add it in, in a way that I thought was new or different from what I had been exposed to. At least me, what I had been, what I had read previously. And, you know, most of the analyses focus on field goal attempts. And it's the idea that. It's the idea that whether you're shooting Field goals or free throws, these are dependent events so that when a player makes a shot, he increases the probability he's going to make his next attempt. Right. And then. And then, if you also are successful on that attempt, now, the probability of your next attempt is even higher than it would otherwise be. So the fundamental question is, are these dependent events? Right. Or alternatively, are they independent, like coin flips? Right. So if I flip a coin and it comes up heads, that doesn't influence what the outcome of the next coin flip is going to be. Right. But in my book, because others had looked at field goal attempts, which I always found kind of odd, because an open layup is very different from a contested three. Right, Right. And so how do you. How are you making judgments when the field goal attempts can be so different? Right.

[B]

Valid point. Because I'm a big man. So, like me going five for ten is way different from a guard shooting threes. Five for ten.

[A]

Right.

[B]

You know?

[A]

Right. So I looked at free throws. Every free throw is exactly the same. They're taken in a very controlled environment.

[B]

Yeah.

[A]

And I also looked at a probability theory from a French mathematician named Pierre Simon Laplace, and he called it the rule of succession. And it was based around very small sample sizes, which is what you get from a single player in a single game on free throw attempts. Right. Even a player like Sean Harden or James Harden. I'm sorry, you're Sean. You're probably better than him, actually, in many ways. I wish you're still talking about very, very small sample sizes. Right. Because the Hot Hand only applies to a single player within a single game. Right. It doesn't transcend across players or across games. And he. He made the. He. He plotted it out that if you had. If you truly had dependent events and one success led to another, which led to another, you would see basically a probability curve that comes very close to 100%. It's never 100%, because there's always some chance that you could have a miss or you could have some. Anything that's deemed a failure. And so when I looked at the data, okay, I figured if there really is a hot hand, and I plotted it out for players that took a fair number of free throws, like someone like James Harden, it should resemble a Laplace curve.

[B]

What's that?

[A]

It's the rule of succession, where your probability of making a shot is predicated on if you made the previous shot. So with each successive make, your probabilities are incrementally increasing, but at a diminishing rate.

[B]

Got it.

[A]

So If I plotted out free throw makes and misses, would it look like that curve? Not exactly. But would it resemble it? Or alternatively, would it look more like coin flips that are independent events? Right. And to me, they look more like coin flips.

[B]

Free throws.

[A]

Oh, yeah. Oh, yeah. Even, you know, if you were to flip a coin 10 times, a fair coin, same coin 10 times, we know that there's always a 50% chance it's going to come up heads, 50% chance they'll come up tails. That doesn't mean that with 10 flips you'll always get heads five times or tails five times doesn't mean you'll get heads once, then tails, then heads. You might even get heads, let's say six times in a row. So there are streaks. Even with random independent processes, there are streaks. That's how free throw shooting to me looks. It looks more random and independent with periodic streaks that someone might mistake for a hot hand.

[B]

That makes sense.

[A]

Yeah. Yeah.

[B]

It's easy to say hot hand in the moment when they make like four or five in a row.

[A]

Right, right, right. Those who support the hot hand idea seem to believe that there's a dependent relationship when shots are made, but not necessarily when shots are not facts.

[B]

That's so true because James Harden went like 2 for 16 the other day, you know?

[A]

Right. And he may. And he may regress to the mean and have a really good shooting night the next time out. Right. So. So I don't see it. I. I see the data more looking like coin flips than. Than something that's more purposeful, I guess. So it looks more random to me than anything else.

[B]

Now, how much does home court advantage matter? Because you saw during the pandemic, the Lakers won the championship, but there was no home court advantage. So you start to wonder, does it actually matter a lot in the playoffs?

[A]

I looked at. Well, it's interesting that you mentioned when the Lakers last won, because, you know, the final games in the postseason were played at a neutral site. Right. But the NBA went out of its way to. There were designated visiting and home teams when those final games were played in Orlando. And the arena, wherever they were playing, was decked out. Colors. Yeah, the arena was decked out in the colors of the designated home team.

[B]

Got it. I didn't know that.

[A]

So they were trying to simulate a home game for the team that was designated the home team. I looked at foul calls and so personal foul calls and free throw attempts over several seasons, and what I discovered was home teams are called for fewer fouls. And home teams attempt more free throws as a result than visiting teams.

[B]

Wow.

[A]

And that holds up over the course of an NBA season. And there. I won't get into all the, all the details here, but there are statistical tests you can run to determine if that variance is meaningful or is it just random. Right. And when you run the tests, the probability that those variances are by randomness is really, really low. Really low. So home team. So I think part of the home court advantage is driven by the fact that home teams are called for fewer fouls and home teams do commit or do attempt more free throws. But the bigger part might be. And you know this because you play, foul calls can disrupt the flow of a game 100%. Right. So if you're a star player and you're called for your fourth personal foul early in the third period, you're going to sit. Yeah, that wasn't necessarily the plan. Right. The plan was you're going to play for most of the third period, but you're going to sit if you're called for your fourth foul early on. So I think that's a contributor to home court advantage is the fact that, as I've said, home teams aren't called for as many fouls as they're visiting opponents, and they get the opportunity to shoot more free throws. But yeah, you also leave open the possibility that the flow of. The flow of a game is disrupted. Right. And more impactful to the visiting team more often than otherwise.

[B]

Yeah. You see this. People complain about the Thunder a lot lately with SGA getting a lot of free throws. Do you think that's because the referee. I don't know how it works with the NBA, but. Are the referees local? Do they live in that city?

[A]

No, no. But no, they travel like the players do. The teams do.

[B]

So they're feeding off the crowd then?

[A]

That's what I think. I think they're affected by the crowd.

[B]

That makes sense.

[A]

I mean, the, the, the home, the home team's fans are right there.

[B]

Yeah. They're in your face.

[A]

They're right. Absolutely. So I think there's some impact human element.

[B]

That's why I wonder.

[A]

And it was interesting. Even during the COVID games in Orlando, the designated home team also shot more free throws. Really? Yes. Huh. Yeah.

[B]

So the referee has some sort of subconscious bias or something. Right.

[A]

They, they, they are somehow influenced or impacted or aware of their immediate surroundings.

[B]

That's why I wonder. And I know we're going to talk about AI if sports are going to head more towards AI judging because you're starting to see that creep up in certain sports.

[A]

That's interesting.

[B]

Right. So I don't think it'll completely replace the human referee, but maybe they'll start having a video. I think they already do it in the video review. Right. Somewhat.

[A]

Baseball is. You can challenge balls and strikes. Now, that's not AI Per se, but you're using technology to, you know, overrule plays.

[B]

Now, I mean, baseball, I could 100 see, because it. It is literally a box. So if they could just use AI to determine if the pitch was in the box or not, why do you need a human to do that or

[A]

who's touching the bag first, you know? Right. So, absolutely.

[B]

We'll see what the future looks like. Right. Let's see. Anything else with basketball? Oh, gambling's become a huge market. Right. Have you looked into the. The science with gambling, how these books are making their lines and all that?

[A]

Yeah, I took the angle of. Well, based off my findings, I took the angle that there's wisdom in crowds. Right. So full disclosure, I'm not a gambler, but there, you know, Vegas comes out with opening lines. Right. For every game. And. And then those lines move. Right. Based off who's betting where. Right. And what. And what I found was, so who's smarter? The odds makers, the people who are paid to put the odds down and balance the betting? Or is it the gamblers who have the real skin in the game because they're betting their own money?

[B]

I would assume the odds makers. But I'm curious what you say.

[A]

It's the gamblers, really. So it's the gamblers collectively, not a. So think of. I think of. Let's say you've got a gumball machine and you're asked to guess the number of gumballs in the gumball. Okay. Well, your, Your guess will be potentially way off. My guess will be potentially way off. And we can get others, let's say, on your team, Sean, to do likewise. They all might be off, but if you take the average, the average might be fairly accurate. So an individual gambler might not be smarter than the odds makers, but when you cancel out their errors and aggregate where they're betting, the gamblers collectively are smarter than the odds makers. So when you look at the opening odds versus the closing odds, the opening odds come from the odds makers. The closing odds are influenced by where the money is going. The closing odds, way more often, at least more often than not, are closer to the actual results.

[B]

Wow, that's so interesting to me because you'd assume These guys with their fancy algorithms and everything.

[A]

Oh, yeah.

[B]

Are better than.

[A]

Well, think of it early in the season, too. It's even more pronounced early in the season when you don't have a lot of data to work with. The data you really have is from the prior season, which doesn't necessarily apply.

[B]

Yeah. It's hypothetical data.

[A]

Right. So early in the year, October, November, when you don't have a whole lot of data points to otherwise work with, the gamblers are actually even more accurate than the odds makers. It evens out. Doesn't even out. But the margin becomes a little tighter as the season progresses. But still, at the end of the day, gamblers independently placing bets, but aggregating the results in a collective way, the gamblers know more than the odds.

[B]

Yeah. They're called sharps. Right. People that could beat the bookmakers. Yeah, I know a few of them.

[A]

Okay. Yeah.

[B]

Very rare. But like you said, if there's a lot of them out there, offsets the initial line.

[A]

So, I mean, even the smallest margin can make a huge difference. Right. When you're gambling.

[B]

Yeah.

[A]

So. So that was. That was fascinating. I thought. I went into it thinking that could be the case because going back to the gumball machine example, I appreciate the potential wisdom that you get when a crowd of people are allowed to independently make guesses and do other things. And when you aggregate those results, like the game show who Wants to Be a Millionaire? That's another example. Contestants were allowed to ask the audience when they. When they didn't necessarily know the answer to a question. Right. So they. They delegated the response to the audience. Well, a huge percentage of the time. So the audience is then making their guesses. Right. And everyone's doing that independently. They're not collaborating in any way and way. More often than not, the audience had the right answer.

[B]

Really?

[A]

Oh, yeah.

[B]

Just because the aggregate. Yep. That makes sense.

[A]

Absolutely.

[B]

One last question about basketball. I know you've probably been asked this a lot, but I know this will go viral. LeBron James or Michael Jordan? Purely off data. Who is the better player?

[A]

I think if you looked at data, it would probably suggest LeBron. I'm a Jordan guy. So it goes back to what we were talking about earlier. Right. It's not just the analytics, But I think LeBron just from a pure. And some of it is due to his longevity. Right. No doubt about that. But if you strictly looked at data and you didn't really know anything else, you just looked at the numbers, I think the Data suggests that LeBron has had a Better career. I wouldn't necessarily agree with that. I'm more of a Jordan guy than a LeBron guy. And that's why the data by itself could be misleading. As you were saying earlier. It's part of your analysis. It's not 100% of it, but I think the data would suggest LeBron better career.

[B]

I could see, but I mean, like, better player. I wonder if there's a way that, like, you know, each player has a plus minus.

[A]

Yep.

[B]

I wonder if there's a way that you could determine if someone's an actually better player overall. Because longevity wise. Yeah, LeBron's the goat for sure.

[A]

Yeah.

[B]

Played 23 years.

[A]

Right.

[B]

And counting.

[A]

There's so many other aspects to. To evaluating a player. Right. I mean, win shares is. Is a nice metric that encompasses a player's offensive and defensive contributions, but it doesn't necessarily measure well. Are you making players better? Right. Are you setting your teammates up to also be successful? And in that way are you contributing to their success, which in turn then contributes to your success and the team's success? Analytics aren't necessarily going to capture that, or at least we haven't figured out a way to do that adequately. I think if we get to that point and do some of that analysis, then I think Jordan would maybe come to the forefront versus LeBron. But I think the way we look at data now, I would say it's LeBron, but again, I wouldn't necessarily agree with that.

[B]

Yeah, itest. Right. That's another part of it. Let's move on to AI. And that's your third book. Right, Right.

[A]

So we talked about my first book. My second book is a little bit more, I guess, conventional or traditional, where I do teach different analytical methods fitting different types of models, and I teach other quantitative techniques and demonstrate how to implement those in Python. My first book was written in R. Second book in Python.

[B]

I never even heard of R. Okay. Is that older?

[A]

It's a little older than Python. It's a programming language that was really developed just for crunching numbers.

[B]

Got it.

[A]

Python's more of a general programming language. It can be used for data science and does some marvelous things. But you can also create a website in Python. You can also develop application interfaces in Python. R doesn't try to live in that space. It's really just for crunching numbers.

[B]

My dad was a programmer. I believe he did Python, but he was always saying, every year there's a new one. I was like, damn, you got to keep learning new languages.

[A]

Absolutely. Yeah. They do similar Things, but the syntax is very different. But yeah, so my second book is a little bit more conventional. I do teach different methods and again, how to, how to implement those same methods in Python. So my third book, which is still in development, but it will be released later this year, has a working title of Timeless Algorithms. So the story really behind the book is we have artificial intelligence now, but where did that come from? And it really is being powered or generated by algorithms that in most cases were developed decades ago, even if you want to count Thomas Bayes and conditional probabilities centuries ago. So AI is empowered by, from an algorithm perspective, it's not powered by anything new. It's using algorithms that were developed many, many years ago by originators who had no idea how their creations would eventually be used or used in this way. That's half of it. The other half is those same algorithms are being applied at scale, thanks to engineering breakthroughs. OpenAI demonstrated just a few years ago that you can have more predictive models by training your models on larger data sets and throwing computing power to be able to churn through that data effectively and efficiently. So I like to think of it as brain power versus processing power. Like it's not the brain power that is further powering modern AI, it's really the, the old brain power, I guess, the old algorithms with the processing power, which is why you've got these massive data centers that are being constructed.

[B]

I was just gonna bring that up. Yeah. Kevin o' Leary just. You saw a lot fiasco, right? He got a lot of shit for that. Yeah. But yeah, they're building them up because the old ones can't come, can't keep up with the demand. Right.

[A]

Well, you've, it, it's, it's, it's, it goes back to the findings from OpenAI. It's not to get better performing models. It's not uncovering other algorithms or generating new ones. It's using the same algorithms, but now applying them at a much, much larger scale. So train them on larger data sets. So larger volumes of data, more parameters, more processing power, more hardware to churn through that data. That's really where you get the performance differences and the performance gain. So that's why I say it's more processing power than brain power, because we're just throwing more hardware at it. But there's value in doing that. And again, that's why you've got these massive data centers being built, because OpenAI demonstrated this is how you get better performing models. And so there's a power law relationship between the amount of data that you train a model on and the kind of performance uplift you get from training that model or any model, quite frankly, on additional data and then having the hardware to churn through all of that.

[B]

So do you have your own model or are you using other LLMs at the moment?

[A]

I use others.

[B]

Okay. Yeah, I like Claude. Right now, Manus OpenAI is all right. I feel like they're falling a little behind. What about you?

[A]

You know, I think I've noticed that different language models have different strengths and weaknesses, so I often play them off against one another. Like, I'll use. I'll use one as my primary, but if I'm not really sure of the response that I'm getting, I might share that with another language model.

[B]

I do that, too.

[A]

Yeah.

[B]

I asked the same question to multiple. And.

[A]

Yep. And see what kind of response. Yeah.

[B]

Very interesting how they're all different.

[A]

Oh, yeah. Because they're, because they're trained. Not necessarily in the same data set. Right.

[B]

So you got to be careful, too.

[A]

And why do you say that?

[B]

I'm going to tell you. So my dog. So we went to the vet. My dog like, like, had some issues with his leg. Okay. So we got an X ray. All right. Okay. So I got the opinion from the vet. I'm like, let me, let me run this through. I get a second opinion if he needs surgery or something. So I put in five different lms, and the responses were just so different. One of them was like, get surgery immediately. The other one was like, nothing is wrong at all.

[A]

Okay.

[B]

So, you know, you got to be careful. But. Yeah.

[A]

Oh, yeah, yeah. And that's why when you, when you log into Chat GPT, for instance, there's a disclaimer in small. In a small font on the bottom of the screen. Chat GPT can make mistakes, you know, check, check the results or whatever.

[B]

Very small disclaimer.

[A]

Right, Right. So, yeah, they're covering themselves.

[B]

Yeah.

[A]

And I'm sure the other language models do the same thing.

[B]

Yeah, Yeah. I know from. A lot of people are using it for legal advice, for medical advice, and they probably don't even want to be advertising that.

[A]

Yeah.

[B]

You know. Yeah.

[A]

So, yeah, it's a little. But is that any. I mean, you got, you got, you. You interacted with five different language models. You could have interacted with five different doctors. You also could have gotten five different responses from five doctors.

[B]

Right. Very good point.

[A]

So. So you saved a lot of money and time. I did, but I, I didn't get the consensus that you were probably looking for or hoping for.

[B]

Like three out of the five were pretty similar. So I kind of just took that. And the one and the one were total opposites. So you kind of have to still have your own. You can't just blindly follow AI like you still need your own discernment.

[A]

Right.

[B]

You know,

[A]

in my kind of world, in the data science world, I think that's especially true. I was in a conversation a few weeks ago and I sort of framed it as knowledge versus wisdom. Like I think the language models provide the knowledge, but you still need human wisdom to really diagnose things. I mean, if you were to deploy a model right. In production after you've trained it for however long. Right. And let's say the model doesn't perform in production the way it did during training, who's going to detect that and then who's going to diagnose the problem and who's going to explain what's going on to executives and then who's going to apply the corrective action? I still contend that those are going to be human triggered activities. Now there may be some AI interaction along the way, but I think they're going to be still triggered by humans. The insight, there's still a human need for understanding the models, especially when a model doesn't perform as expected. Like I said, who's going to detect that and who's going to understand why the model is performing now in production the way it did perform when it was being trained? I don't think AI is going to necessarily pick up on that for you. I think that still requires some human understanding. And that's one of the points in the book that I'm now writing is you can't just outsource or delegate everything to AI because that's not going to get you where you want to go. But there are gaps that need to be closed. It's easy nowadays to. Just like you did for your dog, it's easy just to engage or interact with a large language model. But as you said, that can be dangerous and I concur with that. There still has to be some human element involved to sort of counteract what you otherwise get from AI.

[B]

Yeah. Do you fear it taking away jobs? I know that's a common topic these days.

[A]

Yeah, I, I don't. Well, it will, temporarily at least. So let me get into this a little bit. It's, I think there are, there, there are generally two schools of thought in my mind. So there was a paper that was, that was released, I think Just a few weeks ago. I don't know if it came across my X feed or someone shared this with me, but it was co authored by researchers from the University of Pennsylvania and Boston University. What I thought was most fascinating about it was the fact that knowingly or unknowingly, the authors basically said the same thing that Karl Marx said like 170 or 180 years ago. So back in the day, Marx said that. I'm not a Marxist, by the way. Okay. But Marx said that capitalists had an incentive or capitalists were incentivized to reduce their operating costs. Right. And a big part of that was through automation, Right? Not the kind of automation we're now talking about. We're talking back then it was industrial machinery, right? And machinery would displace low paying, low skilled jobs, many of which were quite frankly unsafe and unhealthy. But, but capitalists were incentivized to do that and so they would introduce automation and reduce headcount. But he made the point that workers are also consumers. So if everyone is automating and reducing their headcount, then you're actually reducing the demand for your products and services. And that's what these researchers said in this paper. And I think there's a lot of truth to that. It's more around, well, is that like a, a permanent condition or is it temporary? I think that's really more of the question. Marx made the argument that. He called it a contradiction of capitalism and he said that capitalism eventually would collapse on that contradiction and it would then be replaced by communism. And in fact, Marx consistently said that capitalism needed to precede communism for that very reason. So he was, in my view, he was oftentimes accurate at diagnosing situations and problems. Thankfully, he was not very good at predicting. So what I think we really have is, and we're seeing this now, right? So there have been concerns around automation as long as there's been automation, right?

[B]

Yeah.

[A]

So it's different now. I mean, back in the day, as I was sort of insinuating, it was more around machinery than more recently around computers. And now it's around AI. Right. So there have always been concerns around some form of automation, regardless of what form that has taken historically. Marx said that that would result in a permanent problem that would lead to the collapse of capitalism. But there was another political economist who I think should be really getting more play in the media, but I don't think he is, and that's a guy by the name of Joseph Schumpeter. I might not be pronouncing his name correctly. So I apologize to his descendants if I am. But he made the case that economies go through these periodic resets. So what we're seeing right now is maybe one of those resets. So when you have something like AI come along, we're very good at identifying the jobs that are going to be impacted. Right. So the CEO of Anthropic, he's gone viral because he's talked about all the jobs that, that he's going to displace. Right. And not just him, but you know what I mean?

[B]

Yeah.

[A]

But eventually, I mean, if history is any guide, we eventually find new jobs, we find new roles for people and technology allows for the creation of new jobs, sometimes new industries, but you end up having this gap. So there's a gap between, okay, we have jobs that are impacted and then it takes a little longer for us to identify the new jobs that need to be created.

[B]

Yeah.

[A]

And that's, I think we're in one of those gaps right now.

[B]

I agree. Short term pain for long term benefit.

[A]

Right. Schumpeter called these, these, these phases creative destruction because he said that economies are basically resetting themselves and gearing themselves up for even further growth. But yeah, you go through this temporary phase of turmoil and uncertainty because you have the job loss first. I mean, we're very specific on the jobs that are going to be impacted. Right. Some people say, whether it's Mark Zuckerberg or Jeff Bezos has said similarly recently. Yeah, we're going to create new jobs, but we don't exactly know what those are yet. So we're still in this phase of uncertainty. But I'm optimistic that it's not going to be the Karl Marx scenario or the scenario that was presented in this paper that was released a few weeks ago. It's going to be, yeah, we're going to be in a period of uncertainty and some further turmoil for a while. But I see it as being very finite and very temporary. And if history is any guide, we'll eventually figure out that we need other roles, we need other jobs to be created.

[B]

So.

[A]

And to combine those jobs with the new technology and that will spur economic growth. We just don't know exactly what those jobs are yet.

[B]

I hope so. I really do.

[A]

Oh yeah, I hope so too. I mean, I'm trying to be realistic, but. But I, I'm up. But again, I think if history is any guide, that's what will, that's what will happen.

[B]

Yeah, I hope so because there's all these talks about universal basic income now and how AI is going to make everyone not able to work.

[A]

Yeah.

[B]

And that doesn't sound fun.

[A]

That does not sound pleasant. No, no.

[B]

I don't want my kids growing up in that.

[A]

No.

[B]

You know, AI taking away all their work and stuff. So I lean towards your side.

[A]

And for now I think I'm hedging a little bit just because I don't want to be one of those guys who automatically thinks that. Okay, well, historically it's been this way, so therefore it will absolutely, definitely be this way in the future. History doesn't always. Is not always a leading indicator of the future. But that's why I'm hedging a little bit, but just a little bit. Otherwise I'm firmly in the camp that we're in one of these creative destruction, creative destruction phases that Schumpeter described maybe 100 or so years ago, for sure. And it will eventually end and companies and others will figure out, okay, well, there are some new jobs that we need and here they are.

[B]

Yeah, stay tuned on that, both guys. By Gary, you're also speaking at a conference coming up. When is up?

[A]

I am. So I'm speaking at a conference of data scientists, specifically our programmers. It's in Poland.

[B]

Okay, wow, sounds fun.

[A]

So a good excuse to travel a

[B]

little bit, learn some polish.

[A]

Yeah, well, we'll see about that. I know data science better than no polish. So I'm speaking about project management. Okay. And because I'm speaking to data scientists that specifically R programmers, I'm otherwise demonstrating, or will be demonstrating, how you can use the R ecosystem to plan and manage a project. Right. So that might be counterintuitive to even people who are very familiar with R and use R every day, but because project management really should be very quantitative in nature. I'll get to that in just a sec. R is really a more natural application to plan and manage projects than what others might expect. So project management might sound dull to some people and might sound dull even to data scientists, but in the workplace, Sean, almost everything is done through a project. Right. So. And some of these projects might be very simple and straightforward, others can be cross functional. So you've got several teams and organizations in the business and in it, and they need to contribute to this very same project. And many of these projects are absolute clusterfucks.

[B]

Yeah.

[A]

And so why is that? And from my experience, and based on my observations, projects go off the rails more often than not because of bad project management. And you can otherwise blame execution elsewhere across the project. I don't Think that's the problem? I think project management more often than not is the problem. Interesting. Here's the thing. Companies and other organizations, they make this assumption, okay, someone isn't a good programmer and that same person isn't a good analyst, so therefore they must be a good project manager, which of course is completely illogical. Okay. And then other organizations underestimate the skill set required to be an effective project manager. Okay. So they focus on the soft skills. Hey, we need someone who is well organized, we need someone who can communicate well, and those are great things. But project management also requires hard skills, just like any other role requires hard skills. But those are not understood or appreciated as well as they should be. So what happens is you're getting really unskilled people in project management roles. So you think about all these people, let's say you got a cross functional project, all of these employees, all these other resources from both IT and business assigned to a project, and the person who might be the least skilled is the one leading the project. So you've got companies and other organizations underestimating the need for project management. They don't understand the full skill set that's required to manage a project effectively. So you're not getting a very skilled person into those roles, but yet they're leading the project.

[B]

Doesn't make sense.

[A]

Right. And I think that more than anything else is why so many projects go off the rails. Especially the larger and more complex they get, the more off the rails.

[B]

I can see that.

[A]

Again, it's because of bad project management.

[B]

They'll spend a lot on the actual talent other than the project manager, and they blame the talent for failing, not the project manager.

[A]

That's absolutely correct. That's absolutely correct.

[B]

I've seen a lot of stories of that.

[A]

So yeah, I'm going to be speaking on quantitative techniques that should be applied to project management.

[B]

Nice.

[A]

And coming from the data science world, managing a project is in many ways very similar to other data science methods. Right. So when you think about all the different models that you can develop, for instance, and other quantitative techniques that data scientists sometimes employ, whether that's creating a Monte Carlo simulation or developing a Markov chain, or fitting some kind of machine learning model, they look and sound very different, but fundamentally they're very similar. They're all around reducing uncertainty. And project management is also a very project, I should say a really an uncertain situation. We don't know. We might know the tasks that are needed, but we don't know the relationships between those tasks. We don't know how long each task will take. We don't know what the critical path of the project really is. We can't figure out what the project timeline really is. We can't develop a probability for completing a project by a certain date or within a certain time frame. Those are the quantitative techniques that should be applied to project management. But even seasoned project managers don't know what they are. And the companies that are hiring project managers and assigning them to projects, they don't know them either. And so you've got projects that are, like I said, they're complete clusterfucks. Well, maybe one is resulting or leading to the other, so it might sound dull, but. But I think, like I said, everything in the workplace is done through a project, so it's a big issue since you figure out how to manage those projects effectively.

[B]

Yeah.

[A]

So that's what I'm going to be discussing is those quantitative techniques and how those can be applied within our.

[B]

Awesome. Well, we'll link your LinkedIn's, your. Your platform that you will link that in the video. Anything else you want to link or close off with here.

[A]

Well, then I really, I really appreciate the opportunity to be here.

[B]

Thanks for coming.

[A]

Love, love being here. I love having an opportunity to discuss some of my work and some of my thoughts on some things. So thank you.

[B]

Yeah, of course. We'll link your books, your LinkedIn, and good luck at the conference, man.

[A]

All right, thank you.

[B]

Well, check them out, guys. Check out the links. See you next time. If you learned anything from this episode or got any value at all, please share this episode with a friend. It helps us grow the channel, it helps us grow the podcast, and it means a lot to us. Thank you so much.