I would add, though, that one should be careful expecting calendar years to have the same statistical perspectives as some other randomly selected set of “trading years.” The set of calendar years is a biased selection. While we might not think it should differ from other “years” with random starting days (March 7 for example) other than January 1, we shouldn’t take that assumption for granted. It’s probably much more likely than not that calendar years differ in meaningful ways from random “years.”
-
01
-
Ghost of Igloi wrote:
la gente esta muy loca wrote:
Great book, read that years ago, started reading his The Deluge a few years back, plan to get back to it this year.
I enjoyed it. Loaned it to a friend and he gave up.
I am headed your way the weekend of May 4th. One of my former teams is celebrating a 40th reunion of a cross country conference win, a little early, but doing it in conjunction with the Texas State 3200 meter races. In Buda where one of my, actually two of my former athletes, husband and wife have a home.
Not a single person cares about your pathetic, uninteresting loser lifestyle. You’re a nobody. Keep losing money being a permabear moron lmao.
00 -
Slow old n poor in TX wrote:
Ghost of Igloi wrote:
I enjoyed it. Loaned it to a friend and he gave up.
I am headed your way the weekend of May 4th. One of my former teams is celebrating a 40th reunion of a cross country conference win, a little early, but doing it in conjunction with the Texas State 3200 meter races. In Buda where one of my, actually two of my former athletes, husband and wife have a home.
Not a single person cares about your pathetic, uninteresting loser lifestyle. You’re a nobody. Keep losing money being a permabear moron lmao.
Pathetic is you getting off once again for drug and alcohol DUI. That pending case may finally land you in prison.
22 -
01
-
la gente esta muy loca wrote:
la gente esta muy loca wrote:
As the disclaimer says in plain English, past performance is not indicative of future results. I'm pushing back against the idea that in the past, a bad year for the NASDAQ will always be followed by a gain. Another comparison between a good year and a bad year is the historical average loss and it's standard deviation. A down year after a good year had an average loss of -16.65% and a SD of 14.04; so 1 SD -2.61 to -30.69%. A down year after a down year had an average loss of -25.3% and a SD of 10.4; so 1 SD -14.9% to -35.7%. There was more risk after a down year than an up year.
Some Saturday morning fun, when I was young that meant Andy's Gang, but this AM I going to get a little wonky. Let's pretend the NASDAQ Index is like playing Blackjack (I've taught my grandchildren how to play ) and the probabilities are constant. Overall probability of Index being higher is 75.41% and when higher it's +22.74%; when lower it's -19.05%. After 252 plays, if the Index is higher, the probability it will be higher 252 plays later is 76.08% and when higher it's 20.55% ; when lower it's -16.65%. If after 252 plays the Index is lower, the probability it will be higher 252 plays later is 71.4% and if higher it will be 30.9% ; lower it's -25.3%. I'll make it simpler. Imagine a large spinning wheel, on the 1st spin there are 75.41% higher slots and 24.59% lower slots. If it lands on a higher slot, you win $22.74 for every $100 you bet. If it lands on a lower slot you lose $19.05 for every $100. Now, if the first spin lands on higher, the slots change to 76.08% higher and 23.92% lower. The payout-loss is $20.55; -$16.65. If the next spin is lower, the slots change to 71.4% higher and 28.6% lower. The payout-loss is $30.90; -$25.30
How much of your bankroll would you wager to start the game? After a higher result? After a lower result? The goal is to maximize your return. I'll come back to this Sunday.
I'd wager essentially most of my bankroll and I would keep that constant after both a high result and a low result.
I say that because while the payouts and losses are relatively the same amounts (maybe a range of about 25% difference between the two), the chances of a spin being higher vastly outweigh the chances of it being lower by about 3 to 1 ratio.
So, in a nutshell, vastly greater chance of higher spin, regardless of the whether the previous 252 plays were higher or lower, coupled with the fact that the payout/loss is only slightly different regardless of the outcome. Therefore: much greater chance to win on higher spin, relatively small penalty for being wrong, and this holds true regardless of whether previous 252 spins were higher or lower.
00 -
comfortably retired on the ipad wrote:
I would add, though, that one should be careful expecting calendar years to have the same statistical perspectives as some other randomly selected set of “trading years.” The set of calendar years is a biased selection. While we might not think it should differ from other “years” with random starting days (March 7 for example) other than January 1, we shouldn’t take that assumption for granted. It’s probably much more likely than not that calendar years differ in meaningful ways from random “years.”
"MY assumption of probabilities is MUCH more reasonable."
...............total bs
01 -
seattle prattle wrote:
la gente esta muy loca wrote:
Some Saturday morning fun, when I was young that meant Andy's Gang, but this AM I going to get a little wonky. Let's pretend the NASDAQ Index is like playing Blackjack (I've taught my grandchildren how to play ) and the probabilities are constant. Overall probability of Index being higher is 75.41% and when higher it's +22.74%; when lower it's -19.05%. After 252 plays, if the Index is higher, the probability it will be higher 252 plays later is 76.08% and when higher it's 20.55% ; when lower it's -16.65%. If after 252 plays the Index is lower, the probability it will be higher 252 plays later is 71.4% and if higher it will be 30.9% ; lower it's -25.3%. I'll make it simpler. Imagine a large spinning wheel, on the 1st spin there are 75.41% higher slots and 24.59% lower slots. If it lands on a higher slot, you win $22.74 for every $100 you bet. If it lands on a lower slot you lose $19.05 for every $100. Now, if the first spin lands on higher, the slots change to 76.08% higher and 23.92% lower. The payout-loss is $20.55; -$16.65. If the next spin is lower, the slots change to 71.4% higher and 28.6% lower. The payout-loss is $30.90; -$25.30
How much of your bankroll would you wager to start the game? After a higher result? After a lower result? The goal is to maximize your return. I'll come back to this Sunday.
I'd wager essentially most of my bankroll and I would keep that constant after both a high result and a low result.
I say that because while the payouts and losses are relatively the same amounts (maybe a range of about 25% difference between the two), the chances of a spin being higher vastly outweigh the chances of it being lower by about 3 to 1 ratio.
So, in a nutshell, vastly greater chance of higher spin, regardless of the whether the previous 252 plays were higher or lower, coupled with the fact that the payout/loss is only slightly different regardless of the outcome. Therefore: much greater chance to win on higher spin, relatively small penalty for being wrong, and this holds true regardless of whether previous 252 spins were higher or lower.
Would you change your mind if a told you after 10 spins (years) there is a 22.62% chance that your wager amount is the same or less and a 10.66% chance it's 40% or less? Data is from running a Monte Carlo Simulation on the ETF QQQ. Keep in mind the NASDAQ 100 has outperformed the NASDAQ Index by a wide margin and the simulation uses total returns. Also I believe the experiment should be limited to 40 spins (years).
00 -
la gente esta muy loca wrote:
seattle prattle wrote:
I'd wager essentially most of my bankroll and I would keep that constant after both a high result and a low result.
I say that because while the payouts and losses are relatively the same amounts (maybe a range of about 25% difference between the two), the chances of a spin being higher vastly outweigh the chances of it being lower by about 3 to 1 ratio.
So, in a nutshell, vastly greater chance of higher spin, regardless of the whether the previous 252 plays were higher or lower, coupled with the fact that the payout/loss is only slightly different regardless of the outcome. Therefore: much greater chance to win on higher spin, relatively small penalty for being wrong, and this holds true regardless of whether previous 252 spins were higher or lower.
Would you change your mind if a told you after 10 spins (years) there is a 22.62% chance that your wager amount is the same or less and a 10.66% chance it's 40% or less? Data is from running a Monte Carlo Simulation on the ETF QQQ. Keep in mind the NASDAQ 100 has outperformed the NASDAQ Index by a wide margin and the simulation uses total returns. Also I believe the experiment should be limited to 40 spins (years).
P.S. I should had stated portfolio not used to wager will be invested in 3 month T-Bills. I chose 40 because it would match the average investors timeline.
00 -
la gente esta muy loca wrote:
la gente esta muy loca wrote:
Would you change your mind if a told you after 10 spins (years) there is a 22.62% chance that your wager amount is the same or less and a 10.66% chance it's 40% or less? Data is from running a Monte Carlo Simulation on the ETF QQQ. Keep in mind the NASDAQ 100 has outperformed the NASDAQ Index by a wide margin and the simulation uses total returns. Also I believe the experiment should be limited to 40 spins (years).
P.S. I should had stated portfolio not used to wager will be invested in 3 month T-Bills. I chose 40 because it would match the average investors timeline.
I'm not sure I understand this latest iteration. What is the payoff amount assuming a payout? You only mention the spin in terms of loss.
00 -
seattle prattle wrote:
la gente esta muy loca wrote:
P.S. I should had stated portfolio not used to wager will be invested in 3 month T-Bills. I chose 40 because it would match the average investors timeline.
I'm not sure I understand this latest iteration. What is the payoff amount assuming a payout? You only mention the spin in terms of loss.
What I mean to say, e.g. you're playing your whole bankroll, starting with $100 and keep playing all of it regardless of wins and losses . Using a Monte Carlo Simulator of QQQ historical results, after 10 spins there's a 22.62% chance your bankroll is now $100 or less and a 10.66% it's $60 or less. Playing just $95 of $100 will reduce these loss probabilities; depending on whether you're just betting 95% of your remaining funds or all of your remaining funds after each spin.
There's an an anomaly here between the odds I gave and the results from using the Monte Carlo Simulator of the QQQ ETF. Hint, the NASDAQ Index data starts in 1971 and QQQ data starts in 1999. The real purpose of this experiment is to demonstrate the perils of backtesting. I'll come back to this tomorrow, I have to console my cat after the Lion's loss.
10 -
Waller is a fed reserve guvner.
Hey guess what - it WAS different this time.
Alf
@MacroAlf
*WALLER PLAYS DOWN SIGNAL FROM RECENT INVERSION OF YIELD CURVE
Of course.
As they did in:
- 1989 (recession in 1990)
- 2000 (recession in 2001)
- 2007 (recession in 2008)
2022: ''this time is different''
7:01 PM · Jul 7, 202210 -
Good technical call a year ago.
SP500 +22% since then.
Seth Golden
@SethCL
🎯 Woomp, there it is! A NEW BULL MARKET...
S&P 500 $SPX 50-DMA has CROSSED above 150-DMA for 1st time in current cycle. Historically this is a signal of a NEW BULL MARKET $SPY9:19 AM · Jan 22, 2023
·
178.1K
Views10 -
la gente esta muy loca wrote:
The real purpose of this experiment is to demonstrate the perils of backtesting. I'll come back to this tomorrow, I have to console my cat after the Lion's loss.
This is an important topic. There’s a powerful inclination among investors to “learn from” past market behaviour. The unfortunate truth (IMHO) is that nearly all observable market movement is either truly random or what I will call pseudo-random, by which I mean there might be some underlying deterministic relationship, but all mathematical tests would suggest random behaviour, and deterministic explanations can only be posed relatively accurately and consistently in hindsight. There’s almost nothing useful that the average investor (non-insider) can “predict” about future market behaviour from the past, except that stocks “usually” go up (except when they don’t), and they fluctuate (sometimes wildly) along their (mostly) upward path.
00 -
comfortably retired on the ipad wrote:
la gente esta muy loca wrote:
The real purpose of this experiment is to demonstrate the perils of backtesting. I'll come back to this tomorrow, I have to console my cat after the Lion's loss.
This is an important topic. There’s a powerful inclination among investors to “learn from” past market behaviour. The unfortunate truth (IMHO) is that nearly all observable market movement is either truly random or what I will call pseudo-random, by which I mean there might be some underlying deterministic relationship, but all mathematical tests would suggest random behaviour, and deterministic explanations can only be posed relatively accurately and consistently in hindsight. There’s almost nothing useful that the average investor (non-insider) can “predict” about future market behaviour from the past, except that stocks “usually” go up (except when they don’t), and they fluctuate (sometimes wildly) along their (mostly) upward path.
I wonder what you all think about investing strategies like 'factor' investing.
The sort of thing where a formula picks the 50 stocks with high cash flow yield and then absolute cash flow dollars and rebalances every quarter or year.
they are the children of backtesting but do seem to work. For a while anyway. Until they stop working, which I suppose is the point.
Funds like CALF, COWZ and HFMDX are successful factor funds. There are many others.
10 -
comfortably retired on the ipad wrote:
la gente esta muy loca wrote:
The real purpose of this experiment is to demonstrate the perils of backtesting. I'll come back to this tomorrow, I have to console my cat after the Lion's loss.
This is an important topic. There’s a powerful inclination among investors to “learn from” past market behaviour. The unfortunate truth (IMHO) is that nearly all observable market movement is either truly random or what I will call pseudo-random, by which I mean there might be some underlying deterministic relationship, but all mathematical tests would suggest random behaviour, and deterministic explanations can only be posed relatively accurately and consistently in hindsight. There’s almost nothing useful that the average investor (non-insider) can “predict” about future market behaviour from the past, except that stocks “usually” go up (except when they don’t), and they fluctuate (sometimes wildly) along their (mostly) upward path.
Dow1923: 73.0
Dow2023 (100 years later): 38,200
It has done a lot of "going up."
00 -
Sally Vix wrote:
Dow1923: 73.0
Dow2023 (100 years later): 38,200
It has done a lot of "going up."
True, nobody would suggest otherwise. If I may, though, I will point out that your numbers, if correct, suggest long-term annual returns of 6.4%, if I've done the math correctly. That's not TOO far off the risk-free return from long-term average yield of 10 to 30 year treasuries.
10 -
comfortably retired wrote:
Sally Vix wrote:
Dow1923: 73.0
Dow2023 (100 years later): 38,200
It has done a lot of "going up."
True, nobody would suggest otherwise. If I may, though, I will point out that your numbers, if correct, suggest long-term annual returns of 6.4%, if I've done the math correctly. That's not TOO far off the risk-free return from long-term average yield of 10 to 30 year treasuries.
Are you only considering capital appreciations? This from memory - but over the last 80 or 90 years dividends have provided about 40% of the S & P total return.
00 -
Interesting article in today's Yahoo News about a familiar topic here, are valuations too high and does it matter. Apparently it is up to debate, calling the excessive fear of their importance a "Valuation trap."
00 -
Sally Vix wrote:
Are you only considering capital appreciations? This from memory - but over the last 80 or 90 years dividends have provided about 40% of the S & P total return.
I just took YOUR numbers for the DOW, which I imagine do not represent total returns (i.e., do not include dividends), but you will know better than me where you got the numbers from and what they include.
Of course dividends are important and my comparison was probably too glib. That said, the absence of risk has value, maybe a lot to some investors, varying according to personal objectives, time horizon and psychological fortitude.
00 -
All time highs SP500 and Dow
Well, I looked at my watch, it was 9:54
I said, "Dance ballerina girl, go, go, go go"
And we go, reelin' and rockin'
We was reelin' and rockin', and rollin' 'til the break of dawn10
What People Are Talking About On LetsRun
No top threads at the moment. Check back soon.