📖 ZKIZ Archives

Trading Equity Curve Consilient_Lollapalooza

http://blog.frankyfan.com/2012/09/trading-equity-curve.html
之前談的Money Management，都要主談要根據個別交易去做的，而每個交易的Capital At Risk係預先定好的。

如果推而廣之，其實每個交易的ＣＡＲ，也可以根據組合的回報去更改的，這便是Trading the Equity Curve了：羸就谷，輸就縮。

例如
如果Equity Curve 200ma之上, 就用牛市的CAR, 可能是2%, 如果是低於200ma, 就用1% CAR.


交易風險受限
組合風險受限
無後顧之憂，交易時就可以得心應手了。

S型生命成長曲線 （Throw Your Life a Curve）

http://xueqiu.com/3773003357/23043131
轉載自: HBR。
原文鏈接：http://blogs.hbr.org/johnson/2012/09/throw-your-life-a-curve.html

導讀：探討企業成長規律和人生規律的好文。 對我發展自己的成長企業投資模型和對未來的規劃幫助很大。 （終於找到中文版本的了）　


中文版本：

我們的世界觀是由我們的個人準則為支撐的：觀察構成我們個人的社交體系的各要素（包括人）間的相互作用，並尋找接下來將會發生的事情的預測方法。當體系間體現為線性行為關係且會立即作出反應時，我們的預測會相當的準確。學走步的小孩之所以善於發現燈的開關，是因為開與關的因果關係很直接。小孩一按開關，燈就會立即亮起來。然而，在存在時間遞延或非線性關係時，我們的預測力就會驟然下降。例如，股價下跌時卻實現了比預期還要高的收益，CEO一定會覺得很奇怪。

瞭解一下我的合著人，在麻省理工學院受過戰略工程師培訓、作為新興企業及財富500強企業諮詢師的蒙德斯格雷西亞（Juan Carlos Méndez-García）。據蒙德斯格雷西亞稱，理解非線性世界的最好模型是S型曲線（S-Curve）。我們已經應用該模型來理解顛覆性創新的擴散，而且我們預測可用它來理解人格分裂——我們職業路徑中的必經點。

在像企業或人腦這樣的複雜體系中，因果關係通常不會像開關與燈泡之間的關係那樣明顯。由於存在時間的延遲關係和依賴關係，即使是大量的投入在近期內所產生的收益抑或微乎甚微，或今天的高產出也許就是長期以來的行動結果。S型曲線通過沿途路標來解釋這樣的系統。我們的前提是，那些能夠成功地以這些逐級學習循環與該S型學習曲線為指導甚至加以應用的人，他們將在這種人格分裂（亦稱「自我顛覆」）的時代裡茁壯成長。

查看原圖
在我們面臨新領域專業技能的培養、提升個人學習曲線時，起初的進步會很慢。但我們可通過刻苦訓練而獲得一種牽引力，牽引我們進入一個良性循環，從而推動我們進入能力提升和自信心提升的最有效點。然而，當我們到達精通階段時，惡性循環開始出現：我們做的事情越平常，我們對學習成果「感覺良好」所持的欣賞態度就越低。這兩種循環（良性與惡性循環）構成了上述的S型曲線。

查看原圖
有關S型曲線如何才能幫助我們更好地預測未來，高爾夫運動員丹.麥可勞林（Dan McLaughlin）的經歷就是一個鐵錚錚的例子。2010年4月，從未打過18洞高爾夫球的麥可勞林辭掉了他商業攝影師的工作。經過1萬個小時的刻苦訓練後，終於實現了做一名頂級職業高爾夫球員的目標。在最初的18個月訓練中，他放球、切球、發球的進步很慢。後來，他將各個環節整合、連貫在一起，訓練速度得到了提高並很快進入高速增長期。不過對於他是如何迅速克服訓練障礙的，他未做任何記錄，為此我們很難對他的訓練過程做出相應的S型曲線。他僅僅花了28個月就實現他的計劃。而據美國高爾夫協會數據統計，近2600萬球員在訓練時都會遇到過類似障礙，而麥克勞林克服訓練障礙的能力卻超出了其中91%的球員。

正如我們在學習新知識時，掌握S型曲線可能會讓我們陷入灰心的困境，但它也可幫助我們明白為什麼一旦達到某個高度、處於停滯階段時我們會感到厭倦。當我們達到精通階段時，我們的學習速度開始減慢，而當做事得心應手就意味著有能力實現時，這還意味著我們大腦裡感覺良好的神經介質在減少，興奮的動因已經結束。

查看原圖
當我們的學習到達一定的頂峰時，我們應該不會跳到新的學習曲線，而實際上也許是在迅速下降。但這未必就是指經濟衰落，而是指我們的情志與社交健康將會受到衝擊。企業創新工廠（Business Innovation Factory）的主要推動人索爾.卡普蘭（Saul Kaplan）曾說：「我這一生一直在追求陡峭的學習型曲線，因為只有這樣我才能全力以赴的去工作。當我全力以赴的去工作時，金錢和地位通常也就成了水到渠成的事了」。或者用詹姆斯.歐沃斯（James Allworth）的理解就是，「史蒂夫.喬布斯解決了創新者進退兩難的窘境，因為他關心的不是利潤，而是產品的越來越好」。那麼，請忘記追求利潤的巔峰吧：追求和放大學習型曲線的範疇。

S型曲線構思模型是針對個人分裂而提出的前所未有的實例。面對線性問題時我們也許是預測未來的數學專家，但問題是商業和生活問題都不是線性問題，而我們人腦最終需要的甚至是必需的東西是不可預知的多巴胺。更重要的是，由於我們是生活在一個日益曲折多變的世界裡，那麼能夠甩開競爭的最佳曲線就是你從某一曲線躍遷至下一曲線的能力。

英文原文：

Our view of the world is powered by personal algorithms: observing how all of the component pieces (and people) that make up our personal social system interact, and looking for patterns to predict what will happen next.  When systems behave linearly and react immediately, we tend to be fairly accurate with our forecasts. This is why toddlers love discovering light switches: cause and effect are immediate. The child flips the switch, and on goes the light.  But our predictive power plummets when there is a time delay or non-linearity, as in the case of a CEO who delivers better-than-expected earnings only to wonder at a drop in the stock price.

Enter my co-author, MIT-trained strategist and engineer Juan Carlos Méndez-García, who consults with both start-ups and Fortune 500 companies.  According to Méndez-García, one of the best models for making sense of a non-linear world is the S-curve, the model we have used to understand the diffusion of disruptive innovations, and which he and I speculate can be used to understand personal disruption — the necessary pivots in our own career paths.

In complex systems like a business (or a brain), cause and effect may not always be as clear as the relationship between the light switch and the light bulb. There are time-delayed and time-dependent relationships in which huge effort may yield little in the near-term, or in which high output today may be the result of actions taken a long time ago. The S-curve decodes these systems by providing signposts along a path that, while frequently trod, is not always evident. Our hypothesis is that those who can successfully navigate, even harness, the successive cycles of learning and maxing out that resemble the S-curve will thrive in this era of personal disruption.

Let's do a quick review. According to the theory of the diffusion of innovations — an attempt to understand how, why and at what rate ideas and technology spread throughout cultures — diffusion or adoption is relatively slow at the outset until a tipping point is reached. Then you enter hypergrowth, which typically happens somewhere between 10-15% of market penetration. Saturation is reached at 90%+.

With Facebook for example, assuming an estimated market opportunity of one billion, it took roughly 4 years to reach penetration of 10%.  Once Facebook reached a critical mass of a hundred million users, hypergrowth kicked in due to the network effect (i.e. friends and family were now on Facebook), as well as virality (email updates, photo albums for friends of friends, etc.).  Though we could quibble, depending on our inputs, over when Facebook will reach saturation, there is no question the rate of growth has begun to slow and is now limited, if for no other reason, by the number of people who can access the service.  (Here's some more on Méndez-García's Facebook and S-curve math.)
查看原圖


As we look to develop competence within a new domain of expertise, moving up a personal learning curve, initially progress is slow.  But through deliberate practice, we gain traction, entering into a virtuous cycle that propels us into a sweet spot of accelerating competence and confidence.  Then, as we approach mastery, the vicious cycle commences:  the more habitual what we are doing becomes, the less we enjoy the "feel good" effects of learning:  these two cycles constitute the S-curve.

查看原圖

One anecdotal example of how the S-curve model can help us better predict the future is the experience of golfer Dan McLaughlin.  Never having played 18 holes of golf, in April 2010, McLaughlin quit his job as a commercial photographer to pursue a goal of becoming a top professional golfer through 10,000 hours of deliberate practice.  During the first 18 months, improvement was slow as McLaughlin first practiced his putting, chipping, and his drive. Then, as he began to put the various pieces together, improvement accelerated, consistent with hypergrowth behavior.  While he didn't track how quickly his handicap decreased, making it impossible for us to build an S-curve, 28 months into the project, he has surpassed 91% of the 26 million golfers who register a handicap with the US Golf Association (USGA) database.  Not surprisingly, his rate of improvement (if measured as handicap) is now slowing as he faces competition from the top 10% amateur golfers.

Just as understanding the S-curve can keep discouragement at bay as we build new knowledge, it can also help us understand why ennui kicks in once we reach the plateau.  As we approach mastery, our learning rate decelerates, and while the ability to do something automatically implies competence, it also means our brains are now producing less of the feel-good neurotransmitters — the thrill ride is over.

查看原圖

As our learning crests, should we fail to jump to new curves, we may actually precipitate our own decline. That doesn't necessarily mean a financial downfall, but our emotional and social well-being will take a hit.  Saul Kaplan, Chief Catalyst at Business Innovation Factory, shares: "My life has been about searching for the steep learning curve because that's where I do my best work. When I do my best work, money and stature have always followed."  Or paraphrasing James Allworth, "Steve Jobs solved the innovator's dilemma because his focus was never on profit, but better and better products."  Forget the plateau of profits: seek and scale a learning curve.

The S-curve mental model makes a compelling case for personal disruption.  We may be quite adept at doing the math around our future when things are linear, but neither business nor life is linear, and ultimately what our brain needs, even requires, is the dopamine of the unpredictable.  More importantly, as we inhabit an increasingly zig-zag world, the best curve you can throw the competition is your ability to leap from one learning curve to the next.

Ahead the Curve?