找出最佳決策點

Finding the Decision Making Sweet Spot
瀏覽人數:1515


影片載入中...
史考特.貝里納托(Scott Berinato)《哈佛商業評論》英文版資深編輯,解釋在社會環境中做決策的機會和陷阱,內容取材自山迪.潘特蘭(Sandy Pentland)的文章〈超越決策一言堂〉(Beyond the Echo Chamber)。

嗨!我是《哈佛商業評論》資深編輯貝里納托,也是與潘特蘭教授合作〈超越決策一言堂〉的編輯之一,這篇文章刊於11月號,內容是關於社會環境中的決策。我推薦你在紙本雜誌或hbr.org上閱讀這篇文章。這篇文章優異之處,在於它把通常屬於軟性主題的「決策」,搭配應用硬性的數學,成為這種形式。這不是印壞的文件,而是散點圖,上面顯示許多投資決策,共有幾百萬個決策,分布在社群交易平台eToro上。

eToro的使用者在進行交易時,可以參考其他使用者的投資策略。使用者的策略若是被他人仿效,每次都能賺取少量佣金。每個交易者的最終績效,其他人都能看到。eToro提供潘蘭特、博士後研究員亞舒勒和博士生潘巍(音譯),一個現實世界的實驗室,在其中探索社會決策。決定要模擬誰的交易,以及由此產生的績效,會讓他們知道在社會環境中,哪些做決策的策略可帶來良好的報酬率,哪些策略會導致不好的報酬率。

其實,這種視覺化圖像的美妙之處,在於它呈現投資策略的最佳表現點,以及績效最佳的人。但若要真正了解這個圖,讓我們假設性地放大看看,潘蘭特和他的團隊如何製作這張圖。圖表在X軸和Y軸上繪製了相同的投資人。在全尺寸下,每個軸上有160萬個使用者。

我們查看五個假設的使用者:艾莉森、丹、蘇珊、艾美、艾倫。在X軸上,這些使用者是交易員。他們用自己的錢做投資決策。在Y軸上,他們的交易策略被其他交易者模仿。因此,當艾利森根據丹的投資組合進行交易時,會在X軸上的艾莉森和Y軸上的丹的交匯處創造一個點。

每個交易者的投資,都是根據他們模仿的對象來繪製。丹的交易是根據艾莉森和蘇珊的交易。蘇珊的交易抄襲艾莉森、丹和艾倫。艾美和艾倫則是模仿了所有其他人。這樣做160萬次,你會得到一千萬筆交易。但讓我們再次放大檢視。在分析資料時,潘蘭特和他的團隊發現一種難以忽視的現象。

他們看到決策泡泡。在社會環境中,使用者可能覺得自己得到新資訊,並據此做出好決策,但其實,他們只是取回自己之前放在外面的資訊。它的過程是這樣的。艾倫根據艾美的策略來投資,艾美的交易是根據丹的交易,而丹的交易是根據蘇珊的交易。事實證明,蘇珊的投資是根據艾倫的策略。其實,艾倫的策略基礎,就是跟隨那些跟隨艾倫自己的使用者。

但他看不到這現象,因為他已經被其他三個使用者過濾。再次縮小。以大比例尺繪製時,像艾倫這樣的循環策略,會顯示為清晰定義的密集交易領域,你可以在這裡看到。這數萬個交易者中,有很多人自認正在多元化自己的策略,但其實只是在遵循自己和自己身邊的人提出的想法。這與艾倫的問題一樣,只是在更為複雜的一系列交易之中。

因此,更難辨別這樣的循環。在光譜的另一端是艾莉森。如果你還記得的話,她只根據別人的決定進行交易。她屬於根本沒有仿效很多策略的群體,如左下角的稀疏點所示。如果說右上方密集區的交易者是處在回音室,艾莉森所屬的群體就是處在隔離室中。他們的決定根本沒有考慮其他人的想法。

中間的交易者表現出一定的密度,但分布更廣,他們遵循許多不同的策略,但彼此沒有相似到足以讓他們進入回音室。他們尋求許多想法,做出的決策不只反映自己的想法。他們處於最適點。我們怎麼知道他們處於最適點?這就是交易平台的好處,我們知道交易者的報酬率。處於最適點的交易者,績效優於回音室和隔離室裡的人,高出30%。

為了進一步確認,哪些決策模式可產生最佳報酬,潘特蘭和亞舒勒進行干預。他們向孤立的交易者建議了一些策略,讓他們的位置提升到最適點。他們告訴艾利森那群人,去跟隨更多不同的交易者。他們也提供建議給回音室裡的人,告訴艾倫那群人要多角化投資,遵循更多不同人的策略。讓他們向下移動到最適點。

在這兩種情況下,研究人員都改善了移入最適點交易者的投資績效。在互相密切連結的世界中,社會決策已成為常態。交易平台讓潘特蘭和他的團隊運用硬數學,來測試自己的假設。但他認為,同樣的想法適用於所有決策環境,即使是那些較不易衡量的決策,包括制定企業策略、押寶創新、雇用員工。任何管理決策,都有一部分屬於社會決策。決策者必須避開隔離室和回音室,找到那個最適點。

(劉純佑譯)


Hi! I'm Scott Berinato, a senior editor at Harvard Business Review. I was one of the editors who worked with Professor Sandy Pentland to develop his article in the November issue of HBR called “Beyond the Echo Chamber,” which is about decision making in a social context. I urge you to read his article in the print magazine or on hbr.org. What makes it great is the fact that it takes a normally soft topic like decision making and applies hard math to it, in the form of this. This is not a bad photocopy. It's a scatter plot that shows you investment decisions -- millions of them -- on a social trading platform called eToro.

On eToro, users can make trades based on other users' investment strategies. Users whose strategies are copied earn a small commission each time. And every trader's ultimate performance is available for all others to see. EToro gave Pentland, post-doctoral student Yaniv Altshuler, and graduate student Wei pan a real-world lab to explore social decision making. The decision of who to emulate trading, along with the resulting performance, would show them what decision making strategies in a social context led to good returns and which ones led to bad returns.

In fact, the beauty of this visualization that it shows you the sweet spot of investment strategies and those who perform the best. To really understand this graph, though, let's hypothetically zoom in to see how Pentland and his team created it. The chart maps the same investors on both the X- and Y- axis. At full size, there are 1.6 million users on each axis.

Here we're looking at five hypothetical users – Allison, Dan, Susan, Amy, and Allen. On the x-axis, these users are traders. They are the ones making investment decisions with their money. On the y-axis, they are the traders' strategies being copied by other traders. So when Allison trades based on Dan's portfolio, a point is created at the intersection of Allison on the x-axis and Dan on the y-axis.

Each trader's investments are plotted against the person whose trade they emulated. Dan trades based on Allison's and Susan's trades. Susan's trades copy Allison, Dan, and Allen. Amy and Allen copy everybody else here. Do this 1.6 x million times over. And you arrive at this -- 10 million trades. Let's zoom in again, though. In analyzing the data, Pentland and his team found a phenomenon that's hard to ignore.

They saw decision making bubbles. In social contexts, users may think they're getting new information and making good decisions based on that, when in fact, they are simply getting the information back that they put out there in the first place. It happens like this. Allen invested based on Amy's strategy. Amy's trade was based on Dan's. Dan's was based on Susan's. And Susan's investments, it turns out, were based on Allen's strategy. In fact, Allen's strategy is based on following users who are actually following him.

But he can't see that because he's been filtered through three other users. Zoom out again. When mapped on a large scale, circular strategies like Allen's show up as clearly define dense fields of trades, which you can see here. Many of these tens of thousands of traders think they're diversifying their strategies, when they're simply following the same ideas that they themselves and those close to them are putting out there. It's the same as Allen's problem, only in a far more complex array of trades.

So it's even harder to discern the circularity. On the other end of the spectrum are the Allisons. If you remember, she only made one trade based on other people's decisions. She's part of a group that don't copy many strategies at all, as seen by the sparse population of dots here in the lower left. If those traders in the dense field in the upper right are in an echo chamber, Allison and her cohorts are in an isolation chamber. Their decisions are simply not taking into account other people's ideas.

The middle traders, who shows some density but across a far wider swath, are following lots of different strategies, but not so many so close that they've entered into an echo chamber. They're seeking many ideas and making decisions that aren't really just reflections of their own ideas. They are in the sweet spot. How do we know they're in the sweet spot? Well, this is the beauty of the fact that this is a trading platform. We know the traders' returns. And those in the sweet spot perform 30% better than those in the echo chamber and those who were isolated.

To further confirm the pattern of decision making that produces the best returns, Pentland and Altshuler intervened. They suggested strategies to the isolated traders that would move them up into the sweet spot. They told the Allisons to follow more different traders. And they also offered advice to those in the echo chamber, telling the Allens of the world to diversify and follow the strategies of more and different people. Move them down into the sweet spot.

In both cases, the researchers improved the investment performance of the people they moved into the sweet spot. In a vastly interconnected world, social decision making is becoming the norm. A trading platform gave Pentland and his team hard math to test his hypotheses. But he believes that the same idea applies to all decision making settings, even the ones that are less easily measured -- setting corporate strategy, making innovation bets, hiring. Any management decision is, in part, a social decision. And the deciders need to avoid both the isolation chamber and the echo chamber and find that sweet spot.



本篇文章主題決策