Level of conviction

[mikazo]

Hey,

 

I was just thinking yesterday about an investment and how sure I was of its prospects. Reading pessimistic opinions and reports online can definitely erode at that sure feeling, but I wanted to find out why the business was so down in the first place.

 

So I'm just wondering, how much certainty do you guys establish before placing money on an investment? Do you spend an amount of money proportional to your certainty? I fully believe in the idea of investing within one's circle of competence and knowing that circle's limits. But even within my circle, I have doubts sometimes. I know the whole idea of successful investing is to place big bets on sure things when the market thinks otherwise, but I'm having trouble deciding how big of a bet to make.

 

Thanks for any thoughts on the matter.

 

-Mike

[ragnarisapirate]

It depends on the stock. Something like SVU, I am not super convinced about, but think that the risk reward make it attractive enough for me to be long. I a have a lot of conviction about SYTE, though, averaged in at prices less than it is trading at. Every situation is fluid and requires analysis that you can live with.

 

To put it in perspective. If you have 100% conviction (or rather correctness) that you are sitting on a 70 cent dollar, and 20% conviction that you are sitting on a 2 cent dollar, what do you allocate towards? I would argue that you do what makes you comfortable. While 20% on a 2 cent dollar isn't exactly Ben Graham, under the right circumstance, there is a place for it in certain portfolios (being relatively young, with time on my side, I have a few of them making up very small positions). I wouldn't bet the farm on it, but, I would likely be willing to make a really small bet.

 

Thoughts?

[racemize]

I kind of like the Kelley bet, which is 2*conviction - 1 for how much to bet/invest.  e.g., an assigned 60% chance of success gets a 20% position (120%-1), though I might set the bar a little higher in general.

 

Edit: or have a very high bar for "success"

[enoch01]

 

Thanks for any thoughts on the matter.

 

-Mike

 

Opportunities arise when lots of holders sell en masse, which is an act of capitulation.  Owners feel it after giving up on all reasons to hold.

 

I try to invert this process, and think about purchases in terms of capitulation: what am I compelled to buy, after giving up on all reasons for not holding?  This dramatically shrinks the universe of worthwhile purchase opportunities, which it seems to me, is the way it ought to be.

[racemize]

Opportunities arise when lots of holders sell en masse, which is an act of capitulation.  Owners feel it after giving up on all reasons to hold.

 

I try to invert this process, and think about purchases in terms of capitulation: what am I compelled to buy, after giving up on all reasons for not holding?  This dramatically shrinks the universe of worthwhile purchase opportunities, which it seems to me, is the way it ought to be.

 

That's interesting, though it seems like there's always a base reason to not be in cash, e.g., devaluation of the dollar/inflation. 

[oddballstocks]

 

Thanks for any thoughts on the matter.

 

-Mike

 

Opportunities arise when lots of holders sell en masse, which is an act of capitulation.  Owners feel it after giving up on all reasons to hold.

 

I try to invert this process, and think about purchases in terms of capitulation: what am I compelled to buy, after giving up on all reasons for not holding?  This dramatically shrinks the universe of worthwhile purchase opportunities, which it seems to me, is the way it ought to be.

 

I've gone this route, I can't find any possible reason to not invest, I can't kill the idea.  Even if there is a queasy factor (usually good) then I throw some money at it.  I've done the worst it seems when I've felt the best about an idea, I'm usually over confident.  If I feel queasy about something but I know I've done my research that's usually a good sign.

 

I remember hearing a quote saying something to the effect that a value investment that doesn't make you question your sanity might not be a real value investment.  I've purchased a few of those situations, they've turned out really well.  My slam dunks...uh, well let's forget about those.

[beerbaron]

In theory one should apply the Kelly formula to size it's investment to optimize returns. The problem is not determining the potential gain but the odds of winning (level of conviction). It's not because it's hard that one should not use it tough!

 

BeerBaron

[scorpioncapital]

"I kind of like the Kelley bet, which is 2*conviction - 1 for how much to bet/invest.  e.g., an assigned 60% chance of success gets a 20% position (120%-1), though I might set the bar a little higher in general"

 

I'm not sure how correct this formula is in the real world. Conviction = 100%* 2 -100 = 100%. But this is crazy, the amount to bet in this situation in a margin account should always be > 100%.

[rranjan]

"I kind of like the Kelley bet, which is 2*conviction - 1 for how much to bet/invest.  e.g., an assigned 60% chance of success gets a 20% position (120%-1), though I might set the bar a little higher in general"

 

I'm not sure how correct this formula is in the real world. Conviction = 100%* 2 -100 = 100%. But this is crazy, the amount to bet in this situation in a margin account should always be > 100%.

 

Another point to note -

 

Card game is closed loop where you can calculate probability and apply this formula but stock market might not work same way. In my opinion, it can be used as reminder to bet more in proportion, based on probabilistic approach. Since probability is only an estimate with outside factor influencing it, following the formula blindly might not work so well.

[tombgrt]

Overconfidence and wrong calculations is why some use a half kelly. You get a lot less volatility and 3/4 the return of a normal Kelly.

 

The formula is actually =edge/odds.

 

I think it goes like this:

Odds measures the profit if you win. Imagine a stock that grows fivefold if you win. That's 4 to 1 or 4.

Edge is how much you expect to win. So if you established that you have a 1/2 chance of winning, on average you would end up with 2.5 times your money. (ex. 5*$100/2)

 

So the formula is 2.5/4. You should bet 62.5% of your account on that bet for one kelly. That way you optimize returns over the long haul & you can never go broke. You could of course be wrong in your assesment that you have a 1/2 chance of winning. Overbetting would lead to lower or even negative returns. I would only use it to get a raw idea of how much I should put in which stocks and then take half of that at most.

 

 

Edit: something isn't making any sense. Got to go now, will check it later. :p

 

 

[berkshiremystery]

There is a third book written about the Kelly criterion, also good to read like "Fortune's Formula"  by William Poundstone and a much easier read than the mathematical book "The Kelly Capital Growth Investment Criterion" by Leonard C. MacLean, Edward O. Thorp and William T. Ziemba. The book is called "Scenarios for Risk Management and Global Investment Strategies", also written by William T. Ziemba. The entire book is about the Kelly criterion, Buffett, Ed Thorp and optimized betting.

 

Scenarios for Risk Management and Global Investment Strategies

(The Wiley Finance Series)

by William T. Ziemba and Rachel E. S. Ziemba

 

http://www.amazon.com/Scenarios-Management-Investment-Strategies-Finance/dp/0470319240/

 

Hardcover: 334 pages

Publisher: Wiley; 1 edition (January 8, 2008)

Language: English

ISBN-10: 0470319240

ISBN-13: 978-0470319246

 

[Kraven]

I find the only problem with the Kelley formula, and it's not that big of one, is that the numbers are just kind of made up.  I mean that they are pulled out of the air.  So with that kind of precision there's no doubt of its usefulness.