REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
michaelsieg
Posts: 563
Joined: Mon Jan 07, 2013 11:02 am

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by michaelsieg » Thu Jan 10, 2019 10:58 pm

2630.1

bo105954027
Posts: 66
Joined: Wed Mar 30, 2016 4:00 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by bo105954027 » Thu Jan 10, 2019 11:11 pm

2737.24
Time in market beats timing the market.

raymclean
Posts: 65
Joined: Mon Mar 19, 2012 11:53 am

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by raymclean » Thu Jan 10, 2019 11:21 pm

2921
Thanks for running this contest again this year, Taylor and Ralph.
Ray

thomas716
Posts: 9
Joined: Sat Mar 12, 2011 4:31 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by thomas716 » Thu Jan 10, 2019 11:33 pm

2382.51

User avatar
tadamsmar
Posts: 8419
Joined: Mon May 07, 2007 12:33 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by tadamsmar » Thu Jan 10, 2019 11:38 pm

3477.77

User avatar
rarvesen
Posts: 70
Joined: Tue Feb 20, 2007 4:34 pm
Contact:

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by rarvesen » Fri Jan 11, 2019 12:54 am

The tracking site has been updated. Please let me know if I got your entry incorrect... I wrote a script that collects a lot of the data, but some of it is a manual process and I don't want to make a mistake with your guess. :-)

Tracking site
https://www.lostoak.com/ls/diehards/contest

Miriam2
Posts: 2605
Joined: Fri Nov 14, 2014 11:51 am

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by Miriam2 » Fri Jan 11, 2019 1:21 am

rarvesen wrote: The tracking site has been updated. Please let me know if I got your entry incorrect... I wrote a script that collects a lot of the data, but some of it is a manual process and I don't want to make a mistake with your guess. :-)

Tracking site
https://www.lostoak.com/ls/diehards/contest
Ralph, the tracking site used to give us the option of seeing the contest by placement OR by name in alphabetical order. Does this site no longer allow us to view by name?

User avatar
rarvesen
Posts: 70
Joined: Tue Feb 20, 2007 4:34 pm
Contact:

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by rarvesen » Fri Jan 11, 2019 1:30 am

Miriam2 wrote:
Fri Jan 11, 2019 1:21 am
Ralph, the tracking site used to give us the option of seeing the contest by placement OR by name in alphabetical order. Does this site no longer allow us to view by name?
Sorry, that is correct, it only displays in placement order. I was not sure if anyone was using the other sort features and they were removed when the site had a slight interface change.

If helpful, you should be able to use your browser find / search on page feature to find your entry on the tracking site. It looks like you have the guess of 2240.00, let me know if that is incorrect.

ABQ4804
Posts: 407
Joined: Sun Jul 24, 2011 4:08 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by ABQ4804 » Fri Jan 11, 2019 1:35 am

Rarvesen, thank you for supporting all of us in putting together this fun contest! Entry deadline has passed, let the fun of our “calculations” begin.

User avatar
LadyGeek
Site Admin
Posts: 55838
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by LadyGeek » Fri Jan 11, 2019 8:59 am

I see several duplicate entries, so there's more work to be done.

According to the rules, the earliest entry is the only one that counts. Can the duplicates be removed? If anyone is running stats on the distribution, the duplicate entries will skew the results.

I suggest posting the duplicate names so the member will know the entry has been disqualified.

Also, the closest entry is +/- (above and below) the value. A tie is possible.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.

User avatar
ResearchMed
Posts: 9062
Joined: Fri Dec 26, 2008 11:25 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by ResearchMed » Fri Jan 11, 2019 9:11 am

LadyGeek wrote:
Fri Jan 11, 2019 8:59 am
I see several duplicate entries, so there's more work to be done.

According to the rules, the earliest entry is the only one that counts. Can the duplicates be removed? If anyone is running stats on the distribution, the duplicate entries will skew the results.

I suggest posting the duplicate names so the member will know the entry has been disqualified.

Also, the closest entry is +/- (above and below) the value. A tie is possible.
How can a tie be possible, if only the earliest entry is left, when duplicates are removed?
Even if timing is "close", one will have been posted before the other, etc.

(I've already reported a problem with our numbers: Mr.ResearchMed and ResearchMed were both given the same entry.)

RM
This signature is a placebo. You are in the control group.

Scooter57
Posts: 940
Joined: Thu Jan 24, 2013 9:20 am

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by Scooter57 » Fri Jan 11, 2019 9:40 am

Are those last 4 entries typos? The order of magnitude is way off. They might be skewing the average. It would be nice to see the median and standard deviation, too.

User avatar
hornet96
Posts: 504
Joined: Sun Nov 25, 2012 6:45 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by hornet96 » Fri Jan 11, 2019 9:48 am

ResearchMed wrote:
Fri Jan 11, 2019 9:11 am

How can a tie be possible, if only the earliest entry is left, when duplicates are removed?
Even if timing is "close", one will have been posted before the other, etc.
If the ending S&P value is 2500, and:

*Person A guessed 2501 for the year-end value, and
*Person B guessed 2499 for the year-end value

Then they would tie for the winning spot since they are both 1 point away from the ending value (one is over, one is under). I don't think they're using the Price-is-Right rules for this contest. :wink:

User avatar
Topic Author
Taylor Larimore
Advisory Board
Posts: 28502
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by Taylor Larimore » Fri Jan 11, 2019 10:47 am

Ralph:

We appreciate your posting all the guesses so quickly. I like the new shaded blue to emphasize our top-five leaders.

Interesting facts about the 2018 Contest:

* We had 479 Bogleheads enter the contest in January 2018. Only 78 (16.3%) could even guess the direction of the S&P 500 Index (December ending lower than January).

* The price of the S&P Index on December 31, 2018 was 2506.85. The average Boglehead guess in January 2832.49 was better than the average Wall Street expert guess 2840.82.

Lesson Learned: No one (except by luck) can forecast the stock market.

Ralph, your computer program has helped make the Boglehead Contest the best of all similar contests on the web!

Thank you and best wishes.

Taylor
"Simplicity is the master key to financial success." -- Jack Bogle

User avatar
rarvesen
Posts: 70
Joined: Tue Feb 20, 2007 4:34 pm
Contact:

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by rarvesen » Fri Jan 11, 2019 10:51 am

Can the duplicates be removed?
I'm not the one that decides, but in my opinion, I don't think duplicate entries should be removed and all entries should be tracked. The contest makes the point that no one can predict the market and it's for fun. If two people are tied for first place since they entered the same guess, then I think they both won the contest. The person that entered their guess first would get the prize assuming there is only one prize available.
+/- (above and below)
The same for the +/-, more then one person could win the contest even with different guesses. I think it's the same... they both won the contest and the person that entered their guess first would get the prize.

I'm not saying this is how the contest works, just my opinion / how I view it.
Ralph

User avatar
rarvesen
Posts: 70
Joined: Tue Feb 20, 2007 4:34 pm
Contact:

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by rarvesen » Fri Jan 11, 2019 10:53 am

(I've already reported a problem with our numbers: Mr.ResearchMed and ResearchMed were both given the same entry.)
Thanks for letting me know, this has been updated, let me know if you still see any problems.
Ralph

User avatar
LadyGeek
Site Admin
Posts: 55838
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by LadyGeek » Fri Jan 11, 2019 11:20 am

^^^ From an administrative perspective, I confirm that Mr.ResearchMed and ResearchMed are indeed husband and wife.

I also want to be clear that I have no influence on the contest manager decisions. I'm just a contest entrant expressing an opinion.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.

User avatar
BolderBoy
Posts: 4393
Joined: Wed Apr 07, 2010 12:16 pm
Location: Colorado

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by BolderBoy » Fri Jan 11, 2019 12:15 pm

tadamsmar wrote:
Thu Jan 10, 2019 4:42 pm
BolderBoy wrote:
Thu Jan 10, 2019 4:08 pm
Hmmmm. One downside I notice is that one has to read through 13 (presently) pages of guesses to see if someone else has used the one you plan to use. Since the rules (from last year) are that the earliest duplicate submission wins over a later one and you can't change your guess once submitted, it makes the duplicate guess 1) wasted and 2) useless.

Isn't there a better way to run this popular game?
I just thought of a better way. Add LadyGeek's instructions on how to search for duplicates to the Original Post (OP) of this topic and make it conspicuous so it would be hard to miss:

viewtopic.php?f=10&t=268318&p=4303407&h ... h#p4303407
An excellent suggestion! LadyGeek...?
"Never underestimate one's capacity to overestimate one's abilities" - The Dunning-Kruger Effect

User avatar
VictoriaF
Posts: 18893
Joined: Tue Feb 27, 2007 7:27 am
Location: Black Swan Lake

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by VictoriaF » Fri Jan 11, 2019 12:51 pm

LadyGeek wrote:
Fri Jan 11, 2019 8:59 am
I see several duplicate entries, so there's more work to be done.

According to the rules, the earliest entry is the only one that counts. Can the duplicates be removed? If anyone is running stats on the distribution, the duplicate entries will skew the results.

I suggest posting the duplicate names so the member will know the entry has been disqualified.

Also, the closest entry is +/- (above and below) the value. A tie is possible.
Barclays (Maneesh Deshpande), BTIG (Julian Emanuel), and Goldman Sachs (David Kostin) are tied at 3000. Should we vote two of them out? Or let all of them lose to some lucky Boglehead?

Victoria
WINNER of the 2015 Boglehead Contest. | Every joke has a bit of a joke. ... The rest is the truth. (Marat F)

User avatar
hornet96
Posts: 504
Joined: Sun Nov 25, 2012 6:45 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by hornet96 » Fri Jan 11, 2019 12:55 pm

tadamsmar wrote:
Thu Jan 10, 2019 4:42 pm

I just thought of a better way. Add LadyGeek's instructions on how to search for duplicates to the Original Post (OP) of this topic and make it conspicuous so it would be hard to miss:

viewtopic.php?f=10&t=268318&p=4303407&h ... h#p4303407
Ok, well I just discovered that the "search this topic" function is not exactly that robust, as when I search for the first four digits of my entry, I get this:

Image

However, apparently Triple Digit Golfer entered the same guess as me, but his result didn't show up in my search (which I did before I submitted my entry), as his included the comma.

I don't really care that much as I won't win anyway, but unfortunately the "search this topic" function isn't necessarily the panacea to ensure no duplicate entries are submitted.

User avatar
VictoriaF
Posts: 18893
Joined: Tue Feb 27, 2007 7:27 am
Location: Black Swan Lake

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by VictoriaF » Fri Jan 11, 2019 12:58 pm

hornet96 wrote:
Fri Jan 11, 2019 12:55 pm
However, apparently Triple Digit Golfer entered the same guess as me, but his result didn't show up in my search (which I did before I submitted my entry), as his included the comma.
LadyGeek's instructions are to conduct two searches, with and without a comma.

Victoria
WINNER of the 2015 Boglehead Contest. | Every joke has a bit of a joke. ... The rest is the truth. (Marat F)

BigJohn
Posts: 1739
Joined: Wed Apr 02, 2014 11:27 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by BigJohn » Fri Jan 11, 2019 2:03 pm

I just search on the first four numbers and then scrolled down the list of a few posts to ensure no duplicate.

Wings2525
Posts: 1
Joined: Fri Aug 03, 2018 9:06 am
Location: Michigan

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by Wings2525 » Mon Jan 14, 2019 9:42 am

2669.69

User avatar
LadyGeek
Site Admin
Posts: 55838
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by LadyGeek » Mon Jan 14, 2019 4:36 pm

Wings2525, Welcome!

FYI - The contest entry deadline has passed.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.

bkweathe
Posts: 200
Joined: Fri Feb 06, 2015 5:02 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST

Post by bkweathe » Fri Jan 18, 2019 11:03 am

Scooter57 wrote:
Fri Jan 11, 2019 9:40 am
Are those last 4 entries typos? The order of magnitude is way off. They might be skewing the average. It would be nice to see the median and standard deviation, too.
I suspect that those four are guessing the DJIA, not the S&P 500.

Brad

livesoft
Posts: 67398
Joined: Thu Mar 01, 2007 8:00 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by livesoft » Sun Apr 28, 2019 3:52 pm

I am reading Daniel Crosby's The Behavioral Investor which discusses Overconfidence and its subtypes. These contest threads should be changed to help people overcome some behavioral traps rather than enhancing them.

One way to improve this contest would be to have people provide RANGE for the closing price of the S&P 500. I know that this contest won't change, so I will have to start another contest after figuring out a good way to score responses.
Wiki This signature message sponsored by sscritic: Learn to fish.

User avatar
tadamsmar
Posts: 8419
Joined: Mon May 07, 2007 12:33 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by tadamsmar » Mon Apr 29, 2019 7:05 am

livesoft wrote:
Sun Apr 28, 2019 3:52 pm
I am reading Daniel Crosby's The Behavioral Investor which discusses Overconfidence and its subtypes. These contest threads should be changed to help people overcome some behavioral traps rather than enhancing them.

One way to improve this contest would be to have people provide RANGE for the closing price of the S&P 500. I know that this contest won't change, so I will have to start another contest after figuring out a good way to score responses.
The contest is not about the estimation* of the SP500 regardless of how you change it. It is about optimizing your chances of getting the prize.

The fact that most players don't see this is an entirely different behavioral anomaly.

* However, having the best possible estimate of the probability distribution is important for optimizing your chances.

User avatar
tadamsmar
Posts: 8419
Joined: Mon May 07, 2007 12:33 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by tadamsmar » Mon Apr 29, 2019 7:13 am

livesoft wrote:
Sun Apr 28, 2019 3:52 pm
I am reading Daniel Crosby's The Behavioral Investor which discusses Overconfidence and its subtypes. These contest threads should be changed to help people overcome some behavioral traps rather than enhancing them.

One way to improve this contest would be to have people provide RANGE for the closing price of the S&P 500. I know that this contest won't change, so I will have to start another contest after figuring out a good way to score responses.
You could have a contest where people provide a probability distribution for the closing price of the S&P 500. Then you could score the contest on the probability that they assigned to the closing price. Usually, contests of this sort use the negated natural log of the probability (aka "log loss") as the score since the probability itself will typically be an exceedingly small number. The log loss is like a golf score: lower is better.

See, for instance, this contest:

https://www.kaggle.com/c/mens-machine-l ... evaluation

as compared with the typical bracket pool were you just try to guess the outcome.

PS: But even this revised contest is not really about estimating the S&P 500. See my other post.

User avatar
tadamsmar
Posts: 8419
Joined: Mon May 07, 2007 12:33 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by tadamsmar » Mon Apr 29, 2019 7:24 am

livesoft wrote:
Sun Apr 28, 2019 3:52 pm
I am reading Daniel Crosby's The Behavioral Investor which discusses Overconfidence and its subtypes. These contest threads should be changed to help people overcome some behavioral traps rather than enhancing them.

One way to improve this contest would be to have people provide RANGE for the closing price of the S&P 500. I know that this contest won't change, so I will have to start another contest after figuring out a good way to score responses.
You could treat the range as a uniform probability distribution over the range. The log loss scoring would be that the person who specified the narrowest range that contained the closing price would be the winner. If the closing price was outside of everyone's range, then the contest would be a tie. So, you might want to include a tiebreaker rule.

Or you could treat the "range" as a specification of a mean and standard deviation of a normal distribution.

But it would be more creative to allow contestants to be free to specify any probability distribution.

livesoft
Posts: 67398
Joined: Thu Mar 01, 2007 8:00 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by livesoft » Mon Apr 29, 2019 7:46 am

Thanks for the responses. I am very familiar with probability distributions, log loss, etc and was thinking of something not dissimilar. I will start a new thread later this summer on such a scoring and a new contest.

I guess that as far as this thread goes, there is no nudging toward avoiding behavioral finance traps, but a nudging toward embracing them. Even your statement "It is about optimizing your chances of getting the prize" speaks to that much like winning a lottery. (Yes, I know the probabilities and distributions are different with lotteries.)
Wiki This signature message sponsored by sscritic: Learn to fish.

User avatar
tadamsmar
Posts: 8419
Joined: Mon May 07, 2007 12:33 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by tadamsmar » Mon Apr 29, 2019 9:01 am

livesoft wrote:
Mon Apr 29, 2019 7:46 am
Even your statement "It is about optimizing your chances of getting the prize" speaks to that much like winning a lottery.
I am not sure you understand the issue I am trying to raise. Let me try to explain more clearly.

Quoting from the OP of this thread, one of the rules is:
Forecast (below) the closing price of the S&P 500 Index on December 31, 2019.
I am competing in this contest and I broke that rule. I did not try to forecast the closing price of the S&P 500. I tried to forecast the biggest gap in everyone else's so-called "forecast". A number of other players have reported that they break that rule. For instance, David Grabiner. Their strategy is to not forecast the closing price.

I am suggesting that yo try to avoid misrepresenting the actual goal of your contest when you describe your contest. This contest is not really about forecasting the close of the S&P500. Providing the best forecast is very likely to provide a grossly sub-optimal chance of winning the contest.

A rule that I (approximately) follow is this rule:
Forecast the price that is closer to the closing price of the S&P 500 than any of the prices specified by your competitors.
(But exactly following that rule is a bit cut-throat because you will end up betting .01 above or below someone else's bet.)

Here's an interesting problem: Design a contest with a single prize where the optimal strategy is to forecast the S&P500. I am not sure it's even possible to design a contest that has this property.
Last edited by tadamsmar on Mon Apr 29, 2019 9:10 am, edited 1 time in total.

bgf
Posts: 969
Joined: Fri Nov 10, 2017 9:35 am

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by bgf » Mon Apr 29, 2019 9:08 am

tadamsmar wrote:
Mon Apr 29, 2019 9:01 am
livesoft wrote:
Mon Apr 29, 2019 7:46 am
Even your statement "It is about optimizing your chances of getting the prize" speaks to that much like winning a lottery.
I am not sure you understand the issue I am trying to raise. Let me try to explain more clearly.

Quoting from the OP of this thread, one of the rules is:
Forecast (below) the closing price of the S&P 500 Index on December 31, 2019.
I am competing in this contest and I broke that rule. I did not try to forecast the closing price of the S&P 500. I tried to forecast the biggest gap in everyone else's so-called "forecast". A number of other players have reported that they break that rule. For instance, David Grabiner. Their strategy is to not forecast the closing price.

I am suggesting that you try to avoid misrepresenting the actual goal of your contest when you describe your contest. This contest is not really about forecasting the close of the S&P500. Providing the best forecast is very likely to provide a grossly sub-optimal chance of winning the contest.

Here's an interesting problem: Design a contest with a single prize where the optimal strategy is to forecast the S&P500. I am not sure it's even possible to design a contest that has this property.
are you not confusing the task (forecast the closing of the SP500) with strategy (how to make one's forecast)?
“TE OCCIDERE POSSUNT SED TE EDERE NON POSSUNT NEFAS EST"

User avatar
bengal22
Posts: 1753
Joined: Sat Dec 03, 2011 6:20 pm
Location: Ohio

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by bengal22 » Mon Apr 29, 2019 9:13 am

This contest is designed to show the futility of forecasting the stock market year in and year out. Period.
"Earn All You Can; Give All You Can; Save All You Can." .... John Wesley

User avatar
tadamsmar
Posts: 8419
Joined: Mon May 07, 2007 12:33 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by tadamsmar » Mon Apr 29, 2019 9:16 am

bgf wrote:
Mon Apr 29, 2019 9:08 am
tadamsmar wrote:
Mon Apr 29, 2019 9:01 am
livesoft wrote:
Mon Apr 29, 2019 7:46 am
Even your statement "It is about optimizing your chances of getting the prize" speaks to that much like winning a lottery.
I am not sure you understand the issue I am trying to raise. Let me try to explain more clearly.

Quoting from the OP of this thread, one of the rules is:
Forecast (below) the closing price of the S&P 500 Index on December 31, 2019.
I am competing in this contest and I broke that rule. I did not try to forecast the closing price of the S&P 500. I tried to forecast the biggest gap in everyone else's so-called "forecast". A number of other players have reported that they break that rule. For instance, David Grabiner. Their strategy is to not forecast the closing price.

I am suggesting that you try to avoid misrepresenting the actual goal of your contest when you describe your contest. This contest is not really about forecasting the close of the S&P500. Providing the best forecast is very likely to provide a grossly sub-optimal chance of winning the contest.

Here's an interesting problem: Design a contest with a single prize where the optimal strategy is to forecast the S&P500. I am not sure it's even possible to design a contest that has this property.
are you not confusing the task (forecast the closing of the SP500) with strategy (how to make one's forecast)?
I don't think so. David Gabiner's entry and my entry are very different from the price that would be produced by any reasonable forecasting method or strategy.

The task is winning the prize. The task is not forecasting the close.

FinancialDave
Posts: 1648
Joined: Thu May 26, 2011 9:36 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by FinancialDave » Mon Apr 29, 2019 11:52 am

tadamsmar wrote:
Mon Apr 29, 2019 9:01 am
livesoft wrote:
Mon Apr 29, 2019 7:46 am
Even your statement "It is about optimizing your chances of getting the prize" speaks to that much like winning a lottery.
I am not sure you understand the issue I am trying to raise. Let me try to explain more clearly.

Quoting from the OP of this thread, one of the rules is:
Forecast (below) the closing price of the S&P 500 Index on December 31, 2019.
I am competing in this contest and I broke that rule. I did not try to forecast the closing price of the S&P 500. I tried to forecast the biggest gap in everyone else's so-called "forecast". A number of other players have reported that they break that rule. For instance, David Grabiner. Their strategy is to not forecast the closing price.

I am suggesting that yo try to avoid misrepresenting the actual goal of your contest when you describe your contest. This contest is not really about forecasting the close of the S&P500. Providing the best forecast is very likely to provide a grossly sub-optimal chance of winning the contest.

A rule that I (approximately) follow is this rule:
Forecast the price that is closer to the closing price of the S&P 500 than any of the prices specified by your competitors.
(But exactly following that rule is a bit cut-throat because you will end up betting .01 above or below someone else's bet.)

Here's an interesting problem: Design a contest with a single prize where the optimal strategy is to forecast the S&P500. I am not sure it's even possible to design a contest that has this property.
One way to avoid all this discussion is to just make the entries not visible until after the contest starts. Either by using a PM to the entry person, or any number of other ways which would make it at least difficult to "game" what you are suggesting.

Dave
I love simulated data. It turns the impossible into the possible!

FinancialDave
Posts: 1648
Joined: Thu May 26, 2011 9:36 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by FinancialDave » Mon Apr 29, 2019 12:02 pm

bengal22 wrote:
Mon Apr 29, 2019 9:13 am
This contest is designed to show the futility of forecasting the stock market year in and year out. Period.
Actually, the odds of winning are much better than buying a lottery ticket, because somebody always wins.

It's a fun game people -- not a life lesson. It got me a book signed by Jack Bogle, so I certainly think it was a worthwhile game to play.

:sharebeer
I love simulated data. It turns the impossible into the possible!

User avatar
tadamsmar
Posts: 8419
Joined: Mon May 07, 2007 12:33 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by tadamsmar » Mon Apr 29, 2019 1:44 pm

FinancialDave wrote:
Mon Apr 29, 2019 11:52 am
One way to avoid all this discussion is to just make the entries not visible until after the contest starts. Either by using a PM to the entry person, or any number of other ways which would make it at least difficult to "game" what you are suggesting.
That would make it much more difficult to game the contest. Some version of Gabiner's approach of picking away from the mean by some factor of the variance would still improve one's odds I think. We would still have the historical data to work with. So it does not really completely make the discussion moot. It's still not correct to present it as a contest to forecast the closing price.

User avatar
grabiner
Advisory Board
Posts: 24829
Joined: Tue Feb 20, 2007 11:58 pm
Location: Columbia, MD

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by grabiner » Mon Apr 29, 2019 8:50 pm

tadamsmar wrote:
Mon Apr 29, 2019 9:16 am
bgf wrote:
Mon Apr 29, 2019 9:08 am
are you not confusing the task (forecast the closing of the SP500) with strategy (how to make one's forecast)?
I don't think so. David Gabiner's entry and my entry are very different from the price that would be produced by any reasonable forecasting method or strategy.

The task is winning the prize. The task is not forecasting the close.
And I have even posted that this is what I do. If I expected the market to decline, I wouldn't have most of my portfolio in stock. But the market falls one year in three, and more than 2/3 of the contestants are bulls, so bears are more likely to win.
Wiki David Grabiner

User avatar
tadamsmar
Posts: 8419
Joined: Mon May 07, 2007 12:33 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by tadamsmar » Tue Apr 30, 2019 1:42 pm

grabiner wrote:
Mon Apr 29, 2019 8:50 pm
tadamsmar wrote:
Mon Apr 29, 2019 9:16 am
bgf wrote:
Mon Apr 29, 2019 9:08 am
are you not confusing the task (forecast the closing of the SP500) with strategy (how to make one's forecast)?
I don't think so. David Gabiner's entry and my entry are very different from the price that would be produced by any reasonable forecasting method or strategy.

The task is winning the prize. The task is not forecasting the close.
And I have even posted that this is what I do. If I expected the market to decline, I wouldn't have most of my portfolio in stock. But the market falls one year in three, and more than 2/3 of the contestants are bulls, so bears are more likely to win.
My analysis shows better chances in the bull region. Here's a plot of the win probabilities of a whole range of possible contest entry prices:

Image

This is based on historical mean growth and s.d.

Bears have entries that are about 200 points below the mean. By that definition, there are 190 bears. As you correctly stated, that is about 1/3 of entries.

But consider the other side of the symmetric Bell Curve, entries 200 points above the mean. There are only 94 entries 200 points above the mean.

The two regions are equally probable and there is less competition on the high end. So, it is reasonable to find that the high end is where the best bargains are.

User avatar
tadamsmar
Posts: 8419
Joined: Mon May 07, 2007 12:33 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by tadamsmar » Wed May 01, 2019 7:27 am

A projection of a 0% mean growth and 19% s.d. supports Grabiner's thesis:

Image

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by gilgamesh » Wed May 01, 2019 8:22 am

tadamsmar wrote:
Tue Apr 30, 2019 1:42 pm
grabiner wrote:
Mon Apr 29, 2019 8:50 pm
tadamsmar wrote:
Mon Apr 29, 2019 9:16 am
bgf wrote:
Mon Apr 29, 2019 9:08 am
are you not confusing the task (forecast the closing of the SP500) with strategy (how to make one's forecast)?
I don't think so. David Gabiner's entry and my entry are very different from the price that would be produced by any reasonable forecasting method or strategy.

The task is winning the prize. The task is not forecasting the close.
And I have even posted that this is what I do. If I expected the market to decline, I wouldn't have most of my portfolio in stock. But the market falls one year in three, and more than 2/3 of the contestants are bulls, so bears are more likely to win.
My analysis shows better chances in the bull region. Here's a plot of the win probabilities of a whole range of possible contest entry prices:

Image

This is based on historical mean growth and s.d.

Bears have entries that are about 200 points below the mean. By that definition, there are 190 bears. As you correctly stated, that is about 1/3 of entries.

But consider the other side of the symmetric Bell Curve, entries 200 points above the mean. There are only 94 entries 200 points above the mean.

The two regions are equally probable and there is less competition on the high end. So, it is reasonable to find that the high end is where the best bargains are.
Is there a way to quantify momentum? ... As in, most contestants do not predict a year in advance based on averages, the current position could have a larger influence. When stocks are rising they may predict larger than average, when stocks are sliding they may predict lower than average. So, competition might be higher on the high end than your data suggest during a bull market and vice versa.

Just curious whether that’s possible...

What else?...reversion to mean (negative momentum)? Or is that the same?

Then approximating how many of the boglehead contestants may go by averages, instinct etc. to decide how to apply such data. Then, of course correct for the percentage who would do the same or similar as what you are doing too, lol. But it could be stopped after a certain point.

Curious whether any other factors can approximate this calculations based on averages, even “better”?
Last edited by gilgamesh on Wed May 01, 2019 9:05 am, edited 3 times in total.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by gilgamesh » Wed May 01, 2019 8:35 am

FinancialDave wrote:
Mon Apr 29, 2019 11:52 am
tadamsmar wrote:
Mon Apr 29, 2019 9:01 am
livesoft wrote:
Mon Apr 29, 2019 7:46 am
Even your statement "It is about optimizing your chances of getting the prize" speaks to that much like winning a lottery.
I am not sure you understand the issue I am trying to raise. Let me try to explain more clearly.

Quoting from the OP of this thread, one of the rules is:
Forecast (below) the closing price of the S&P 500 Index on December 31, 2019.
I am competing in this contest and I broke that rule. I did not try to forecast the closing price of the S&P 500. I tried to forecast the biggest gap in everyone else's so-called "forecast". A number of other players have reported that they break that rule. For instance, David Grabiner. Their strategy is to not forecast the closing price.

I am suggesting that yo try to avoid misrepresenting the actual goal of your contest when you describe your contest. This contest is not really about forecasting the close of the S&P500. Providing the best forecast is very likely to provide a grossly sub-optimal chance of winning the contest.

A rule that I (approximately) follow is this rule:
Forecast the price that is closer to the closing price of the S&P 500 than any of the prices specified by your competitors.
(But exactly following that rule is a bit cut-throat because you will end up betting .01 above or below someone else's bet.)

Here's an interesting problem: Design a contest with a single prize where the optimal strategy is to forecast the S&P500. I am not sure it's even possible to design a contest that has this property.
One way to avoid all this discussion is to just make the entries not visible until after the contest starts. Either by using a PM to the entry person, or any number of other ways which would make it at least difficult to "game" what you are suggesting.

Dave
I think something like this is necessary...if a number is close to a previous prediction, then it will need to be corrected up or down until a certain gap from all other entries is found. But after a certain deviation from the prediction, the entry just needs to be disqualified as being too close to multiple other entries.

Without this, it’s obvious the intent of the challenge is totally destroyed.

User avatar
tadamsmar
Posts: 8419
Joined: Mon May 07, 2007 12:33 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by tadamsmar » Wed May 01, 2019 11:24 am

gilgamesh wrote:
Wed May 01, 2019 8:22 am
tadamsmar wrote:
Tue Apr 30, 2019 1:42 pm
grabiner wrote:
Mon Apr 29, 2019 8:50 pm
tadamsmar wrote:
Mon Apr 29, 2019 9:16 am
bgf wrote:
Mon Apr 29, 2019 9:08 am
are you not confusing the task (forecast the closing of the SP500) with strategy (how to make one's forecast)?
I don't think so. David Gabiner's entry and my entry are very different from the price that would be produced by any reasonable forecasting method or strategy.

The task is winning the prize. The task is not forecasting the close.
And I have even posted that this is what I do. If I expected the market to decline, I wouldn't have most of my portfolio in stock. But the market falls one year in three, and more than 2/3 of the contestants are bulls, so bears are more likely to win.
My analysis shows better chances in the bull region. Here's a plot of the win probabilities of a whole range of possible contest entry prices:

Image

This is based on historical mean growth and s.d.

Bears have entries that are about 200 points below the mean. By that definition, there are 190 bears. As you correctly stated, that is about 1/3 of entries.

But consider the other side of the symmetric Bell Curve, entries 200 points above the mean. There are only 94 entries 200 points above the mean.

The two regions are equally probable and there is less competition on the high end. So, it is reasonable to find that the high end is where the best bargains are.
Is there a way to quantify momentum? ... As in, most contestants do not predict a year in advance based on averages, the current position could have a larger influence. When stocks are rising they may predict larger than average, when stocks are sliding they may predict lower than average. So, competition might be higher on the high end than your data suggest during a bull market and vice versa.

Just curious whether that’s possible...

What else?...reversion to mean (negative momentum)? Or is that the same?

Then approximating how many of the boglehead contestants may go by averages, instinct etc. to decide how to apply such data. Then, of course correct for the percentage who would do the same or similar as what you are doing too, lol. But it could be stopped after a certain point.

Curious whether any other factors can approximate this calculations based on averages, even “better”?
I don't feel like I understand what you are getting at.

The only approximations are the mean growth and s.d. and the assumption that the distribution is normal.

Otherwise, the plot is based on the actual entries by the contestants. (I call them "entries" because I think that is a neutral term. The OP called the "forecasts" or "guesses" of the closing price, but I think those terms are theory-laden.)

The plot shows the win probabilities over a range of possible whole-number entries that could be made a few seconds before the entry deadline on the assumption that there will be no additional entries. In the plot, entries identical to opposing whole-number entries in that range are given a zero probability since they would be duplicates that have no chance of winning.

I think part of what you are saying is that the opposing entries might tend toward the high side in a different year. I suppose that could be true.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by gilgamesh » Wed May 01, 2019 12:30 pm

You picked mean growth of 8%, and I think you said grabiner’s hypothesis will be correct with 0% growth (may be, I misunderstood)

What’s the rationale for picking mean growth of 8%...is that the best option?

That’s what I was trying to get at? Thanks!

User avatar
tadamsmar
Posts: 8419
Joined: Mon May 07, 2007 12:33 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by tadamsmar » Wed May 01, 2019 1:04 pm

gilgamesh wrote:
Wed May 01, 2019 12:30 pm
You picked mean growth of 8%, and I think you said grabiner’s hypothesis will be correct with 0% growth (may be, I misunderstood)
You have that correct.
What’s the rationale for picking mean growth of 8%...is that the best option?

That’s what I was trying to get at? Thanks!
The average historic growth of the S&P 500 price is roughly 8%. That's the rationale. (BTW, I find it hard to come up the historic price growth. Not sure what the exact value is.)

I am not sure if it is the best option.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by gilgamesh » Wed May 01, 2019 1:08 pm

Delete

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by gilgamesh » Thu May 02, 2019 7:59 am

tadamsmar wrote:
Tue Apr 30, 2019 1:42 pm


My analysis shows better chances in the bull region. Here's a plot of the win probabilities of a whole range of possible contest entry prices:

Image

This is based on historical mean growth and s.d.

Bears have entries that are about 200 points below the mean. By that definition, there are 190 bears. As you correctly stated, that is about 1/3 of entries.

But consider the other side of the symmetric Bell Curve, entries 200 points above the mean. There are only 94 entries 200 points above the mean.

The two regions are equally probable and there is less competition on the high end. So, it is reasonable to find that the high end is where the best bargains are.
tadamsmar wrote:
Wed May 01, 2019 11:24 am

I don't feel like I understand what you are getting at.

The only approximations are the mean growth and s.d. and the assumption that the distribution is normal.

Otherwise, the plot is based on the actual entries by the contestants. (I call them "entries" because I think that is a neutral term. The OP called the "forecasts" or "guesses" of the closing price, but I think those terms are theory-laden.)

The plot shows the win probabilities over a range of possible whole-number entries that could be made a few seconds before the entry deadline on the assumption that there will be no additional entries. In the plot, entries identical to opposing whole-number entries in that range are given a zero probability since they would be duplicates that have no chance of winning.

I think part of what you are saying is that the opposing entries might tend toward the high side in a different year. I suppose that could be true.

I was getting tired yesterday so just deleted my post...

What I have put in bold above...can you explain what is based on ‘all probable contest entry” and what’s the actual entry.

I assumed the X axis is all possible contest entries and then your probability of winning the contest (the graph) is based on the actual entries.

Anyhow, the main point is, I thought you assumed the actual S&P 500 would follow the historical average for the year in question (2019) to come up with your winning chances (may be I am wrong, and you didn’t). I was wondering whether that assumption of 2019 just following the historical average can be fine tuned via concepts like momentum.

User avatar
tadamsmar
Posts: 8419
Joined: Mon May 07, 2007 12:33 pm

Re: REGISTRATION FOR 2019 BOGLEHEAD CONTEST [Entry deadline has passed]

Post by tadamsmar » Thu May 02, 2019 2:31 pm

gilgamesh wrote:
Thu May 02, 2019 7:59 am
What I have put in bold above...can you explain what is based on ‘all probable contest entry” and what’s the actual entry.

I assumed the X axis is all possible contest entries and then your probability of winning the contest (the graph) is based on the actual entries.
The X axis range meant to show the range of the decision variable. But it is really only part of the range. The S&P 500 can take on any value between 0 and infinity inclusive, rounded to 2 decimal places, that is your decision space. But the X-axis is just the integers between 1275 and 4060 or so.

The actual entries are listed here:

https://www.lostoak.com/ls/diehards/con ... fault.aspx

with the non-competing entries of wall street guru predictions (added by Ralph Arvesen) removed.

The idea is that you are deciding on an entry price just before the deadline so that you likely know all the actual entries of your opponents. In this analysis, it is assumed that you know what all the actual entries of your opponents will be after the deadline.
Anyhow, the main point is, I thought you assumed the actual S&P 500 would follow the historical average for the year in question (2019) to come up with your winning chances (may be I am wrong, and you didn’t). I was wondering whether that assumption of 2019 just following the historical average can be fine tuned via concepts like momentum.
The assumption about 2019 could be based on other foundations. And I suppose those might be better.

Post Reply