Horizontal links

Wednesday, 8 June 2011

Lies, damn lies, and statistics!

Recent IPO statistics on the yield:

FY 2011 = 5.30%

FY 2012 = 5.51%

Take at the look at the charts below.




Big growth for the 1st chart?

Since I prefer round numbers, the same statistics in my mind graph is as the 2nd chart.

The difference between 2011 and 2012 is only 0.21% lah!


Is this example similar in our daily lives? You bet!


In my work, I always remind my colleagues to round up our Key Performance Indicators (KPIs) to a big round number – don’t create the “illusion” of precision; where none is needed.

I am now working in supply chain, so one important KPI is total lead-time. I always hate it when I see lead-time written as 17.65 days!? What is that?

This “long” lead-time means it’s likely by ocean transport  mode. So in reality, we just need to know whether the vessel/container will arrive in 17 or 18 days will do.

Even if it’s airfreight, we just need to know the hours and minutes. Estimated time of arrival is at 22:30 hours. Do we need to put in the seconds? You mean airport so precise?

Those who like to “add feet to snake” are usually people who have been working in the air-con office for far too long, and may have lost touch with reality on the ground… (Poor thing. I must bring them out more often!)

Of course in some industry, precision is the difference between life and death – like working in the nuclear industry or when undergoing brain surgery. But in most other industries, we don’t really need precision. It’s all marketing hype!

Let’s take photography. We compare the sharpness of lenses by the number of eyelashes visible in professional magazines. Do we count eyelashes when looking at photos in real life?

Don’t get me started with pixels…. If the biggest print-out we ever print is A4 sized, can you see the difference between 6 and 12 mega-pixels?

And for gamers, the video cards are evaluated by frames per seconds. But can the human eye see the difference between 40 fps and 60 fps? For cartoons, we only need 24 fps to get the natural motion effect ;)

Yup, the same marketing trick is used in finance and trading. All we want as marketer (many moons ago I studied marketing…) is to get you to “up-size”. And McDonald does it best! “Would you like to up-size sir?”

Just be aware next time you look at statistics. You may want to take a look at my other post on percentages: "I dont want percentages I want money!"

I am very happy to see many finance bloggers put their portfolio performance monthly. I find it a great way to gauge the mood of the market – I am more interested to see the asset allocation (how much cash, how much equities, how much bonds, etc); or to see what stocks they are vested in (growth, defensive, cyclical, etc).

I don’t really care whether it’s positive or negative. Or increase by what %. Our portfolio sizes are different, strategies are difference, personalities are different – what’s there to compare?

But I am sure some readers will get a kick out of this “benchmarking”, and declaring I beat so and so by 0.21 percent! It’s the Singapore habit – we need to be number 1!

Hey! In the words of my hokkien army friends: “Whatever makes you happy!” or “Hua hee du ho!”



9 comments:

  1. haha but sorry SMOL, can't agree with you on this one.

    i'll stick to relative rather than absolute.

    we run faster than the ants - absolute. the ant run faster than us - relative.

    but still a great post thanks.

    ReplyDelete
  2. Hi SMOL,

    I agree with you. Numerical precision gives the illusion of certainty. Telling you the support of STI is 3101.23 sounds more intelligent than just plain old 3100. In practice, the number of significant figures we need depends on the usage.

    I find it quite amusing to see target prices of brokerage reports giving it up to 2 decimal places. It gives too much 'certainty' in the face of uncertainty, haha!

    ReplyDelete
  3. One time snap shot of statistics may lie, damn lie.

    But, a series of regular and consistent presentation of statistics may tell you a pretty good story of what is going on there and can't really lie.

    ReplyDelete
  4. It is better to be vaguely correct than to be precisely wrong

    ReplyDelete
  5. Hi SMOL,

    It is very easy to lie with statistics, not only the chart which you mentioned. There are many things which people can do to lie.

    Source of data, relevancy of data, data tools for audience visualisation(the case which you mentioned) etc.

    Thus it is never good to trust the numbers or graphs posted in newspapers or reports. Better to churn deeper into the number before making any conclusion on the number

    Great post!

    ReplyDelete
  6. Hey SMOL,

    your charts also remind me of what big pharma likes to do.

    Drug A from Company A has a 2% chance of curing a patient of whatever disease.

    Drug B from Company B has 3% chance of curing the same patient from the same illness.

    Company B will market their drug as "50% BETTER THAN DRUG A!"

    ReplyDelete
  7. 1. Cheers Coconut!

    Hey! It's cool to have an alternative view! I can understand where you come from. I like to lurk amongst hard-core traders' blogs to get a feel of the market. It always amaze me how scalpers can make money with 1 or 2 ticks' profit!? You guys are the hares! Jump in and out :) In your case, relative is more important!


    2. Hi LP,

    Spot on! I also wonder where analysts get this 1.23 "extra precision" support from!?


    3. Hello CW8888,

    Agree. That's why I nowadays stay away from IPOs... There are so many other stocks in SGX with a longer track record. Harder for them to mislead me.


    4. Hi Isaac!

    Hey! That's a great example on the pharma companies! 50% increase!!! Must use your example in my future posts ;) Steal with pride!


    5. Hello OT83

    I always like to say: "Never trust a statistics that you have not manipulated yorself!" LOL!

    ReplyDelete
  8. Hi SMOL,

    I dug some old posts related to this topic. Perhaps you'll like them?

    1. Statistics lie! http://bullythebear.blogspot.com/2008/03/statistics-lie.html

    2. I lie to statistics!
    http://bullythebear.blogspot.com/2008/03/i-lie-to-statistics.html

    Enjoy!

    ReplyDelete
  9. LP, thanks!

    It's funny how somethings never change :) You were 2 years ahead of me!

    I was laughing when I saw the yield chart from that "property trust" on Wednesday and can't help poking fun at it :)

    ReplyDelete

Related Posts Plugin for WordPress, Blogger...