Hacker News new | past | comments | ask | show | jobs | submit login
Moneyball teams (bg-blog.com)
383 points by bdg on May 1, 2017 | hide | past | web | favorite | 169 comments

I wish I could triple like this! Hire the best is lazy thinking.

* I have been on teams that were not considered the best until they had success, then everyone on the team was a "star".

* I have worked places where they only hire "superstars"-- and were pretty serious about paying for good talent and being picky-- and still failed for all kinds of organizational reasons. Teams with more role players, or more cohesion of goals would have performed better.

*I have seen people recommended because they worked at "X" or worked on the "X" project, and then turn out to be completely underwhelming. Great teams don't always only have great people, and sometimes the great team has people on it who are one promotion away being at their level of incompetence.

At a former (large, organizationally deficient) company, there was this small team of 5 people with very different specialties (and personalities). They were a "strike team" of sorts, tasked with helping out at a technical level different projects that were in trouble. They were excellent. They saved my team's integration of a camera co-processor in an SoC.

But finally a re-org came around and some high-level manager didn't have a box under which to put them (or didn't have the right budget item to fit them in), and the team was dissolved.

The guys were good in the teams they each landed in, but the synergy[1] was dead.

Nowadays, my primary criteria for joining a team is the quality of the people and of the manager. I wish I could say I've seen other teams of the quality of that first one.

[1] yeah I said it

One of my most enjoyable periods while working for someone else was at CapGemini when I was between teams and lent out to do short term fixes for teams where they didn't have enough developer resources/the right resources. Typically just a week or two to go in figure out what was going wrong and patch their systems then get swapped. I don't know if it was the change of pace or the fact that you kind of got to feel like an office superhero that I liked more but I was really young so probably equal parts both.

I had a similar stint helping teams with system performance issues. I think I found it so enjoyable for two reasons:

1) Solving the problems involved the focused exercising of all of my software dev knowledge – often around database access, sometimes around garbage collection and memory allocation of the specific platform.

2) The impact was highly visible and appreciated (by the team, management, and system users), and the result was usually objectively measurable, e.g., the global search now completes in 1/50th of the previous time.

That said, when I reflect on the 10 or so projects I helped in this time, they were all problems the senior dev(s) on the team could have figured out. So, it wasn't really that I was some rockstar who had special knowledge. It was more that I was given time to focus on the issues, while the regular devs on the team had bugs to fix, features to work on, etc.

Like you, I now work for myself, but if I were offered a job doing small-medium performance optimization projects (without travel), that would be tempting.

If there were one key benefit of consulting / being an internal developer at large it would be the ability to leave after finishing a project.

Sometimes that's sad. (When the client / team is awesome)

But usually it means not having to deal with any of the worst parts of office politics. (E.g. de facto stack ranking, non-merit-based favoritism jockying, office management politics, dysfunctional management priorities, waiting for someone to die / retire for a promotion, etc)

Are you still doing consulting? I find that the upside of consulting is that there is a well defined finish line before you are deployed.

When you land in a role that has no finish line, it kind of removes the benefit of being a consultant.

Not really. I moved more into product work but I have one existing contract still going for a few hours a week. No finish line but it's a steady income stream and minimal hours so I've not minded it. It helps that the client is fantastic.

That reminds me of a (proably apocryphal) story regarding IBM's "Black Team" [0]:

> The individuals who made up the team were not exceptionally intelligent or talented, but they all enjoyed testing software and were better than average at it. When these like minded individuals were assembled, they they spent their working hours, lunches and sometimes free time collaborating on how to better find software defects.

> Soon the members of team were twice and then dozens of times more effective than their peers, and they began to view their jobs not as testing software, but as breaking software. Team members took a well-deserved pride in their abilities and began to cultivate an image of villainous destroyers. As a group, they began coming to work dressed in black and took to calling themselves "The Black Team."

[0] https://news.ycombinator.com/item?id=2472219

How do you measure and quantify a team's quality when joining? Especially if you are moving from a different company, I feel like it is probably impossible to know as you may only meet a small subset of the team you work with, if you even get to meet anyone on your future team at all.

I would say: - Look at what they have accomplished. But be sure you are looking at accomplishments of the actual team, not people who are taking the credit (that is a little trickier). - See if there is relaxed communication and trust. - Ask everyone individually what the goal of the team or the goal of the project is. Really talented teams sometimes put their individual goals ahead of the overall goal. As in: the front end engineer says a responsive front end is most important, an AI dev. says deep learning is the key feature etc. All can be talented and well meaning. But if everybody doesn't agree on the trade offs, effectiveness will suffer.

You can't measure it before you join, just as they can't measure your effectiveness before you join. The best you can do is hit the eject button after 6 months after you've assessed what's really going on. Same goes for the employer's side.

> You can't measure it before you join

You can generate heuristic data though. This really should be a priority for you during the interview process. They have the right to ask you very difficult questions, and you have the right to ask them very difficult questions as well.

What's your criteria to measure "quality of the people", in this context?

Can I communicate with them well? (which is a huge topic in itself) Are the people I talked to focused and flexible? (this is not a contradiction) Do the team members refer to each other, and if so, in positive terms? Are the people interesting and curious?...

> Nowadays, my primary criteria for joining a team is the quality of the people and of the manager.

That's mine too, but nowadays teams and especially managers last about 6 months in one place. What is one supposed to do?

But, that's leveraging A talent across an entire organization. Why pay for 2-3x for all A players(that are going to piss all over each other anyway) across the entire org when you could with 1/2 hr of help a day use a bunch of B and C players. and save a lot of money.

thats ok, its still a real world and a real thing if its used in a meaningful context

No worries. It was a reasonable deployment.

Yes, people really need a well-organized environment to thrive, and the environment needs to be tolerant of humanity.

You can build a supercluster of massive computers, but if the coordinators are not able to map good tasks onto the appropriate resource or reduce the output into something coherent and comprehensible, your cluster is going to be sitting at 99.9999% idle time.

And just as you can't stick supercomputers exposed to the elements in the Mojave with a 3G connection and expect them to last more than a few more minutes without crashing, you can't stick good coders in an environment that's not conducive to their group and individual welfare and expect them to do anything except crash.

Humans have much less consistent performance output than machines, no matter how "10x" people think you are. They are even more sensitive to their environments than delicate computing devices. They require a lot of coordination and organization to output something useful.

Everyone ebbs and flows. Focusing on productivity multipliers usually ends in disaster. In most cases, today's raw skill is much less important than a realistic, productive worldview.

This extends beyond hiring. For example, selecting a spouse based on present qualities is much less intelligent than selecting a spouse based on their attitude and worldview and the changes that worldview is likely to precipitate, regardless of the current state.

Ahhh sure very nice...of course a "meta-selecting" system would almost surely outperform over time, but my experience seems to tell me that employers aren't really into "ebb-and-flow" productivity...they want consistency and dependable "X done in y time" workers.

I got a good perspective on this from little league coaching.

Sometimes that "superstar" kid is individually talented, but a net-negative on the team if the kid is a ball hog or doesn't share.

Coaching gives you a great perspective on team building. I'm still a novice coach, but my main goal with a new team is to get them to think like a team: it's not about the individual, it's about furthering the team's goal. Last game a really great hitter on the team had the opportunity to win it in the last inning: tie score, we had the last bats, no extra innings, and a runner in scoring position. With a simple ground ball, she could have driven in the run. Instead she went down swinging for a home run. That tells me I have a ways to go in developing her sense of team.

NOTE: We did win when the next batter came up and got the hit we needed.

yes! coaching helps you become a better team member, and you better understand the need for meshing personalities and having the right blend of specialties, not just stars. in basketball, i like to emphasize passing, moving without the ball, and help-and-recover defense, which are crucial skills to winning, but are often neglected at the expense of individual offensive skills. there are parallel sorts of skills in product development--you don't just need features, but also testing, documentation, visual design, analytics, etc.

i recommend coaching heartily. just don't get overly invested in winning (it's an important goal, but it's not everything).

Imagine if you picked a sports team by the most paid players in their league.

The football team would be half quarterbacks, the hockey team mostly goalies and so on.

The soccer team would be Real Madrid.

The baseball team the New York Yankees.

They seem to do OK.

The baseball team currently with the highest roster salary is not the Yankees, but the Los Angeles Dodgers by almost 20%. They're currently in third place, behind teams with the 16th and 26th highest roster salaries (out of 30 MLB teams)[1].

Obviously it's early on the season, but we can look at last year's season to get a better heuristic. The top two teams in terms of roster salary were once again the Dodgers and Yankees. The Dodgers did come in first in their division but finished the season 12.5 games behind the eventual World Series winners, the Chicago Cubs, whose roster salary was barely 2/3 of the Dodgers. Oh, and the Yankees came in fourth place (of five) in their division, a whole nine games out of first place.

So yes, spending money helps you be a contender, but spending the most money is not a good metric for the best team (at least, not in baseball, and I'd imagine that being smarter about spending your money rather than spending more is useful in other fields as well).

http://www.spotrac.com/mlb/payroll/2017/ http://www.spotrac.com/mlb/payroll/2016/ http://m.mlb.com/standings/?yyyy=2016&view=divisional

I'd imagine the majority of sports have a fairly loose correlation between money spent and winning due to (a) limited sample size leading to high variance (even with a superior player selection strategy), (b) limited number of games played (though baseball does better than other sports here), and (c) games consisting of more dependent than independent events (leading to low chance events having cascading repurcussions).

At least in MLB, 40 multi-year players isn't that large a sample!

I don't disagree, but I also don't think that's indicative that paying the most money is the optimal strategy. I'd much rather play for a team that does win rather than one that "should" but doesn't. My point, which I alluded to in the last paragraph, was not that teams shouldn't spend a lot of money, but that spending a lot of money is not a good substitute for spending money intelligently.

I can't speak for the New York Yankees, but the 'Galacticos' policy of Fiorentino Perez, where Real Madrid would try to sign one expensive, world-famous footballer every year, is widely considered to have been a failure despite initial success - partly because of a lack of interest in signing defensive players (they didn't want to pay Makelele's salary despite it being relatively low), locker room egos, and constant interference from the Club President over hiring.

To throw another example on the list, the New England Patriots are 30/33 for salary money spent. The Miami Dolphins are 5th. Spending big money for rockstars is neither necessary nor sufficient.


Sports is a terrible analogy for work. A sport is a game with rules that are created to create an entertaining event. The rules are defined. The outcomes are always instances that never build to a lasting change to the structures that created the outcome. Work is not like that in any way.

Well the US Olympic Basketball Dream Team did crush it.

Even that team was fielded based on position. They didn't have 12 centers.

They didn't have 12 Cs [0], but can you think of any absolute snubs? The team only had 1 true center and 1 true PG.

Isiah Thomas was deserving, and Laettner was picked over Shaq. At the time, both were tossups. Johnson was older and so was Bird, but I assume they had high salaries per the topic.

[0] http://www.basketball-reference.com/olympics/teams/USA/1992/

On the other hand, a star quarterback or goalie has a disproportionate positive effect on the team's winning chances, and chances to win championships.


But even great quarterbacks struggle if the O-line is dysfunctional.

And you definitely wouldn't be better off with 6 star quarterbacks than 1 average quarterback and 5 O-line players.

That's an argument for snatching up one of these people per team, but not an entire team of them. The analogy being made I believe is that there's simply no benefit to having two or three quarterbacks on a single football team. It doesn't even make sense.

With exceedingly diminishing returns.

This is probably true of a quarterback, but it's not generally true of hockey goalies. You'd much rather build your team around Sidney Crosby than any goalie to ever play the game.

That's not true. You can take a look at the Rangers right now, whose only true "superstar" is the goalie- Lundqvist.

OK, but that's my claim. My claim is that you'd still rather have Crosby, because there's a much bigger talent gap between him and his hypothetical league-average replacement at center than between Lundqvist and a league-average goalie.

Obviously I meant to say that's not my claim :)

Ok, fair enough! Though, has anyone ever analyzed this question rigorously with data? I.e. is there a Bill James equivalent in the ice hockey world?

The larger point is that individual stars definitely matter a lot in pro sports. And recruiting and retaining them with extremely disproportionate amounts of money doesn't seem to be irrational for most sports.

See my link to an attempt at a hockey WAR (wins above replacement) elsewhere in this thread.

Look into Corsi and Fenwick.

Goalies are the highest paid players in hockey? Interesting.

No, this is not true. The top paid players in professional hockey are almost all high-scoring forwards.

4 of the top 50 highest paid players in the NHL are goalies.

I didn't do my research there, sorry.

You hear a lot about how a good goalie is half the team in modern hockey, so I thought that would translate to salaries.

Anyway, I hope I made my bigger point clear, despite the questionable example.

They're out on the ice the entire game, and usually the only person on the team that can play that position at any acceptable level. That on top of goalies being your last line of defense, having to take 25-35 shots per game.

The same reason your team fights a lot when the other team messes with your goalie; can't afford an injured goalie.

I'm not a fan of hockey, so I'm asking this question out of pure curiosity and ignorance. I'm assuming advanced analytics have made their way to hockey. Have the analytics supported this?

No, not really. Top-line centers are usually the most valuable player on a team, followed by other scoring forwards. Scoring is a premium in the NHL, which means that a very good scorer is harder to replace than a good defensive player, all else equal.


I do like the Belbin Team Inventory as a way of looking at team dynamics I was Plant Chair Shaper when I took it a few years ago,


I'm in complete agreement, but to add to the thought of "Hire the best..." as lazy, I think it's just incorrect.

The best what? If you're talking about a software development team, are you hiring the best software developers? The best devs in what niche? I consider myself a very good developer, but I would be shit for quite a while if I joined a team focusing on embedded systems. Or if I was writing an OS kernel. I would likely get up to speed and surpass some people, but I may never be as good at Y as I am at A, B, and C now. I may just not have the passion for it that I do for those other things.

To agree with your point more, assembling a team of 'rockstars' who are all good at just about the same things means that your team is likely going to have very large blind spots (say: testing, for instance), and contention over people's preferences for work. We can't all be the architect of a greenfield real time data processing service.

I fundamentally believe someone who is engaged, enjoys critical thinking, likes to learn, and is generally good (or at the very least conscientious) at their craft can move from technology stack to technology stack if it aligns with their personal interests and/or motivations. I do, however, believe we generally have our sweet spots where we are most apt. Maybe due to our preferences, possibly our aptitude, likely a combination of those and our personal preferences.

I consider myself stack agnostic, I float between a half dozen languages a month between professional and personal projects and depending on the problem trying to be solved. But if I sat down and thought about it, a couple of languages would float to the top as my preferred or best, and the rest would be because I'm good at solving a certain type of problem with them.

We can't all be the best at everything, but we can be really damn good at some, and together we can go full Voltron and kick ass. It's just too hard for someone in HR/Talent Acquisition to figure out who should be the leg or the arm (or even what Voltron is), and they're likely the filter most candidates must get through.

EDIT: To add, I'm not just ragging on HR/Recruiters either. I generally know what I'm looking for in developers but still consider it hard to interview and hire. I don't think I'm very good at being the interviewer in the formal interview process. Now, if we had all the time in the world and left the office to grab a pint and just talked like human beings I think I could feel it out better, but now I can't get any work done and I'm going to be drinking 5 days a week until I fill the position.

>I have been on teams that were not considered the best until they had success, then everyone on the team was a "star".

Yes, it's very important to realize that this is how people think; most of our decisions are subconscious, and the frame/bias that informs everything else about our perception is determined rapidly, within seconds of evaluating a situation or becoming acquainted.

If you bear the tokens of success (i.e., if you "look" successful), people will assume that means you know what you are doing, and if you don't, people will assume you don't. There is a pronounced halo effect on association. Anchor yourself against something generally considered successful and your perceived intelligence will go up precipitously.

Entrepreneurs especially need to take this lesson to heart. You will be a daydreamer and a slacker who can't handle having a boss right up until your company appears successful, and then you'll be a visionary that everyone is falling over themselves to try to imitate.

But but...then what will the MBAs and C-suites do if they can't stack rank and yank people, saving some $$ and getting their own promotions?

Another way to look at it is: What causes projects to succeed or fail?

I've seen my share of failed projects throughout my career (as a participant and as a co-worker watching it happen). By a huge margin, most of the failed project did not fail because the engineering team was weak, or because they didn't hire "rockstar" programmers, or because the engineers didn't go to Stanford. They failed because of things like requirements growth, unrealistic schedules, emotion-based/gut-based (rather than measurement-based) product decisions, and external factors (market changing and company couldn't react). If companies spent more time worrying about these things and less time worrying whether their engineering interview candidates can balance a binary tree on a whiteboard, we might see hiring practices change.

I agree with everything you've said so I won't add to it.

There is an interesting anecdote that I love. Some number of years ago there used to be a show called EcoChallenge. It was made by Marc Burnett but before Survivor. It was simply an adventure race but no politics and no contrived voting. Each team of 4 must be co-ed. There is a starting point, there is a stopping point and there are way stations along the way. Simply go from start to end hitting the way stations. Everything else was left up to the teams. The one big twist is that you can't win unless ALL of the team hits each way point. Teams are allowed to race 24x7.

One of the last years it was on tv, a team of navy seals bombed out and a team of 3 playboy playmates went further then they did. The team that won (from NZ) had a sick teammate and spent two days nursing their teammate into passable health. They loaded up all that person's gear split 1.3x, 1.3x, 1.3x, 0x and set off. They rocketed past all of the other teams and won by over a day.

This is what real success looks like. A determined team laser focused on the exact same goal and doing whatever it takes to get there.

A bit more on topic, I was asked to read a book on leadership and teamwork that's a bit of a cliche these days but Coach Wooden's Leadership Game Plan for Success covers a lot of these ideas specifically in a sports context. A lot of leadership texts revolve around sports out of sheer ubiquity in subject matter among business leaders culturally but beyond some pop sociology type of research I really don't pay much attention to them in a business setting. Why? In much of the world of software, day-to-day work is not a race against a competing team towards the exact same goal as much as it is a race against oneself and against market inertia whether that may be the default mode of failure or convergence to an equilibrium point that needs to be readjusted as the team discovers the hidden rules of a game.

In fact, a lot of business is probably a lot closer to something more sadistic and cruel in gross mechanics like from Takashi Miike's 2014 As the Gods Will (and the manga of the same) - arbitrary filters where the rules are very clear only until after people have died. Some stages test raw individual skills while others test cooperation. And in some cases it didn't actually matter if you had either of those because "luck" is the bigger factor (the fact few in practice agree upon these slack variables to achieving business success shows how random things are for survivorship moreso than controllable factors like skill and effort)

> By a huge margin, most of the failed project did not fail because the engineering team was weak...

YMMV here. I've seen many projects fail because teams unnecessarily accumulated technical debt of various sorts. I guess you could blame management for not knowing about it or prioritizing correctly or something, but good engineers know that you have to include technical cleanup in the feature estimate or you'll end up like this.

I've also seen many unforced errors that created more technical debt than was worth it (over-complicated from-scratch frameworks, betting on unproven demo-ware, etc.).

I've seen some of those failures but made a different conclusion: None of those teams had the balls to say 'no' any time management got even a little bit upset about anything.

Now, part of that is on the engineers, but the only solution I know of is solidarity. But a lot of it came from systematic abuse from the project managers, wearing them down and then calling them incompetent when they couldn't deliver.

Just like this article says farther down, everybody is good at something. I've only had to talk about firing a handful of people (rather than reassigning) and it was always because they were belligerent to progress, or because of Dunning-Kruger Effect - they would not admit they weren't good at something and keep creating 2x as much work for everybody else who has to clean up after them.

Punishing people because they aren't good at everything is just some sadistic bullshit that makes small men feel important, and a lot of those are attracted to middle management. It's twisted and wrong and only rarely does a company succeed with that kind of corporate culture. And when they do there are always memoirs from the victims.

One of my favorite movies, Miracle about the 1980 U.S. Olympic hockey team that upset the Soviets, had this dialogue that actually happened between Head Coach Herb Brooks and Assistant Head Coach Craig Patrick. I feel it applies here.

Patrick: This is the final roster? You're kidding me, right? You're missing some of the best players.

Brooks: I'm not looking for the best players, Craig. I'm looking for the right ones.

Which implies something major that everyone is glossing over:

If the "right people" aren't the same as the "best people" are the criteria we're using for "best" correct?

I think the real lesson - and the premise behind Moneyball - is that when you have a complex objective, you can't evaluate people in a vacuum. You have to optimize at the team level. You can try your damnedest to come up with heuristics and metrics to evaluate individuals, but if you don't allow yourself to consider the context that they'll be working in, you'll miss a lot of potential.

I don't think this is the central thesis of Moneyball. I'm a huge baseball nerd, including on-field and statistical side. The idea behind the WAR (Wins Above Replacement) stat is that you should adjust for context, then evaluate players in that vacuum.

Moneyball was about identifying and exploiting market inefficiencies. It actually rejects the team level optimization ("he's a team player" is explicitly rejected as a valuation of a player).

Instead, Billy Beane identified what components led to team success, then found which of those the market doesn't pay for. Then he got those guys.

Used to be thought that a good baseball team had a speedy leadoff hitter. Teams would play objectively worse players because they thought their style better "fit" their role.

Moneyball was the start of the revolution that you just want to get the best players. Then accept that playoffs don't really mean anything other than randomness with small sample sizes.

Agreed. It's also worth observing for those who might not really be baseball fans that baseball is really less of a team sport than most major sports. Yes, team dynamics can reach toxic levels and sometimes teams seem to outperform their individual roster members. But other than some degree of matching players with positions (pitchers in particular), you mostly want players who would be best/best value on key stats under any coach or any team.

The entire offensive side of the game pretty clearly violates the notion of independence. In particular, the value of a batter with particular stats is dependent on the batters that come before and after him. This is the reasoning behind stacking people with high on base percentages in front of sluggers. If it were not a team sport, you'd see no clear patterns in line-up. This applies at the team selection level as well. For instance, a player who can draw a lot of walks is worth different amounts to different teams depending on their ability to convert those into runs.

The fact that saving a run is nearly equivalent to scoring a run means it bleeds over into the defensive side as well. A phenomenal batter with a high error rate may be of enormous value to one team since their pitching staff strikes out more batters than average while being a terrible addition to another team.

If you're still unconvinced, I could throw together some simulations later that might make this clearer.

"In particular, the value of a batter with particular stats is dependent on the batters that come before and after him. This is the reasoning behind stacking people with high on base percentages in front of sluggers."

This is the older, common wisdom that has essentially been refuted by modern analytics (SABRmetrics, if you will).

Team-dependent statistics like R + RBI (and pitcher ERA!) are dependent on the performance of the rest of the team. They also correlate much less with a team winning games than other stats, like OBP.

That's one of the central theses of Moneyball and the SABR movement: it's better to ignore stats like (pitcher) Wins, RBI, and R, because they measure the teammates contribution more than the player you're tryign to measure.

One popular advanced pitching statistic is literally named Fielding Independent Pitching, and is calculated using only BB, K, and HR. This tracks more with contributions to a team winning than stats that include fielding skill, randomness, sequencing, etc.

What you described for a good hit / bad field team does work around the edges, but these effects are not very strong. The latest research suggests the biggest impact might be matching your pitching staff tendencies with fielders; if you have a flyball heavy staff, invest in better OF defense (Mariners, 2017). If you have worm burners (Houston, esp w/ Keuchel), invest in your IF defense.

The goal of the modern stat movement is to evaluate a player independent of team context using statistics that do not depend on their teammates. This is because estimating a players "true talent level" is much more useful than evaluating RBI.

It seems like we're more or less in agreement...Team performance and individual performance are interdependent, particularly in naive statistics like the ones you've mentioned are traditionally used. You're talking about making team-independent estimations of player value. That's completely valid. What I'm also saying is that in a similar vein, when you are considering how a player's addition will impact a team, you need to correct for his interaction with the team as well.

Yes, agreement! Thank you for nudging me to recognize that.

However, the studies I've seen (regular Fangraphs and BP reader) suggest these interactions between player skill sets is minimal, if such an effect can be shown.

Take lineup position. It's been studied ad nauseam, by some of the brightest in the field. Turns out, lineup order, over the course of a season, doesn't really matter that much. An intentionally suboptimal lineup underperforms an "optimized" one by maybe a couple wins per year. Almost all lineups actually implemented are more like fractions of a win, which is generally within error bars.

By and large, baseball is a game where you just assemble the best talent and they will win. (This is in regard to on field talent; I do not believe "clubhouse culture fit" is as silly as in the tech world, and is usually retroactively defined.)

This kind of reasoning go only so far. While you are technically correct, the difference that something like batting order makes is marginal in the long run. The advantage of a fully optimized vs. a totally random batting order would equate to approximately 1-2 additional win(s) over the span of a full 162 game season.

That being said... one win can be the difference between a playoff spot or not, so it clearly does matter. I'm just trying to emphasize that the teamwork aspect of the offense is minimal, as a lineup of hitters with solid statistics is more important than the order in which they bat.

However, the defensive side of things is much more tricky to measure. Baseball statistics still have a difficult time calculating the value (for better or worse) of a player's defense and nobody is arguing that it doesn't significantly affect a team's performance.

That's fair enough. Although I'd argue that it's more a case of assembling an appropriate portfolio than it is a "team" in the way that people normally use that term.

(I'd also argue that overall offensive ability trumps a lot of optimization around base stealing or the type of hits.)

ADDED: i.e. I'd argue that if populating a roster with about a half dozen All Star sluggers were financially viable, I'd argue that would probably be a pretty effective team however unbalanced.

As covered in Moneyball (the book), the "sluggers" tend to swing at more. So although they may hit home runs, they also tend to strike out a lot too.

The actual approach was to get a bunch of people who worked to extend the inning by not getting out.. aka getting on base safely. You don't need many big hits when you can get singles and doubles consistently.

A couple years ago, I met a small scale angel investment group that complained "we haven't had a unicorn" but when we chatted more, I found out 60% of their investments were acquired in under five years with them making ~5x each time.

Not as sexy but sounds like a good strategy to me.

"Instead, Billy Beane identified what components led to team success, then found which of those the market doesn't pay for. Then he got those guys."

Let me ask you - as I am not a baseball nerd - if there is any counter-argument to the BillyBeane/Moneyball/WAR movement ?

I don't mean knee-jerk, irrational objections - but real, reasoned counter-hypotheses ?

Yeah, the counter argument is that you mess up the team chemistry. This is extremely hard to define and measure so can sort of always be used.

Billy Beane is constantly attacked for destroying the looker room chemistry in the local press.

>>you mess up the team chemistry.

True, only if you have a great team already and you don't desire messing it up. If that is the case, you wouldn't need that method at the first place.

On the other hand, teams are pretty dynamic these days given the interfaces and responsibilities are well defined. Or at least are largely defined well. So replacing team members is never really a problem.

I'm not sure real counter-hypotheses so much as data only takes you so far.

- You can't always predict the future about a given player's performance.

- We're still arguing about what the right measures actually are (although there's more data all the time)

- There clearly are sometimes cases where personalities and attitude mean some player needs to leave even if his historical stats were good.

If you have a team of 6, sometimes the best individual 6 people are not the same group of 6 that you put out there to solve a problem.

Let's continue to use hockey as an example, because I love ill-fitting sports analogies and I also love hockey. For sure, we already know we need 3 forwards, 2 defensemen and a goalie, so for the sake of argument, even if your 6th best skater is Wayne Gretzky, he's going to sit because you need to put Neuvirth in goal (to pick on my own team for a second). But let's think about forward lines only for just a second. Pittsburgh has Crosby and Malkin centering different lines, even though they're the best two guys on the entire team - they do the same job, they're playmakers first, goal scorers second (even though Crosby is the league's top goal scorer). Washington has Ovechkin playing the wing with Backstrom, because even though he's their best player, his job is to take a pass and fill the net.

That's a really long-winded way for me to say that "best" has a situational component, and teams have compositional elements that often involve roles. The 10 best people at that level in the company may all fill the same role, but on a team of 6, maybe there's only 3 of that role, and the other 3 are complementary parts. Best is situational.

Gretzky is great example because he was not the best player because he scored the most. He was the best because just him being on the ice lead to others scoring as well. Just being out there he was able to help the rest of the team score as well. Contrast this with a player like Kobe Bryant who was a great individual player but horrible team player. Kobe did not get along with his team mates and other than the points he put up on the board Kobe did not really contribute much to the Lakers. Kobe was such a ball hog that they refer to a missed shot that is rebounded and scored by the offense as a kobe assist because it was only way he was going to give up the ball

Kobe is top 30 all time assists. What are you talking about?

I detest the man, but you can't use statistics to argue against him.


But he played ball for 20 years. I think average number of assists per game is a better stat here, where he was ranked 136th according to this link [1].

[1] http://www.basketball-reference.com/leaders/ast_per_g_career...

There have been over 3000 people to play in the NBA. Ranking 136th still makes him better than the vast majority of people to have ever played the game. Especially as a two guard on a team that primarily ran the triangle, an offensive system that devalues traditional pick/roll dribble creation and generates most of it points on secondary actions (eg hockey assists) off high post play.

When people either can't or don't want to measure the performance of who they picked and who they didn't, they resort to stories that sound good. Even if you have "rigor" around how you score candidates, if you don't connect those scores to subsequent performance it's all still just a narrative.

That's just a folksy way of explaining local vs global optima. Best in one context is often not best in other contexts.

There is so much statistical noise in any short series of 5 to 7 events that any coach whom thinks they have some magical eye at picking the "right" players is simply delusional.

Yes it worked out for Brooks this time and god-bless-his-success, but really...the whole idea behind Moneyball is that people totally suck at it.

Presumably you have teams of Web-stack ("Fullstack") developers.

This caveat should have been put up front.

Since "Best" is a vague loaded term it's more accurate to understand the problem that you're trying to solve.

If you are creating a new web-standard or industry shifting technology then you absolutely need the best. If you are creating a new web based CRM for firefighters, then you need someone who is good, but probably doesn't do vector algebra in their head or [insert favorite discriminator for something hard.]

> Since "Best" is a vague loaded term it's more accurate to understand the problem that you're trying to solve. > > If you are creating a new web-standard or industry shifting technology then you absolutely need the best.

There's a tension between these two sentences which is worth more consideration about what “best” really means and how it intersects with scalability.

If you're working on a web standard, “best” tends to mean things like “experienced”, “works well with others”, and “broad and deep understanding of what the users do”.

If you're working on a major new technology, it might be more likely that you really need someone who is incredibly good at a particular hard skill but it's still likely that if you have a problem in that class it will exceed any individual's capacity and the limiting factors will continue to be coordination between people, ability to maintain clean separation of responsibilities, etc.

Think about something like LLVM – there's a ton of really hard work in compilers and optimization, it's had a significant impact in multiple parts of the industry, and while there are some people who rightly deserve a great deal of respect for work on those hard problems, a key part of the success which hasn't been virtuoso work at that level but rather keeping a design which is clean and extensible enough for so many different organizations to use and contribute to.

Yes, fair point. Ended up stepping on myself there a bit.

Maybe "Top of their discipline" would be a better way to say it.

Yeah, it's hard to talk about this since we're using a single term to convey multiple complex concepts.

This reminds me of a post on Less Wrong about the meanings of words and the idea we're targeting. http://lesswrong.com/lw/np/disputing_definitions/

Or, hire the best, and make a team out of them. Or hire people with high potential and push them to reach it. Or, hire people who can get the job done.

One thing for sure: if you have mind-numbing processes, "the best" will leave, and others will not reach their potential.

In my view a good environment where people are allowed and challenged to improve will always beat any other alternative. A lot of people have potential if that potential is being nurtured. You can hire superstars but if your processes are bad, they lose their edge soon. Or leave.

Yeah, these "don't do [hip thing], do [contrarian but normal/common sense] thing" articles (or the reverse of them) are becoming a bit tired.

You should have good team-building and organizational skills and good technical hiring skills.

Everybody thinks they're doing the common-sense thing and not being contrarian, even if it's the hip thing.

Very true.

I'm sorry to hear you didn't enjoy the article format. I agree that one should have good team-building and technical skills. I also think there are other and possibly better ways to improve development team throughput besides "hiring the best".

but... but.... management is HARD. [/snark]

Moneyball and Six Thinking Hats are excellent approaches to for people and roles.

But they aren't enough. The key driver of top-tier teams is excellent ongoing safe communication.

To build communications, use team goals (e.g. OKRs & KPIs), team ways of working (e.g. TEAM FOCUS), and team positions (e.g. scorecards).

My notes about these are open source: https://joelparkerhenderson.github.io/

Edit to explain abbreviations:

* OKRs = Objectives & Key Results

* KPIs = Key Performance Indicators

* TEAM = Talk, Evaluate, Assist, Motivate

* FOCUS = Frame, Organize, Collect, Understand, Synthesize


For those wondering: Collect, Aggregate, Reduce and Generate Objectives to Cultivate Understanding Learning and Teamwork

Totally, just imagine how people who don't know anything about tech feel when they don't grok our jargon!.

Ha! You're right, these can sound like cargo cult if someone doesn't know what they mean.

These teamwork ideas are worthwhile IMHO. YMMV. :)

Is "TEAM FOCUS" an initialism, or is it there for EMPHASIS?

Hiring the best doesn't scale. Building teams does. The problem as with hiring the best is that almost no one knows how to build good teams. You will probably spend way more time resolving conflicts and dealing with politics than you'd like and in the end the software product will still be a steaming pile of technical debt.

upvote x 1M

This is a pretty important lesson that gets lost at times. Teams with people of diversified passions work well; if everybody is super into data structures, UX, security, etc., then debates can rage and feelings can get hurt. When people on a team specialize differently, everybody can be a mentor to everybody else. To me, that is a great team to be on.

Yes, and I also suspect this has lead to "we don't hire jerks" policies, which are great when you find someone who enters a conversation and talks down to everyone. That person fails a culture fit.

However, there's another kind of person who just talks over you, doesn't put in the effort to communicate their ideas, or refuses to accept other ideas. These can remain hidden until tested in the depths of some technical discussion. They might pass a culture fit.

I agree, same sentiment from W. Edwards Deming: "A company could put a top man at every position and be swallowed by a competitor with people only half as good, but who are working together."


It happens now and again that the best team has very few of the best players:

- Denmark at Euro 92. The only world class player, Michael Laudrup, decided he was boycotting the coach. His brother was a fairly decent striker, and future Man Utd legend Peter Schmeichel was in goal, but everyone else was a journeyman.

- Greece at Euro 2004. Nobody from a top team in a top league, but Otto Rehagel managed to coach them to victory. Not amateurs, of course, but certainly not superstars.

- Leicester City last year in the Premiership. Nobody on the team was a real target for a top club, except maybe Schmeichel junior, also in goal. But they played a great season and one of them got voted player of the year. And Vardy beat the scoring streak record. What's interesting is how they fell apart this season under the same coach, who then got sacked and replaced by someone who has partially brought back the magic.

The rest of the time, the winners are built around a small number of world class players, and competent squaddies. This is typical of US teams, where you have a salary cap, meaning the superstars will be spread out, and you need to be efficient in paying for the squaddies that are appropriate for that star's game.

So once in Premier League since it's inception (1991 IIRC) and twice in Euros since I don't ever remember when should be that surprising?

My model for ideal hiring are the Spurs. They consistently hire for talent and potential, provide the proper mentoring and conditions for the potential to fully materialise.

"As you sow, so shall you reap". But few have the patience to sow nowadays. We rather buy ready-made in the grocery store.

I like the Spurs (San Antonio) as a better goal to work towards than the oft-cited A's because ultimately, Oakland didn't win the World Series, and lost in that year to the big spending Yankees. Even though the idea of "Moneyball" has merit in many applications, it always bugged me as a sports fan because Oakland never won any titles with that approach.

>It always bugged me as a sports fan because Oakland never won any titles with that approach

There is too much variance (and too much league parity) in the Playoffs short series format.

Moneyball gets you there, where all the gate is gravy.

Most companies cannot have their employees sign multi-year contracts. Outside of sports, there is a much bigger risk of your competitors poaching people after they've developed.

> Most companies cannot have their employees sign multi-year contracts.

No, most companies prefer not to make multiyear commitments to employees. They could enter into multi-year contracts with employees (and they do, with execs, with fairly expensive buyout provisions), but they don't want to for most non-executive employees.

Presumably that's always a risk regardless of whether you hired someone junior and trained them or whether you hired someone already skilled. As long as you're doing the standard things that help you retain talent (like paying market rate, providing good opportunities, etc.), there should be no reason to worry, right?

The only companies that should be worried about "poaching"[1] are the ones, for example, who are not paying market rate, have crappy projects, provide no career advancement opportunities. I've never heard of a talented employee leaving a company for literally no reason.

1: Which by the way, is a pretty offensive term--we're not deer or wild game owned by some feudal lord.

> The only companies that should be worried about "poaching" are the ones, for example, who are not paying market rate....

Well, "only" is implying it's not almost universal. Companies routinely pay market rate for new hires and then base raises and bonuses on KPIs for individuals, teams, or the whole organization.

Are there companies that say, "Well, we had a mediocre year, but market rate for devs went up 8%, so I guess that will be the baseline raise for everyone."? If a company does anything less, eventually employees are paid below market rate and they're forced to negotiate a bigger raise (convincing management to ignore KPIs and pay attention to market rates) or find a new job.

Or they leave for a company paying new hires market rate.

The job market doesn't care if your company had a mediocre year, any more than your suppliers care. If the price of screws or cables goes up 8% you either pay it, look around for something cheaper, or negotiate. Somehow, when it comes to labor, companies think paying “what we paid last year plus N%” makes sense.

I guess the take-away here is: Companies by and large do not worry about poaching. Rather, they are willing to live with it because they see the alternative as too expensive.

>As long as you're doing the standard things that help you retain talent (like paying market rate, providing good opportunities, etc.), there should be no reason to worry, right?

Right, that's the point: sports teams benefit from being able to pay below-market rates to players they have developed. If you're doing the standard things to retain talent either way, why not just hire someone skilled?

I'm not arguing about the moral case here. It would be great if companies invested in people out of a sense of responsibility. But I don't think "As you sow, so shall you reap" translates from professional basketball.

> sports teams benefit from being able to pay below market rates to players they have developed

They also pay to keep players off the payroll of their competitors. There's not strong incentive to develop the players in those cases.

Why would you be worried that your employees will leave after you've helped them grow and develop? If you're hiring employees for potential and fit rather than just going after people with big names on the resume (Stanford, Google, Apple, etc) you're not going to have to pay exorbitantly to get them in the first place. Or at the very least, you'll be able to hire faster and spend less money chasing a very small group of devs.

Put another way, you'll spend less money overall if you can identify, hire, and retain undervalued talent. If you build an exceptional workplace (and pay well) most will want to stay. And if you've done that, it'll be far easier to attract talent in the future.

Sure, you can't prevent your employees from leaving, but you can definitely structure the compensation so that they are highly incentivized to stay, which is what many companies do via stock grants with vesting periods.

I'm four paragraphs in, and I don't want to get too hung up on grammatical errors, but it's making it hard to follow along on what point they are trying to make.

> and was won against with teams spending over $100 million

What does that mean? They spent less on their teams, and they lost? So they should spend more!

> Once upon a time the ambitions of your company were exceeding the throughput capacity of your development teams. This is the ultimate goal of hiring...

This sentence ends with the ultimate goal, but "this" seems to be referring to the previous sentence, when it really isn't.

> If ambition growth constant

I'm guessing they mean "is" constant?

I am a native English speaker and I had a hard time following the point as well. I surmised from the title that the thesis was going to be "x number of decent developers who work together well can produce a product better/faster/cheaper than x great developers who don't get along"

I think a bit of proof reading might make it easier to follow the argument in this article.

I had hoped to convince people that since hiring is about improving team throughput there may be better ways to do that besides simply saying "hire the best".

Thanks for spotting those problems. I've updated the post. I've debated hiring a proof reader and may do that now as I passed this around my circle of friends before publishing.

I am a bit surprised if this is new information to anyone who has managed teams and/or thought seriously about the challenges that job has to face. When ever I hear an executive making the assumption that "software engineers" are a fungible resource and they will just move 6 over from the project that is on track and humming along to the group that is behind and in trouble I know they are clueless about managing high performance teams. Every team needs a mix of people and when done well the talents and skills of the individual members complement each other and create an effective whole. Too many people trying to drive, or too many people unwilling to drive as an example can both leave the team floundering and unproductive.

When someone asks me what I think their priority should be as a manager, their day to day job if you will, it is looking for ways to improve this spread of talents to minimize overlap and cover gaps.

Only hiring the best is inherently unsustainable, because it implies a need for "other companies" to train people into being the best, and only then are you willing to hire them.

The greatest companies are those who offer an atmosphere in which good developers can grow into the best developers. The worst companies are those who believe they can hire people from great companies, and consequently transform into great companies themselves. Great companies produce value; poor companies consume it.

This is an interesting point. Companies that try to "hire to greatness" have a steep fee to pay for their hiring approach, and it might not work for reasons that cannot possibly be squashed into, and recognized from, the sloppy "greatness" scalar used to score the possible team members.

As you said, cultivation of expertise is a better approach, and I'm happy that where I currently work fosters this carefully and deliberately.

One thing that struck me from Ed Catmull's book on Pixar, Creativity Inc[1] was how much effort they put into trying to get people to work together effectively and how important that was in the final results, versus the talents of the "super stars".

After the Disney acquisition, they kept the companies separate, but basically got rid of a bunch of red tape at Disney Animation and injected a bunch of Pixar cultural practices. The DA team went from mediocre results (many people thought Pixar would just shut them down) to producing hit movies.

[1] https://www.amazon.com/dp/B00FUZQYBO/

I used to play basketball almost every week at the university gym. There were these guys in their 60's who still played with 19-25 year olds and routinely won. I asked one of them if he saw one particular thing young people tended to ignore. He said, without hesitation, "the team game."

There is probably some bias, the young guys probably didn't play quite as hard as they would have otherwise, but I'm really not convinced that accounts for the entire story.

Maybe it's just as simple as: interaction effects on teams are often stronger than main effects.

The 'best' is so subjective as to be meaningless. I have worked with probably 500 software engineers, and I have been the best of all of them. (Well, one guy was probably as good as me.) But, maybe this is only because I value things that a lot of other people don't? My criteria are just different.

I have led teams that developed software resulting in 4 successful (another subjective word, but in this case, acquired) startups. I designed and implemented 14 products. All together these products were used by hundreds of millions of people. Some of them are still used 10+ years later.

But, I still can't always convince hiring managers to hire me. I don't have a blog (don't care). I don't have a github (don't have time). After being turned down some say that I'm 'too mellow' or that I'm too abstract or too specific or that I didn't answer their questions (usually one specific) the way that they wanted me to. Again, it's all subjective.

At the end of the day, it's just a fashion show. And, when it comes to work, I'm not into fashion (I have a laser focus on the product, its market suitability, its functionality, and and its market success).

Basically, people hire the ones that they like. And, I believe the evidence shows that most managers prefer to hire people who are worse than them, or at least only better in certain ways. Corporations are just glorified hierarchies, remember.

So, if you get turned down a lot, take it as a compliment.

    > I have been the best of all of them
What's your metric for that? How would you have known if you weren't?

Well, that's the problem isn't it? All you have are confirmation and survivor biases to work with. You're going to get a lousy outcome.

Someone with far more practical experience than you might sound paranoid about some things you don't even think about, and blasé about others (because they have ways to mitigate the problem if it actually happens). You won't know enough to realize why they aren't on board with your priorities, and you will make conclusions that have no basis in reality.

You've only known this person for 45 minutes. You've known yourself for decades. Which one are you going to assume is the idiot in this situation? How often are you going to be wrong? And how would you ever know if you were wrong? You just kicked this person out of your life because you didn't like their answer. If they become lead developer at your biggest competitor, are you even going to remember their name?

Haha - exactly.

This reminds me of Reed Hasting's culture deck (he even uses a baseball team as a metaphor when describing Netflix's hiring practices). To further the analogy, Netflix would best be described by the Yankees--pay top-of-market for those who have proven they are the best at their role. This has been effective, for both the Yankees and Netflix. But, the interesting caveat, as alluded to by the OP and the moneyball A's (and other teams), is that this might not be an optimal strategy.

I think the Yankees model is a (possibly reaching) metaphor for capitalism. All the capital concentrates in a few hands. But where are all the "best" people supposed to prove themselves?

That's why you see so many posts in this thread like "I worked on the perfect team for a short period, but alas, it was short lived." Building the gestalt is always going to flame out. Right at the moment when you realize you've assembled something special is when your team members start getting traded to the Yankees.

That's why the only sustainable model is culture, culture, culture. Always be striving for the perfect combination of team members. Keep the talent pipeline active, and if you have someone who doesn't fit, get them out of there.

Interesting, to belabor the sports team metaphor, culture does indeed seem to be an alternative to max salaries. C.f., Spurs being one of the most winningest teams in the past couple decades, Tom Brady taking pay cuts in order to maintain and foster a dynasty, and Steph Curry doing similar. In all these cases, a great coach (read, CEO) has fostered a strict but burgeoning cultures which rewards its members through a shared success in what they love and not an individual's compensation.

Of course this analogy can be dangerous. Athlete's get paid extraordinary sums, even while sitting on a bench 99% of the season. Whereas many employees have to worry about their own compensation simply to survive, and such selling of culture as a supplement to compensation won't always put food in your mouth (and thus can be exploitative).

Nonetheless, I appreciate your comment, you brought up an interesting view on workplace culture that I never thought about.

One metaphor I've been kicking around lately is that a long-term sustainable software team needs a mix of "gas" (folks whose instincts are to churn ahead -- "move fast and break things", if you will) and "brakes" (folks whose instincts are to keep things under control, slow down and do things "right"), and an organization that knows how to balance the two.

I remember when michaelochurch was on HN. I think even nostrademons admitted that closed allocation sucked.

I think we need to be critical of the Moneyball narrative - and not just because Oakland hasn't actually performed that well under the MB regime ...

The understood narrative is that dummies were choosing players based on broken heuristics and that proper, sober heuristics would uncover hidden value.

Put another way: If you remove human decision-making and replace it with algorithms you will achieve better results.

I am skeptical of that conclusion in every other realm and I don't think baseball (or hiring, as is the case in this article) should be exempt from that skepticism.

Having read MB the book (and seen the movie, for what that's worth) I wonder if Billy Beane is missing the potential for deeper insights that only human brains can put together. It is the thing that we are good at, after all ...

> Having read MB the book (and seen the movie, for what that's worth) I wonder if Billy Beane is missing the potential for deeper insights that only human brains can put together. It is the thing that we are good at, after all ...

The human brain is also very good at misleading itself. Statistics is how you avoid that.

Do remember that Moneyball happened in an era where statistics was just becoming more ubiquitous. Bill James was collecting and collating statistics from a bunch of volunteers who used to snail mail him the scoring sheets from the games. He only quit doing that in 1988 because suddenly there was a flood of statistics.

So, from weak data to Moneyball in less than 10 years is actually fairly impressive.

>So, from weak data to Moneyball in less than 10 years is actually fairly impressive.

And it's taken over every sport thereafter: soccer, hockey, football and basketball.

>and not just because Oakland hasn't actually performed that well under the MB regime

That can't be used as any form of argument. The ideas behind Moneyball revolutionized baseball. However the financial challenges that Oakland had in the book are still exactly the same today. The reason for their lack of ultimate success is that they gave away their first mover advantage by allowing that book to be written. Now every other team has adopted the Moneyball mindset and Oakland is back to being a cash strapped team with little competitive advantage. Even considering that, they have still been one of the more successful teams when compared to their financial peers.

> If you remove human decision-making and replace it with algorithms you will achieve better results.

Not really. We're not talking about computers inventing algorithms from raw data. People typically invent the heuristics and algorithms, so it's still people being clever and creative. They just have computers codify their decisions (i.e., code).

In other words, a spreadsheet didn't come up with the OBP correlation. A spreadsheet was just used to do the analysis more efficiently.

Reminds me of this great article from last year about a couple of stats nerds who started managing a minor league baseball team using sabremetrics:


This reminds me a lot of Dan Luu's article, in which he makes the same "Moneyball" comparison:


> Software Development is often about human factors as much as it is the code.

Totally true that. But the title almost suggests you can build a winning team out of a random bunch of average developers.

Although the best developer is an illusion as it doesn't exist as such, just 'building teams' doesn't really say much to me too. I've seen very terrible teams that should have been replaced with just 1 good developer..

Personally I prefer to hire the best fit for the job, whether he or she will be part of a team or not.

Thinking aloud...I wonder if the ancient Greek methods of sortition and ostracism have any place in building teams at companies.

Obviously election (selecting people from a group) is hit-or-miss anyway...so maybe reverse-election (eliminating people from a group) could help.

As a kid, one of the comic books I loved was the sci-fi Scrameustache [1]. One of the recurring characters was a society of aliens called "Galaxians", who's society was based on extreme sortition. Every week (or day?), they traded posts. One would be a squad leader one day and the janitor the next.

Even as a kid, I realized this was a flawed concept, because it assumed experience and knowledge were fungible. Still, it was fun to think about...

[1] https://en.wikipedia.org/wiki/Scrameustache [2] http://www.bedetheque.com/serie-682-BD-Scrameustache.html

Seth Godin actually talks about this in his book the dip where he mentions that the best means the best in their world (he goes on to explain the concept of their world). Highly recommend the read.

Thanks for the recommendation, I'll add it to my reading list!

Excellent start to a post but really needs a follow up. I'd love to hear OP's ideas on practical steps to take in order to effectively execute his team based hiring model.

The Moneyball metaphor is useful. The "best" player in baseball (Ty Cobb) batted .366 over 24 seasons.

For every 10 attempts (at bats), he "failed" 6.5 times.

Moneyball teams are really about hiring professionals who consistently bat .300.

If each player is batting .300, then the results will compound (more RBIs and runs for the team).

Sooo I totally agree with this concept: that teamwork is more important than individual contribution; that we should focus on teams rather than metrics; and that teams are simply more resilient than other types of organizational structures.

Buuut, then the author starts talking about some corporate-bs mumbo-jumbo "skills matrix". Moneyball is about "black swans" as Nassim Taleb would put it: unexpected outcomes that defy the current knowledge base. In other words, a conclusion that we could not have reached via deduction (but only perhaps abduction).

But that goes completely against the "skills matrix" concept. The whole point of Moneyball is that the Oakland players were flawed given standard metrics -- in fact, most were (very calculated) gambles.

>The final touch is to match these against levels of mastery. A lower level of skill may be someone who is largely uninformed of the subject, a middle level may be someone who can constantly use that skill well, and a high level is someone who is a mentor and thought-leader.

Oh god, he really did miss the point.

>With your new chart you are well prepared to decide where to make an investment. Should you hire for a specific weakness? Which team members get what training? Can we run peer mentorship? Do I have unchallenged experts who have no room to grow? Your answers are straight-forward now.

Yep. Whoosh. The whole idea of a strategy like Moneyball is that answers are not straightforward. Moneyball is all about intuition vs. evidence. In the case of the A's, intuition seemed to trump evidence, but the author keeps telling us to collect more evidence (in the form of matrices, skill trees, or whatever else). Not only is that the categorically wrong approach, it's also not how you build good teams.

I wouldn't have said that Moneyball is all about intuition vs. evidence, with intuition trumping evidence. Rather the other way around: the A's acquired talent that the evidence suggested was underpriced, disrupting the traditional intuition-based scouting approach. It seems like you've got that backwards.

I really disagree, and I Podesta would also:

[1] https://www.psychologytoday.com/blog/plato-pop/201109/moneyb...

[2] http://www.datacenterknowledge.com/archives/2011/09/23/the-l...

According to Podesta: "Subjectivity ruled the day in evaluating players ... We had a completely new set of metrics that bore no resemblance to anything you’d seen. We didn’t solve baseball. But we reduced the inefficiency of our decision making."

I'm coming in only knowing what those two links say (having no experience with the Moneyball media franchise or concept), but based on those, it sounds more like DePodesta is saying that the subjectivity was involved in selecting the relevant data points, and that once those data points were selected, they were used as the basis of evaluation -- they didn't override their formula to benefit "gut feelings".

From later in [2]:

>"We turn to data as our flashlight in the cave – our guiding light,” DePodesta said. “We said ‘unless we can prove it, we’re not going to believe it.’"...

Essentially, from these articles, it seems that scouting used to be a random occurrence based on the biases (or "intuition") of the scout, and that DePodesta's "innovation" was an attempt to link that process to something measurable. This "innovation" is so obvious in other fields that it's not even mentioned, and that people who depend wholly on unjustifiable random chance deserve punishment and/or ridicule.

Like most things, it boils down to subjectivity, because while you can find ways to quantify and analyze virtually anything, there is still a human who has to make a subjective judgment call as to what quantifications are significant to achieving their end goal. DePodesta says they would make those determinations and then refuse to change the criteria over undefined feeling-based intuition. That's, uh, good, I guess?

That supports exactly what I was saying: "In the competition between intuition vs. evidence, evidence wins."

So it's not clear why you are saying "In the case of the A's, intuition seemed to trump evidence, but the author keeps telling us to collect more evidence (in the form of matrices, skill trees, or whatever else). Not only is that the categorically wrong approach, it's also not how you build good teams." (Emphasis mine.)

Evidence may or may not win (that wasn't my point), my point was that Moneyball favored intuition over evidence. Even Podesta concedes that. So using it as the example in the original article makes no sense.

> Moneyball favored intuition over evidence. Even Podesta concedes that.

Both the articles you've linked contradict you. Traditional scouting favored intuition over evidence. DePodesta and the Moneyball approach turned that on its head, favoring evidence.

> DePodesta: “We said ‘unless we can prove it, we’re not going to believe it.’"

> But that goes completely against the "skills matrix" concept. The whole point of Moneyball is that the Oakland players were flawed given standard metrics

If anything I think I'm advocating against using "standard metrics", and opting to create ones that are meaningful to what it takes to ship software in your company. Where standard metrics are the ones that result in hiring the traditionally "best" developers.

> corporate-bs mumbo-jumbo

Mumbo? Perhaps. Jumbo? Perhaps not!

These topics are orthogonal. You can hire the best and build teams at the same time.

I agree with that. I attempted to describe a way for a company to improve throughput that isn't simply "hire the best".

Boss: I only want to hire the guy who worked on XXXX Project. me: That project was 2 years late and 1.2 million over budget. Boss: Still, he did something right.

If you say "the best" that also includes being good at teamwork. Great programmers write great interfaces anyway and are highly flexible.


> So grow up HN, and stop this trend of B.S. articles about what works and what doesn't that have no associated analytical or statistical proof.

Be part of the solution. Participate in the community in a constructive, productive way. Submit articles you find substantive. Contribute good comments. Upvote submissions and comments you find useful. Flag submissions you find not useful. Each of us has a role to play in curating HN content.

I'm glad that you read the entire post and it excited you enough to create an account and participate on HN.

I'm a big fan of having data. What data would you have liked? What collection method would have satisfied you?

Registration is open for Startup School 2019. Classes start July 22nd.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact