Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
×
Math Stats

When Graphs Are a Matter of Life and Death (newyorker.com) 122

Pie charts and scatter plots seem like ordinary tools, but they revolutionized the way we solve problems. From a report: John Carter has only an hour to decide. The most important auto race of the season is looming; it will be broadcast live on national television and could bring major prize money. If his team wins, it will get a sponsorship deal and a chance to start making some real profits for a change. There's just one problem. In seven of the past twenty-four races, the engine in the Carter Racing car has blown out. An engine failure live on TV will jeopardize sponsorships -- and the driver's life. But withdrawing has consequences, too. The wasted entry fee means finishing the season in debt, and the team won't be happy about the missed opportunity for glory. As Burns's First Law of Racing says, "Nobody ever won a race sitting in the pits."

One of the engine mechanics has a hunch about what's causing the blowouts. He thinks that the engine's head gasket might be breaking in cooler weather. To help Carter decide what to do, a graph is devised that shows the conditions during each of the blowouts: the outdoor temperature at the time of the race plotted against the number of breaks in the head gasket. The dots are scattered into a sort of crooked smile across a range of temperatures from about fifty-five degrees to seventy-five degrees. The upcoming race is forecast to be especially cold, just forty degrees, well below anything the cars have experienced before. So: race or withdraw?

This case study, based on real data, and devised by a pair of clever business professors, has been shown to students around the world for more than three decades. Most groups presented with the Carter Racing story look at the scattered dots on the graph and decide that the relationship between temperature and engine failure is inconclusive. Almost everyone chooses to race. Almost no one looks at that chart and asks to see the seventeen missing data points -- the data from those races which did not end in engine failure.

This discussion has been archived. No new comments can be posted.

When Graphs Are a Matter of Life and Death

Comments Filter:
  • Graphs (Score:5, Interesting)

    by war4peace ( 1628283 ) on Thursday June 17, 2021 @06:06AM (#61495796)

    One of my two jobs revolves around generating graphs and data tables, as well as analyzing the outputs.
    In business, at least, very few people understand them, and even fewer are making decisions based on those graphs and data tables.

    • Re:Graphs (Score:5, Interesting)

      by jellomizer ( 103300 ) on Thursday June 17, 2021 @07:24AM (#61495924)

      Often I find the person in charge is often the least able to interpret the data to make informed decisions.

      For most businesses the Owner is also the founder of the organization. They often got their company by grit, force of will and making gut decisions. This is usually perfectly fine for an organization until it reaches a particular size (usually to a point where the boss cannot manage every person in the organization (around 100 employees) After that point, you are going to need to to be able to understand data in aggregate. Then know enough to ask questions and find problems that are creating the numbers. (To de-average the data)

      For too many bosses when a company gets past being a small company, to a mid-sized company, when they get graphs and charts, they will often be happy because the line for say profit is still going up. or they will freak out when an other metric goes down. They are not accustom to seeing so much data, and knowing what is the normal variance, and how to judge shape of data.

      So if I were to see a company profits going on a log scale over time, it is still going up, I would be concerned because, or at least start asking questions, because I would hope it would be more linear. But I am not the boss, I am a data guy. I may recommend my interpretation of the data to them, however still the Boss who is often the founder, who got there with gumption and instincts may not really listen to it, and go on their own way. Until the company fails.

      • Re: (Score:2, Interesting)

        by Canberra1 ( 3475749 )
        I find politicians are blind when they see that they have to spend money now. So the reject the unpleasant. See Public Housing, Public Hospitals, New schools, new water and sewerage plants, crossing lights, and mental health. Instead they got rid of public toilets, rubbish or litter bins, sold parks and public car parks to developers, cheap. On television, like tobacco executives, they swear blind they had no idea of latent demand.
      • Re: Graphs (Score:5, Interesting)

        by PPH ( 736903 ) on Thursday June 17, 2021 @09:41AM (#61496364)

        Often I find the person in charge is often the least able to interpret the data to make informed decisions.

        Yeah. Not so much the founders of a business. But after a few generations, management tends to be populated by people with high verbal/social skills and less math/logic/visual. The smart ones recognize their shortcomings and keep a few math geeks close by on their staff. Not so smart tend to get corralled onto the golf course by the court jesters and steered by politics.

      • Re:Graphs (Score:4, Funny)

        by chispito ( 1870390 ) on Thursday June 17, 2021 @12:15PM (#61496866)

        Often I find the person in charge is often the least able to interpret the data to make informed decisions

        Something tells me that's not exactly a controversial stance hereabouts.

    • You want to see pain, go to reddit's wallstreetbets community and look at the lines people draw on graphs there. It'd be hilarious if it wasn't for all the people throwing money at stuff based on them.

    • Re: Graphs (Score:4, Insightful)

      by reanjr ( 588767 ) on Thursday June 17, 2021 @02:38PM (#61497272) Homepage

      "Look, I can graph this data, but it would be helpful to know what you want it to show so we don't waste time trying to present objective truth"

    • In Business, many decisions are made by marketing/sales. Do you SERIOUSLY expect any of that lot to know how to review data and make an intelligent decision?
  • Hannah Fry is an Associate Professor in the Mathematics of Cities. And goddam gorgeous https://hannahfry.co.uk/ [hannahfry.co.uk]
    • You're so stuck on looks you forgot to mention she's also an all-around badass as decided by Hanna Fry, who is a Mathematician, science presenter so you know it's true.

    • by MagicM ( 85041 )

      She has a number of videos on Numberphile. Entertaining and educational.

      https://www.youtube.com/playli... [youtube.com]

  • News for Nerds (Score:4, Interesting)

    by MancunianMaskMan ( 701642 ) on Thursday June 17, 2021 @06:11AM (#61495800)
    the news for nerds is not that "Statistics is an exact science/helpful in many fields to guide decisions" we knew that already.

    The real news for nerds is "Hannah Fry landed a New Yorker Article", congrats. Pity about the pay (or spam subscription) wall, I didn't get much beyond the author name and the TFS copypaste.

    • the news for nerds is not that "Statistics is an exact science/helpful in many fields to guide decisions" we knew that already.

      Have you seen the level of discourse on Slashdot recently? I think it's about time we get back to basics. The nerd is dead.

      • Guess they moved to Reddit?
      • Re: (Score:2, Interesting)

        by jellomizer ( 103300 )

        Exactly,
        Lets get back to the esoteric stuff that other news sources don't cover.
        The general news, (especially Social Media and Cable based sources) are based on entertaining, not informing, the people. So when a news article reaches their desk they are going to spin it to a point where we are forced to get an emotional reaction from it, then their competing news source is going to spin the same thing to get an opposite emotional reaction from it.

        If the topic isn't covered by these entertainment "news" site

  • Yep (Score:5, Interesting)

    by Anonymous Coward on Thursday June 17, 2021 @06:20AM (#61495814)

    15 year ago, we had this case study on the first day of an MBA class. My first thought: don't race, the risk / reward tradeoffs are bad. Second thought: hey this sounds like what happened to Challenger. Most of us with an engineering background had the same reaction.

    Among non-engineering students, not only did most fail to spot the connection. Most students were like "Wait, what happened to Challenger?"

    • Re:Yep (Score:5, Interesting)

      by garyisabusyguy ( 732330 ) on Thursday June 17, 2021 @09:59AM (#61496426)

      For about a decade I worked with blood manufacturing software with a large blood bank

      We used six sigma tools to monitor, identify and correct process failures because a mistake in any step of the donation, processing and delivery steps can KILL people

      In order to do this we had to employ strict documentation and 'honesty' standards because statistical analysis is shit if people are not documenting problems or lying to you about them

      That is where most companies fail when trying to use statistical analysis for themselves. GIGO Garbage In Garbage Out

      imo, most organizatios

      • oopsy

        imo, most organizations fail to collect good data, so analysis is unlikely to produce usable results

      • That is where most companies fail when trying to use statistical analysis for themselves. GIGO Garbage In Garbage Out

        This is why I despise telemetry. It's not just a privacy issue. I refuse to believe it actually works, and companies just slurp up data because they can, it makes for nice reports, and gives managers a higher sense of knowledge and control then they really have.

    • Maybe I'm just generally distrustful of the level of competence and honesty in graphs, but the first thing I do with a graph is parse the exact language and labels, check how the graph is framed (start/end date framing in economics, for example), and generally try to tear it apart.

      Most people don't understand math or statistics well enough to create objective graphs. And I know full well it's usually going to be a graphics artist and not a statistician.

      Graphs are in news stories primarily for morons who can

      • by BranMan ( 29917 )

        Ha ha! Me too - in fact if I see any graph or chart of numerical data the does not include the 0,0 point (i.e. only showing from X1 -> X2 and Y1 -> Y2) I immediately assume it is trying to lie to me, and proceed from there.

        Sounds like you have the same reaction.

    • by Tablizer ( 95088 )

      Most [MBA] students were like "Wait, what happened to Challenger?"

      And after they learned probably thought, "Gee, they couldn't find a way to blame it on somebody else."

  • Why not just assign _The Challenger Launch Decision_? My guess being that it is a 400 page book with footnotes and MBA candidates prefer everything in a simple chart, without the nuance of contributing factors such as organizational psychology, personal ambition, "don't bring me problems" cultures, etc.

    • by Megane ( 129182 )
      Or just read the chapter about it in the Feynman biography "Genius".
    • by Rei ( 128717 )

      More to the point, their "blowout" scenario actually sounds like they're describing the Challenger Launch Decision, just thinly disguised.

    • by CaptainOfSpray ( 1229754 ) on Thursday June 17, 2021 @07:08AM (#61495888)
      The best reason for doing it this way is to give the students a smack-in-the-face experience.

      "Congratulations, you just killed seven people - because you blindly accepted incomplete data".
    • by bws111 ( 1216812 ) on Thursday June 17, 2021 @08:53AM (#61496196)

      It obviously IS the Challenger launch decision. But if you called it that, then gave students the data and asked for fly/no-fly, they would have to be complete morons to select 'fly', whether they understood the problems with the data or not. So they fictionalize it so the outcome is not already known.

    • The Challenger disaster, as Feynman discovered, was more due to a culture discrepancy between the political management and the pragmatic engineers. Concerning the notorious O-Rings, the upper management was operating under a distorted notion of the term "safety factor" where, in fact, there was none.

      Alas, what you describe is closer to what happened with Columbia. Here we have a case where the problem was known and investigated, but the ball was dropped at the last critical minute of decision-making where t

  • Only the failures? (Score:5, Insightful)

    by quonset ( 4839537 ) on Thursday June 17, 2021 @06:26AM (#61495828)

    This sounds like the opposite of Survivorship Bias [worldwarwings.com].

    • by alanw ( 1822 ) <alan@wylie.me.uk> on Thursday June 17, 2021 @07:09AM (#61495890) Homepage

      +1
      You beat me to it. I'll just mention Abraham Wald [wikipedia.org], who worked on this.

    • The main difference is having all the data vs missing data. In the case that is cited in the story, they have the temperature readings until failure and they have the parts to inspect after failure. The survivorship bias uniquely demonstrates the problem with how to treat missing data points.
    • by sfcat ( 872532 )
      It is actually a type of sampling bias [formpl.us]. BTW, this same statistical issue is why the COVID numbers are so wonky and why your stats professor always added the bit about randomly sampled data. If the data isn't randomly sampled, then you likely have a bad analysis due to sampling bias.
  • by Musical_Joe ( 1565075 ) on Thursday June 17, 2021 @06:39AM (#61495844)

    I find it odd that the headline is about how pie charts are now being used in a "matter of life and death", with the example given being data presented in the context of racing cars. It's almost as if the writer of the article doesn't know that pie charts - specifically pie charts - were popularised by (if not invented by) Florence Nightingale; she decided it was the easiest way to communicate her statistical analysis of injuries, deaths and recovery rates in the Crimean War. Surely that *original usage* was far more a matter of life and death than the example given of how we are "now" using pie charts to determine how much money a company will make from a potentially dangerous event...

    • by Entrope ( 68843 ) on Thursday June 17, 2021 @06:58AM (#61495870) Homepage

      If you RTFA (I know, I know) you will find that the "race car engine" failure data were actually from Space Shuttle O-rings, and that the first data set graphed -- showing when failures occurred, but not successes -- was presented (in tabular form) the night before the Challenger launched.

      The Fine Summary gives a terrible sense of how the article begins.

      • by burtosis ( 1124179 ) on Thursday June 17, 2021 @07:29AM (#61495934)
        Pie chart? I’m gluten intolerant you insensitive clod! Now let’s launch this puppy!
        • Pie chart? I’m gluten intolerant you insensitive clod! Now let’s launch this puppy!

          Damn - where are those mod points when I need them? LMAO!

      • I was about to say. As I was reading that, my mind immediately went to Challenger’s O-rings, but I figured that it must have just been a coincidental similarity. I apparently haven’t had enough caffeine yet this morning to realize it was an intentional similarity formulated around that same time.

      • If you RTFA (I know, I know) you will find that the "race car engine" failure data were actually from Space Shuttle O-rings, and that the first data set graphed -- showing when failures occurred, but not successes -- was presented (in tabular form) the night before the Challenger launched.

        The Fine Summary gives a terrible sense of how the article begins.

        I did not, but it seemed impossible there existed an engine gasket issue where it went bad before the engine started because it was cold. And that this was never detected, seemed odd in the extreme. Knowing this now it makes more sense. There were only a handful of data points, not hundreds as in racing car high end machinery, or millions in street cars.

      • I RTFA, but did a search for the use case anyway.

        I found this slide presentation.
        https://slideplayer.com/slide/... [slideplayer.com]
      • I was wondering what was up with that summary. When have you heard about a car racing team skipping an event because they're afraid their car might blow up? I though that was the whole point of racing.

      • Paywall, dude. The only thing worse than a bad summary is a bad summary of an article you can't read.

        Personally, the only reason why I still visit Slashdot is because of the commentary, but if this place insists on making bad summaries of paywalled articles, even the commentary won't be worth reading anymore.

    • Florence Nightingale developed polar area diagrams(see https://en.wikipedia.org/wiki/... [wikipedia.org]) they are similar to pie charts but more sophisticated. She did not invent the polar area diagram but made excellent use of it.
    • by Jamlad ( 3436419 )
      Pie charts have been one of my least trusted forms of data representation for years. It's too easy to fudge- humans are notoriously bad a judging angles, areas, or arcs. Pie charts are easy to subvert to sell a narrative because of this. Linear XY charts have their flaws too (misleading scales, offsets, truncation, etc.) but it's a bit more apparent.
  • by geekmux ( 1040042 ) on Thursday June 17, 2021 @07:21AM (#61495916)

    Every other racer is going to have to endure the same weather conditions.

    And who the hell, gets into racing, to find guaranteed profit? With the amount of money they blow out the tailpipe, it's a business expense and tax writeoff more than anything. You should have deep pockets to even dream of this hobby.

    Go big or go home.

    • How do you make a small fortune in racing? Start with a large fortune.
    • That's why the race car analogy was stupid. The interpretation of statistics differs depending on what kind of problem you're trying to solve.
  • This sounds exactly like the story of the graph of the o ring issues on the USA's space shuttle, even down to the same sort of problem. Is this one of those names changed to protect the innocent kind of thing?
    • Damn, I came too late to make the same comment. So at least let me link the chart on page 45 (printed) of this PDF file [williamwolff.org].
      • Worse, I realized later that the summary doesn't mention the article actually is about the space shuttle incident and for some reason they tell the race car story at first instead of the real story.
        • Which is kind of funny, because I didn't initially read the article since I have zero interest in race cars. A wonderful way to write a misleading summary that will make me not read an article!
    • by bws111 ( 1216812 ) on Thursday June 17, 2021 @09:01AM (#61496236)

      It is the Challenger. Not changed 'to protect the innocent', but changed so that the students being asked to look at the data don't already know the outcome.

      • by codrus ( 35604 )

        Problem is that changing the context changes the underlying assumptions. It is very unusual for a race car engine failure to risk anyone's life -- unlike a spacecraft a car will just coast to a stop when the engine fails. The tradeoffs and risks aren't even close to being the same.

        • by bws111 ( 1216812 )

          They state that the engine failure DOES risk the driver's life (and they also throw in a financial consideration), so that 'assumption' should not be made. And in any case, how does that change the fact that you are looking at incomplete data and shouldn't make a decision based on that?

          • But everything in auto racing does risk the driver's life.

            I remember this little exercise from a business class. When asked specifically about other data points (exactly the point the article is making) and quantification of the additional risk, the professor told us that this is all that we had available, then was smug when we made the "wrong" choice.

            • by bws111 ( 1216812 )

              I'm confused. You're saying you didn't have enough information, and you realized that, and you still chose the riskiest option?

              • As I said, everything about auto racing is risky, so given that the problem summary details how the livelihood of the whole race team was at stake and that we couldn't assume that this was significantly riskier than what he would face otherwise, we chose the "wrong" answer.
  • by kmahan ( 80459 ) on Thursday June 17, 2021 @08:01AM (#61496038)

    The example I like to use is that of the mathematician Abraham Wald and his work on armoring planes. He asked a similar question. Before him planes that made it back were looked at and where the bullet holes were most frequent that's where the armor went. His method was "Those planes made it back. Let's think about the planes that didn't -- did they get shot where the bullets holes aren't on the returning planes?"

    https://www.wearethemighty.com... [wearethemighty.com]

    • Re: (Score:3, Interesting)

      by Anonymous Coward
      During World War I, British soldiers were being killed by the rocks raining down on them after an artillery shell exploded nearby. British high command decided to issue the men helmets. But then a few weeks later they noticed that there was a big increase in the number of men being hospitalized - so they told the men to stop wearing the helmets. Men refused, because they knew the guys going to the hospital would've been killed and buried on the battlefield without the helmets; instead, they were going to th
  • Fight Club where he beat him self up in the bosses office.

    Ok to not tell any one about what I know my new job will be to just collect pay checks and not tell any one.

  • Well, duh... (Score:4, Insightful)

    by TomGreenhaw ( 929233 ) on Thursday June 17, 2021 @08:36AM (#61496140)
    The human visual system is highly effective at recognizing patterns. Its obvious that graphs and data visualization exploit this capability. This isn't news at all.

    Contemporary machine learning technologies, e.g. gradient descent & linear and logistical regression as well as normal distribution and standard deviation of datasets that highlights important patterns of vast amounts of information are much more interesting.

    Of course they are only as good as the quality and quantity of the underlying data.
  • have access to a graphing program there are other ways of understanding data that works (in some cases).

    https://www.nytimes.com/2021/0... [nytimes.com]

    In the NYTimes article "Who's Afraid of Big Numbers?" (sorry if there's a paywall), another way to understand (big) numbers is by scaling them down to "normal" (not Jeff Bezos) size. For example by setting the total U.S. revenue at $100,000 and all other numbers appropriately you can see that, for example, the Department of Defense spends $17,310 whereas NASA gets $549 (no political commentary there at all folks :).

    Despite the fact that Graphs are making me(!) money because of my investment in a successful bioinformatics data visualization company, I have to admit that there are times when, as the NYTimes article shows, they are not always the best way of understanding data. Using the above as an example, if you were to plot how much money was spent on school meal programs ($5) it would be a tiny tiny sliver that you wouldn't even be able to see, whereas by presenting the data in terms of a household budget you could at least understand it.

    Still, Graphing data can be incredibly useful and, if I may be so bold, is sometimes beautiful. May I refer you to a classic? Edward Tufte's "The Visual Display of Quantitative Information". Highly recommended :)

  • you can put any old crap in agprah without being fires - so long as it's not the truth or scientificas that hurts their feelings.
  • Pie charts have a lot of inherent issues. In many cases a bar chart is a much better choice, though pie charts can be useful if certain conditions are met.

    For more insight into why pie charts are often a poor visualization choice, read Steve Wexler's recent book, The Big Picture.

  • Life and death pie charts are a bad idea. We have a hard time comparing circular segments. Its a well known limitation. https://www.data-to-viz.com/ca... [data-to-viz.com]
  • For an article extolling the virtues of plots and visual representation of data, the article is devoid of significant plots. Not even number of plots over time in New Yorker magazine.

    This article is meant for humanity majors who like reading large blocks of text. There are much better articles and books that talk vividly about visual representation of large data sets, multi dimensional data etc.

  • The entire 'Carter' study centered on a bad graph. It could have shown both, with red being blow outs and green being good outcomes.

    The major problem is caused not by bad graph reading, but by bad graph making. Too much work is put into making it simple and pretty, not enough into making it valuable.

    My personal pet peeve is the lack of zero-axiing. If you line/mountain graph something with a high number, people tend to start the graph near the number, rather than at 0. This makes it easier to see the s

  • Lies, damned lies, and statistics

  • by dcw3 ( 649211 ) on Thursday June 17, 2021 @09:31AM (#61496326) Journal

    Before I retired a couple years ago, one of my functions was to prepare a monthly report with about 25 slides showing metrics for upper management. When I'd present these (as I was required), I'd occasionally get a question or two, but almost noting ever of any significance. Note that it typically took 15-20 hours of my own preparation, along with input gathered from a dozen customer sites where our employees had to typically spend a day gathering their data to send to me.

    I made the case on several occasions to do away with this process, and was poo-pooed each time, even when mentioning the cost in time/money for something that wasn't used. On the plus side, I got to work from home when doing this, and it kept me employed through the recession.

    • Before I retired a couple years ago, one of my functions was to prepare a monthly report with about 25 slides showing metrics for upper management. When I'd present these (as I was required), I'd occasionally get a question or two, but almost noting ever of any significance. Note that it typically took 15-20 hours of my own preparation

      (emphasis added)

      You were overdue for retirement, then. Automate the process FFS! I wrote R-code in less than your 15-20 hours to take the raw data and generate a full powerpoint report deck with one click. You can do the same using MATLAB toolboxes or even roll-your-own executables based on MsoftOffice APIs.

      • by dcw3 ( 649211 ) on Thursday June 17, 2021 @12:33PM (#61496930) Journal

        You're making an assumption that I didn't try to. I left out what I thought were unnecessary details.

        This is part of the problem when you work for a F500 bureaucracy...funding, agreement to allow it to be done (it's the customer's data), etc. Our program's chief engineer had promised me he would get it worked on for well over a year...nothing. I also tried to offload the task to three others over the years, spending untold hours training each of them. By the time I'd punched out, we had gotten about half of the task semi automated with Excel macros, and then not long after I departed they stopped doing the job because nobody else could/would do it...it was not only a pain in the ass, it was thankless...honestly one of the worst tasks I had in 37 years with that company. I had ~50 people to manage at four customer locations, but couldn't give any of them the tasking because they were funded directly by customers, while this was for internal use. There's plenty more to the story, but I'm sure you'll find some other reason to fault me.

  • by mtaht ( 603670 ) on Thursday June 17, 2021 @09:34AM (#61496334) Homepage
    If there is any one event that shaped my life as an engineer, it was that. I still have nightmares about it every few months [blogspot.com], still see the explosion in my head - and every time I get stressed out I end up playing the song I wrote about it, ("Rhysling and me") over and over again, until I can get up and go back to work.

    This studio version of the song [taht.net] includes samples from richard feynman's famous demonstration of what had gone physically wrong and much more

    In the song I firmly put the blame for the disaster on the management types that didn't understand science, and were too busy posing for the cameras or trying to please their bosses to notice what feynman and one brave engineer [wikipedia.org] were trying to say.

    Over the last decade...

    In working so hard [blogspot.com] to avert a worldwide bufferbloat congestion collapse disaster [bufferbloa...beyond.net] across the internet, and knowing damn full well that how the internet actually works is still so persistently mis-understood [apnic.net], in particular, I've pulled out all the stops to attempt to kill the L4S vs RFC3168 disaster still unfolding in the ietf, and I'd laid down another version [youtube.com] after successfully - at least temporarily - halting it, 2+ years ago. L4S is still alive and going into last call at the ietf, despite all the evidence gathered against it [github.com] to date, backed by some of the biggest firms in the industry.

    I don't know if it's possible to "enjoy" this song, but when you've tried writing books, code, presentations, and screaming at the top folk running the internet to "do something", what else can you do? Have a listen.
    • Have you considered a mix of MDMA and psilocybin?

      Perhaps in a relaxed setting, like a mountain music festival... it does wonders for ptsd

      Gave the song a listen, you really captured the pain/angst of an event that I remembered as signaling the end of an era, or at least my own optimistic ideas for the future

      so yeah, find a music festival, make sure you have a lot of water and take a hippie flip, you will find comfort

    • In the song I firmly put the blame for the disaster on the management types that didn't understand science, and were too busy posing for the cameras or trying to please their bosses to notice what feynman and one brave engineer [wikipedia.org] were trying to say.

      Feynman wasnt brought in until after the disaster, and he wasnt reporting to NASA he was reporting to Congress.

      They brought Feynman in not to get to the bottom of it, but to explain the mess to dumb Congressmen, because Feynman was the Greatest Teacher To Ever Live. The fact that Feynman figured out exactly what went wrong himself is just icing on the cake. Brilliant person. Its too bad you didnt do any research.

  • Figures do not lie, but liars figure.
  • It's called the 'Quad Chart' - a daily time series of heart rate, blood pressure, respiratory rate, and patient temperature plotted together. It was developed by William Halsted at Johns Hopkins in the late 1800s. Makes it very easy to note the temporal relationship of changes in physiology, which in turn drives diagnosis.
  • Honestly, I'm a bit surprised that most people would choose to race. Without even seeing the graph, let alone one supplemented with additional relevant data, just based on the text description here of the general shape, it already sounded obvious that lower temperatures meant higher rates of failure.
  • Engine failure "will jeopardize sponsorships -- and the driver's life"

    I mean, who would risk sponsorships?

  • If you look at the graph with the missing data, all failures occur below 65F but there are so few attempts that one could argue that it's still not compelling. Certainly with the missing data points, one is less likely to launch/race, but the failures at 75F can't be discounted. Given the high-stakes one could argue don't launch at less than 75F and that would be reasonable. But what about a launch at 70F?
    • One data point is likely NOT conclusive, but it points you in a direction to investigate

      Let's say that you were able to interview a person involved in that single data point, and they revealed that the engineering staff had doubted that it would work in the first place, was outside of design specs and therefore not necessary to take additional datapoints in that range

      So yea, gotta place Sherlock sometimes

  • Sounds like the story of aeroplanes during WWII.

    People were looking at where the ones that got back had been shot and so decided to add more armour there.

    Until someone pointed out that these had got back - so it was the ones that were shot elsewhere on the fuselage that had been shot down and hence the place to increase the armour was where those that returned had not been shot.

Top Ten Things Overheard At The ANSI C Draft Committee Meetings: (3) Ha, ha, I can't believe they're actually going to adopt this sucker.

Working...