Don’t trust expert predictions

I’ve always thought that experts had some value at helping you to see the universe of possible futures, although not at telling you which possible future would occur. In this respect, they help to remedy two human behavioral/psychological limitations: (1) people tend to think in self-reinforcing patterns and to have difficulty challenging (or even articulating) their own assumptions and (2) people find it much easier to reason about concrete, semi-plausible stories rather than vague abstractions or unknown unknowns.

For example, many businesses suffer from myopia, focusing on one happy imagined future. If an expert with lots of inflated credentials can convince a business manager to consider the possibility of even one alternative future, that expert will have created positive value.

An example of part (2) would be the use of expert forecasts to train emergency-response teams — if you are running a corporate information security department, or a financial risk management office, or a local fire department, you may use expert-generated scenarios in your training and planning exercises. It is 100% certain that the real emergency will not follow what you planned for, but, you still need something concrete for your planning exercise.

How about experts USING extrapolation algorithms? Do the experts worsen the result of the extrapolation?

Most of the things that people are interested in hearing forecasts about are inherently hard to predict. That’s why we are interested in them. Being able to predict when the sun comes up tomorrow is a highly useful skill, but it won’t get you on TV to argue with other experts about it. In contrast, consistently predicting who is going to beat the point spread in NFL games is practically impossible because the point spreads are established in self-correcting markets, but because it’s so hard to do, there is much demand for the opinions of people who can go on TV and act like they know what they are talking about.

What’s amusing is how little interest there is in being able to model the past correctly. For example, eEconomist Steven D. Levitt’s abortion-cut-crime didn’t predict the past accurately, as an hour of checking Google showed, but it still helped make him rich and famous. Now we know his theory was based on simple incompetence on his part, but it was pretty obvious it failed to predict the past when it was unveiled in 1999.

I’ve not had a chance to look at Tetlock’s book, but I have read Klein’s “Sources of Power” (which was about decision making). My opinion is that while experts make far better decisions than a non-expert, many experts have been co-opted by people with money. Like when the administration hired James Guckert/Gannon, Armstrong Williams, Dave Smith, and Karen Ryan (among others) to advertise for them. The book “Trust Us, We’re Experts” is also about “scientists” for sale and the pseudoscience that gets bandied about in the media. Remember the brokerages that had to pay large fines for pushing stocks that they were being bribed to push?

“By definition, an index fund should earn (before fees) precisely the market-cap weighted average of the before-fees performance of all of the actively managed funds in the same basket of securities”
No. An index fund will carry the basket of securities that make up the S&P500, or Russell 2000 index, or whatever index the fund is based on. They are NOT averages of other funds. They are based on a single list that changes semi-annually (in the case of some of the Russell indices), or less frequently (in the case of the Dow Jones Industrial Average). When the word “core” appears as part of the name of the fund, it will be predominantly the index with a few stocks missing. Index funds usually have lower management fees since there is no need to manage them, nor do any form of analysis. And it is the management/12b-1 fees that can make a lot of difference in a fund’s performance: which is why they tend to outperform managed funds. Very few indices are “based on fund managers” and those are indices with “growth” or “value” in their names, as G&V are subjective judgements of the fund managers (we think X is under priced, so that makes X a “value” stock).
http://www.russell.com/US/Indexes/US/methodology.asp

Large funds sometimes close themselves to new investors to keep their performance up. If they take in too much money, there aren’t usually enough investments to be able to keep their performance up. It isn’t the smaller funds that underperform, sometimes it is the large ones, as they grow too big for effective management, and have too much money to be able to find good invesments for it.

Small cap funds were popular for some time, and it looks like folks are trying to pitch Large cap funds as the next “big thing.” I used to work for a company that ran websites for mutual fund companies, so I’d see the performance numbers all the time. The remark that an index would outperform 2/3 of funds is quite accurate, although the funds it would outperform might differ each year. What surprised me was that some funds would underperform the index every year, year after year, and would still have investors.

houyuping 07年7月11日

google排名
google排名
wow gold
wow gold
powerleveling
powerleveling
wow powerleveling
wow powerleveling
wow power leveling
wow power leveling
power leveling
power leveling
wow power level
wow power level
Ddo plat
Ddo plat
FFXI Gil
FFXI Gil
GW Gold
GW Gold
GW Power Leveling
GW Power Leveling
RS Gold
RS Gold

RS Power Leveling
RS Power Leveling
Lotro Gold
Lotro Gold
Lotro Power leveling
Lotro Power leveling
翻译公司
翻译公司
上海翻译公司
上海翻译公司
北京翻译公司
北京翻译公司
监控
货架
货架
货架
涂层测厚仪
涂层测厚仪
翻译公司
仓库
仓库
仓库出租
仓库出租
物流园区
物流园区
红外测温仪
红外测温仪

超声波测厚仪
超声波测厚仪
超声波探伤仪
超声波探伤仪
频闪仪
频闪仪
涂层测厚仪
涂层测厚仪
电火花检测仪
电火花检测仪
google排名
门禁
门禁
门禁系统
门禁系统
搬家
搬家公司
北京搬家公司
光盘制作
光盘印刷
光盘印刷
搬家公司
激光打标机
搬家公司
北京搬家公司
hongkong hotel
hongkong hotel
beijing tour
beijing tour
集团电话
集团电话
wow power leveling
wow power leveling
wow powerleveling
wow powerleveling
wow power level
wow power level
china tour
china tour
青春痘
青春痘

摆闸
痤疮
痤疮
色斑
色斑
黄褐斑
黄褐斑
青春痘
青春痘
口腔医生
口腔医生
口腔
口腔
口腔医院
口腔医院
口腔科
口腔科
三星集团电话机
呼叫中心
电话交换机
交换机
呼叫中心
测厚仪
测温仪
停车场
道闸

wanyamin 07年7月19日
wow gold
wow gold
powerleveling
powerleveling
wow power leveling
wow power leveling
wow powerleveling
wow powerleveling
world of warcraft powerleveling
world of warcraft powerleveling
world of warcraft power leveling
world of warcraft power leveling
cheap wow powerleveling
cheap wow powerleveling
cheap wow power leveling
cheap wow power leveling
beijing tour
beijing tour
beijing
beijing
china tour
china tour
great wall
beijing travel
beijing travel
网站设计
网站设计
多媒体
光盘制作

光盘印刷
光盘印刷
超声波探伤仪
超声波探伤仪
频闪仪
频闪仪
涂层测厚仪
涂层测厚仪
电火花检测仪
电火花检测仪
搬家公司
搬家公司
北京搬家公司
google排名
wow gold
翻译公司
翻译公司
上海翻译公司
上海翻译公司
北京翻译公司
北京翻译公司
货架
货架
货架
猎头
猎头
软件测试工程师
软件测试工程师
软件测试
软件测试
软件测试培训
软件测试培训
仓库

仓库
仓库出租
仓库出租
物流园区
物流园区
痤疮
痤疮
色斑
色斑
黄褐斑
黄褐斑
青春痘
青春痘
三星集团电话机
呼叫中心
电话交换机
交换机
呼叫中心
测厚仪
测温仪
cheap Silkroad gold
cheap Silkroad gold
Silkroad gold
Silkroad gold
Silkroad Online gold
Silkroad Online gold
sro gold
sro gold
Archlord gold
Archlord gold
cheap Archlord gold
cheap Archlord gold
Archlord Online gold
Archlord Online gold
buy cheap Archlord gold
buy cheap Archlord gold
Dofus kamas
Dofus kamas
cheap Dofus kamas
cheap Dofus kamas
Dofus gold
Dofus gold
Dofus money
Dofus money
Rappelz Rupees
Rappelz Rupees
Rappelz gold
Rappelz gold
cheap Rappelz Rupees
cheap Rappelz Rupees