Predictions for 2012

I have none to offer.

It may sound strange to hear me say that I don’t make predictions even though I often talk about how things will change and even provide some forecasts. The difference is one of degrees. A prediction to me is a very specific, time-sensitive and materially valuable recommendation. An observation about the future is an imprecise, intuitive hunch based on pattern recognition. It’s mushy. It’s theoretical. It’s the difference between saying a company is great and recommending to buy its stock with a price target in a time frame.

But it gets even weirder.

One of the recurring themes in Apple analysis has been that independent analysts, as a group, have been, on average, more accurate in quarterly forecasting than highly paid Wall Street Analysts. Many have asked why.  I believe that the answer is because independent analysts are theorists who make observations while professionals are data gatherers who make predictions.

First, we should understand the reasons why each group does what it does.  Wall Street sell side analysts are rewarded for making Buy/Sell recommendations that cause clients to retain the services of the firm they work for. They are not paid to be accurate with specific forecasts in fundamentals. They are paid to predict how stocks move. The objective is not to be accurate in the quarterly but to support the “call” or prediction. The disconnect from accuracy comes in trying to sustain the call with a “conservative” forecast.

For example,  an analyst with a “Buy” recommendation (and a prediction in the form of a price target) thinks Apple could ship 30 million iPhones this quarter and generate EPS of $10. But he also thinks the stock looks cheap even if they only ship 28 million iPhones and report EPS of $8.50. In this scenario he’s incentivized to publish the lower (less accurate) number. That way he supports his prediction. By dampening expectations, the stock rises on the “surprise”.

Because an analyst sells predictions, he self-censors what he believes.

Contrast that with the motivation of an independent analyst (aka blogger). His motivation is to (presumably) gain a reputation of accuracy and insight. He sells his reputation as a keen observer. Precise predictions or recommendations can turn out wrong quite often. Getting caught with these errors exposes the independent to instant negative feedback. The risk is far greater than the  opportunity.

So here is the difference: Analysts have an incentive to put forth a version of the future that supports their call on the stock. Bloggers have an incentive to put forth the most accurate version of the future. By taking the prediction out of the picture, accuracy in describing the future improves.

So I don’t have any predictions or recommendations. I just hope to help you see patterns that lead you to making your own predictions.

  • That’s the best breakdown between the “pros” and the “amateurs” that I’ve seen.  It’s so clear now that they are doing two different jobs.

    • Don Wilson

      Agreed.  In Horace’s parlance, bloggers and analysts are hired to do different jobs, even though their output is similar (earnings and sales estimates).  

      What strikes me is that AAPL remains where it is, despite the recent uptick, even though Apple’s performance has consistently outperformed those expectations (with a couple of minor hiccups).  

      • Ian Ollmann

        > What strikes me is that AAPL remains where it is, despite the recent uptick, even 
        > though Apple’s performance has consistently outperformed those expectations 
        > (with a couple of minor hiccups).

        Queue Benjamin Graham.
        Assuming Apple can keep it up, the market will eventually be forced to come around. Don’t expect it to be exuberant though.

  • Rob Scott

    Analysis is not complete without some form of prediction and recommendation. Describing patterns and obse versions is cool but not enough. You must have some skin in the game, otherwise how do we take you seriously.
    I, as an analyst I make calls, and recommendation with the goal of providing both insights and accuracy, they are not mutually exclusive.
    Horace won’t cut it as a merchant because he cannot make a call, I can.

    • Anonymous

      You’re welcome to your opinion. I don’t share it. Horace’s reputation IS his “skin in the game”, as you put it. In fact, by not being in someone else’s employ, he has a far better likelihood of achieving objectivity than a “wage slave”, no matter how well paid and respected thay may be.

    • What firm do you work for? 

    • That would be like suggesting we only listen to automobile manufacturers about petrochemical chemistry, because scientists don’t have “skin in the game”.

    • Tatil

      Horace has published his estimate for Apple’s revenue, sales volume of each product line, profits and assuming a conservative P/E ratio, offered a target price at the time of earnings call about a month before today. Considering Apple’s shares were about $90 cheaper than his target price that constitutes a recommendation to buy. He did make his call. What was your call? 

      • gbonzo

        Those sound like the exact same things that a professional analyst does. Horace claims he only does “observations about future” and professional analysts do “predictions” and the difference between these is large.

        What difference?

      • I will never make a recommendation to buy or sell. I admit to making forecasts but my purpose is entirely different. I don’t do it to help you make money but to test my hypothesis or theories of how the market/industry works.

      • gbonzo

        Of course you would not make a recommendation to buy or sell. Why would you? It is not what your audience expects from you.

        What you implicitly have already said is that Apple stock is undervalued in your opinion. A professional analyst would just formulate this as “Buy” because her audience is searching for stocks to buy or sell.

  • Anonymous

    …and then there’s the poster (like me), whose motivations vary all over the map. Some of us pride ourselves on objectivity, while others may be motivated by emotion or even purposeful dissemination of disinformation. Recently, a PR firm hired by Google was caught paying bloggers for positive blogs about Chrome. It’s not much of a jump to posit the existence of paid posters.

    By far the major reason I follow Horace’s blog is because I see in him a blogger dedicated to objective analysis, and I also see a majority of posters of like persuasion. It also helps that both blogger and posters are far above average in intelligence, but frankly, intelligence is over-rated, IMHO. You can be an Einstein and still have your head stuck where the sun don’t shine…

    • A bit off topic but…

      I was impressed by Google’s response to the paid links controversy of Chrome.

      On topic.

      I agree with your view of Asymco.  Horace offers very well reasoned opinions based on solid data.

  • gbonzo

    What about analysts who got fired when their recommendations and analysis were wrong or took too long to materialize (like Per Lindberg at Dresdner Kleinwort)? The prospect of that should provide some incentive to be correct at least for some subgroup of analysts.

    Also, the professional analysts having a bad record on one company does not mean much. It would be different if the same could be shown to be true for a large set of companies.

  • Andrew Tennikoff

    Well articulated. I’ve always seen the two parties as having fundamentally different jobs to be done.
    As regards an analysts motivation, I’d suggest an additional incentive to dampen their reported predictions: If a customer (investor) has followed the analysts “Buy” call, and the results come in higher than the analysts prediction, the customer is still delighted – perhaps even more so then if the results came in merely as predicted. Ie. the average investor, who hires the services of an analyst, cares less about absolute accuracy than buy/sell accuracy.

  • Sander van der Wal

    There is a big difference between observations and predictions. An observation is the examination of the state of a part if the observable world, a prediction is a statement about the result of an observation in the future, based on a specific theory.

    Both bloggers and analysts construct theories, but not in the same way. They both use their theories to make predictions. Calling the predictions of one group observations is not wrong, as Carroll said, but is confuses matters to no end, and we also need a new word for wat observations are commonly thought to mean.

    • Kizedek

      “An observation is the examination of the state of a part if the observable world, a prediction is a statement about the result of an observation in the future, based on a specific theory.”
      I disagree with your linkage here between observation and prediction. I agree with Horace that analysts are predicting but not observing.

      I watched Sherlock Holmes this evening on BBC’s iPlayer. Sherlock said, “The evidence was right under your nose, John. As ever, you see but do not observe.”

      Analysts see. They do not observe. Observation has to do with knowing what you are seeing. It has to do with placing things in context. And to do that, you must recognize the various contexts. Either analysts are placing the data they see in the wrong contexts, or they are totally uninterested in context.

      Horace constantly draws us into any number of contexts: products, markets, history, disruption, technology, social, art and philosophy… Therefore, it is obvious that Horace is constantly observing. When he makes an observation, it is accurate and insightful.

      If one were to draw logical conculsions from his observations, they might well come true, and in retrospect carry the weight of predictions.

      But the predictions of analysts, they are pretty much shots in the dark; when they are close, it’s by accident. They gather data, they don’t know what they are seeing, and they are not interested in the subtleties of context. It’s like consumers who crow over the spec sheets of Android phones, and don’t really get that Apple is all about what you can DO with your device.

      My conclusion is that pretty much anyone can be a Wall Street “analyst”. It seems they just gather data, stick it in a spreadsheet and run it through various pre-determined formula now and again to come up with a “prediction”. If they truly had some talent and were analysts of the romantic nature that one reads about in spy novels, the kind that thinks laterally and makes connections between seemingly unrelated pieces of data, then they would have been recruited by the CIA already.

      • Sander van der Wal

        In the field of theory verification, the meaning of both prediction and observation are well established, and it is a pity that Horace is using his own definitions. If his predictions are based on intuition, he should call them hunches, which are predictions based on intuition.

        Further, I am not surprised that analysts do not check their predictins. These people are most likely belivers in the “efficient market” theory, which states that markets are not predictable. They predict because people pay them for predictions. If theybwere paid for verifications, they would do that too, one assumes.

        And why people pay analysts for predictions these analysts do not believe in, is completely beyond me.

  • Roo_44

    Well said Horace, the independent analysis’ are rewarded for being right, the professional analysis’ are rewarded for not being wrong.  

    • it explains why bearish analysts tend to have relatively higher Rev. and EPS estimates, and bullish ones have  relatively lower estimates. This way if the bears are wrong, they can always fall back to  “it missed my (high) estimate” defense, and if the bulls are wrong, they can say “it beat my (low) estimate”. all make sense now. 

  • Pingback: AOTD: Predictions for 2012 | Dreaming Of Beetles()

  • Anonymous

    This just proves one of my axioms on life, ‘If something don’t make sense, you ain’t looking at it right.” It never made sense to me how a group of people (not just one, but many), with nothing more than an idea, a computer and the Internet, could out prognosticate Harvard MBAs with all the research and tools known to mankind. What’s wrong with this picture?

    Thanks Horace.

    • The funny part is is that it costs about $100,000 to get a Harvard or Ivy League School MBA.  Talk about return on investment.

    • Anonymous

      Isn’t Horace a Harvard MBA?

  • Stefan Sidahmed

    “observation about the future”

    I get what you are conveying, but I find all the semantics confusing and distracting from the point.  They all mean the same thing to me.  The key is whether the analyst is providing a biased or unbiased prediction/forecast/”observation of the future”. 

    Know the agenda of the analyst.  The blogger wants to get it right, the professional wants to be safe.  This is well understood.

    • Kizedek

      It might seem merely like semantics much of the time. But we have different words and phrases for a reason. Unfortunately, language tends to blur over time since the correct or more precise terms are not employed when they should be. Certainly, the connotations between the different terms differ.

      Some of the confusion or distraction can be diffused when you realize a lot of the difference in terms comes down to “intention”. What intention does the analyst or blogger place on his “pronouncement”. Horace covers this in his treatment of independent theorists and professional data-gatherers, and why each group does what it does.

      “Forecast”, and “Prediction” are much more “intentional” terms than “observation” (you might even say, “professional”). If someone puts themselves out there and “makes a prediction”, they should be held to account, just like a “prophet” (a “forth-teller” with a very specific message). Unfortunately, they don’t seem to be held to account for poor predictions.

      “Making a prediction” is as intentional as “placing a bet”. Except, the analyst is doing it with your money.

      That is not to say that “making an observation” is a cop out. It is saying exactly what Horace intimated: “look at my record and draw a conclusion about whether or not I know what I am talking about.” Whereas, “making a prediction” says, “you must listen to me now, no matter what else you hear”.

    • They do not mean the same thing to me. An observation about the future would be a statement that we are all going to die. A forecast is that we will all live to an average age of 82. A prediction is that you will die on a specific day. The remarkable thing in the areas I observe is that everyone is willing to make predictions without agreeing on any common observation of the future. That’s like telling the day of your death without believing that people, in general, die.

      • Stefan Sidahmed

        I concur with your point on the difference between a “casual observation” vs a “predictions” and “forecast”.  Observations can lead to the latter.

        Another observation you had in an earlier podcast was that the pros do not disclose their methods or assumptions, whereas some amateurs put it all out there for peer review.  It can make for some tough criticism, but how else can there be a community discussion among those who want to understand the past, present, and observations of the future in order to make sound predictions and forecasts.

        I now better understand the point of this article.

        Another inherent problem with financial predictions/forescasts:  They are
        point estimates, but any estimate has a distribution associated with
        it.  No one talks about this, probably because it adds a level of
        complexity and sophistication to the estimating process.


  • Janderson

    And this, my dear friend, is why we love you.


  • Just Iain

    Guys (and any Ladies present), you’re going all maudlin on Horace. 🙂

    Horace, I applaud your honesty and humility.


  • Martin R. Schulke

    open, honest, direct, vulnerable
    my kind of world!
    thanks Horace

  • In my work, I look to the “utility function” – what is the payout (benefit or cost) from an action (including buys or predictions) under probable outcomes? Horace has neatly characterized that predicted-v-actual EPS is a tiny part of this function. 

    I would further characterize it that as far as numeric predictions go, the function is that being on the correct side of the consensus (in recent quarters re: Apple, being above consensus) gives a modest positive benefit to the analyst. Being much higher than consensus, still closer to the actual, is a bit better, but not much. Being still higher, and actually overshooting the actual EPS mark, starts nosediving— being by yourself at $1 over actual EPS is much worse than being in the pack at $4 too low.

    Being on the wrong (recently, too conservative) side of consensus is not problematic as long as the consensus itself was too conservative, until you get ridiculously low, which will get you fired eventually.

    Those of us who use statistical techniques for forecasting recognize that “least squares” is comparable to a symmetrical penalty about the actual, with a dollar too high being the same penalty as a dollar too low, both of which are only a quarter as bad as being off by twice as much, ± $2.00. In terms of understanding what’s going on, this is a better quality measure, but as Horace explicates, it ain’t the way that professional analysts get rewarded.

    It’s not true that amateurs are immune to these distorted incentives; others have mentioned ego and those who use page hits as a revenue source have incentive to make outlandish claims, “anti-safe” predictions. John Dvorak comes to mind: somebody who has apparently had a long career of saying obvious nonsense that people can’t help themselves linking to, in order to free-ride on his infamy.

    • larry

      I have just one simple question maybe you can help a amateur (beginner) how can I buy appl stock and when is it the best time to do that if I can??????

      • Anonymous

        There are lots of online services which let you purchase and trade stock for a pretty low fee. You’ll have to do your own research to pick one you like. As for when, Monday would be a good time. Even if Apple retrenches (and it may not – that’s part of the gamble), it’s going to have a good year, IMHO.

  • Thanks Horace for the work you do. Became addicted to 5by5 podcast series, you may want to consider making a visual version of those, I mean, either video versions or presentations. Warm regards. Eduardo

  • berult

    Information theory tells us that observation effects in a fundamental way the object being observed; entropy revisited. What happens if you collect a set of data representative of moments in space and time is that you freeze the information content into a solid state. The very act of it being captured modulate information’s frequency to vibrate in unison with the steady reasoning state of the observer. Information is thus amplified, …augmented, morphing henceforth into cognitive realism.

    Intuition stems from and sits on continuously interiorized factual coordinates. A geometry that links an observer with the object of observation, extending the latter parameters, ..and perimeter, to cover the observer’s intellectual capacity to abstract patterns of idiosyncrasies out of a streaming flow of information. No cheating there… A rigorous and harmonious pas-de-deux between didactical facts and their pedagogical apprehension …right up to the bidding exercise of the sponsorship auctions, where over-the-counter bias pray steadfastly to be bought, as preyed-upon souls duck …’lame-duck’ to be sold…

    • Anonymous

      Agreed, in principle, my old bean.
      But you omitted any mention of the profound influence of the Schwanstucker Inversion Syndrome which applies in all such cases.
      Tut, tut.
      You must try to do better in future.

    • Oma5665

      That was a good read thanks you’ve inspired the artist in me to writ another song

    • Chandra2

      It took me half an hour to parse and apprehend this.. I have to look up dictionary and wikipedia for every other word 😉  Still I am not sure if I got what you are talking about. Quantum Physics, second law of thermodynamics, combination and application of all that?

      • Marcos El Malo

        Berult is disrupting the Time Cube (TM) industry.

  • Abhiraj

    Again, this isn’t any comment that will add any more value to your blog, it’s just one among the many that already populate this space…but I couldn’t resist saying “Thanks a lot Horace!”…it’s really amazing The way u analyze things and put it in front of laymen like in such a lucid and yet compelling style…wanna go on listening to u analyze and tell stories about the entire history of the computing world…

    Thanks again!


  • Anonymous

    Analysts serve their masters.
    Like politicians and their corporate paymasters.
    An analysts comments were never intended for our ears.
    Indeed I have never understood why they make their paid research findings public.

    • Anonymous

      All analysts, like all politicians, should not be grouped together. People like Horace give the lie to that oversimplification….

    • Dajhilton

      Not so sure about that analogy. If you analyze corporate donations to Congress, you will find a very large percentage of those corporate donations actually going not to politicians who allegedly ‘serve’ the interests of corporations, but to those who don’t.  It is very common for corporations, especially large Fortune 500 companies, to donate equally to both parties, and to candidates of all persuasions.  

  • Pingback: Just A Friendly Reminder: If You Sold Your Apple Stock In October, You Were, In Fact, An Idiot | Technology Magazine()

  • Pingback: Just A Friendly Reminder: If You Sold Your Apple Stock In October, You Were, In Fact, An Idiot | Apple World()

  • Anonymous

    Personally I think predictions and forecasts (and estimates) are more or less interchangeable terms and indeed in most dictionary descriptions use one or other to describe another.

    However, I find Horace’s differentiation interesting and his explanation very insightful of the different approaches, motivations and results of “professional” and “amateur’ analysts as DEF describes them.

    Personally, I regard all predictions/forecasts as interesting but unreliable estimates which are indicative. In general terms the “professionals” are the least reliable on fundamentals, as Horace’s and DEF have demonstrated.

    Having employed some very bright economist PHDs and found that economist rarely agree,  I have long preferred to use a range of predictions/forecasts similar to Kathy Huberty of Morgan Stanley. 

    I find this comment by Horace especially interesting: “They [the “professionals”] are not paid to be accurate with specific forecasts in fundamentals. They are paid to predict how stocks move.”

    Although “professionals” are ridiculously conservative in forecasting Apple’s growth prospects, their price target predictions are significantly more accurate, so perhaps we should judge them by the accuracy of their price targets, which is how they clients may judge them? 

    I used to track price forecasts in a spreadsheet until I found through one of my banks a Factset chart (which for copyright reasons I may not be allowed to reproduce) which does does almost the same, but using consensus price forecasts, saving me the trouble. 

    Looking at the Factsest data 

    In Jan. 2009 their 12 month price forecast (i.e. for Jan 2010) was about $240. In fact in Jan 2010 the actual SP ranged between $192 and $215.

    In Jan 2010 their consensus price forecast for Jan 2011 was about $400, compared to the range so far this month of $411 to $422

    Today FactSet’s  12 month consensus price target is around $510.

    However, if Apple significantly beats FY Q1 2012 consensus earnings, as most bloggers or “Amateurs” expect, then we shall probably see a significant rise to the consensus 12 month Share price target. provides a pretty comprehensive list of analysts price forecasts:

    • Anonymous

      I think your dates may be out of sync. In Jan 2011, the high was about $350/share. If they were predicting $400/share, they missed by $50/share.

      It also isn’t clear if your statement that “Today FactSet’s  12 month consensus price target is around $510” is a forecast they made a year ago for Jan 2012 or made “today” for Jan 2013. I suspect it’s the former, but could you please clarify?

      I’d also point out that in 2010, we started to see the beginnings of P/E compression on AAPL, so it makes perfect sense that we would see them “miss” to the high side both times. Two years ago, I also expected to see AAPL much higher than it is now.

      Finally, we must keep in mind that January is far from over, and that Apple’s earnings aren’t due for another two and a half weeks. I think it’s very important for AAPL’s stock price that it has hit an all-time high this far in advance of its earnings release. While I don’t expect AAPL to hit $510/share this month, it’s not completely beyond the realm of possibility. If it did, then it might be signaling that, as Andy Zaky suggested, AAPL’s P/E compression has hit a floor. If true, that would be extremely bullish for the future stock price.

      Andy Zaky also suggests that we carefully watch Apple’s revenue projections as opposed to its EPS projections. According to his analysis, these projections are always in the neighborhood of 15% under the actual revenue, while EPS is much more volatile. I’d also note that Apple has had two fiscal years in a row where revenue increased over 50% per year. In fact, last year it increased 65%. 

      If (1) P/E has hit bottom at 15, (2) revenue increases 50% once more, and (3) EPS stays steady relative to revenue, and (3) P/E has indeed hit bottom at about 15, then the price of AAPL will go up 50% a year from now.

      Lord willin’ and the creek don’t rise….

      • I believe Andy Zaky’s position is the rate of PE compression will slow soon and eventually stop, not that it has hit a floor now. A PE of 10 is still very possible and likely even when PE compresses by .5 every quarter.

      • Anonymous

        You are correct. I only meant to say that Mr. Zaky had predicted a floor was imminent. Bad phrasing on my part.

      • Anonymous


        Yes you’re right. 

        On 20 Jan 2010  Factset’s consensus  12 month share price forecast (for Jan 2011) was around $400. The actual outcome: closing  ranged between $329 to $348. So as you say, the analysts over estimated the 12 month share price target by over $50.

        The Jan 2012 Factset 12 month share price target of $510 is for January 2013.


  • Pingback: Just A Friendly Reminder: If You Sold Your Apple Stock In October, You Were, In Fact, An Idiot | Tech News Aggregator()

  • Pingback: Just A Friendly Reminder: If You Sold Your Apple Stock In October, You Were, In Fact, An Idiot | Startup Help()

  • Pingback: Just A Friendly Reminder: If You Sold Your Apple Stock In October, You Were, In Fact, An Idiot |

  • Pingback: iPhone-verkoopcijfers: de schatting van 37 Apple-analisten - iPhone -

  • for more on the difference between prediction and forecasting, see

    • I wish I had read that before writing my post, though it does seem mostly consistent with how I view the distinction between forecasting and predicting.

  • Pingback: iTuneTV » Just A Friendly Reminder: If You Sold Your Apple Stock In October, You Were, In Fact, An Idiot()

  • Anonymous

    I’m not so sure that analysts make conservative predictions in order to stay on the correct side of their recommendations.

    I mean it may be nice to make a rec to one’s clients to put the client in the right position, but isn’t it overwhelmingly embarrassing to make modelling errors on such a large scale? Their research notes include details on their complex financial models that are supposed to instill confidence in their clients on their prognosticating ability. Doesn’t this all go for naught if their model’s faux complexity comes up with numbers that are so far off? That would totally undermine their credibility.

    Maybe it’s my naïveté, but I would prefer, as a client that the analyst make as accurate a call as possible and let me decide to adjust my position accordingly.

  • Pingback: 一如既往,iPhone 销量预测分左右派 | Jaring's blog()

  • Pingback: 一如既往,iPhone 销量预测分左右派 - 黑色 | 黑色()