September losses

[Silk McCue]

Curious if anyone has seen an on-line back tester than could answer the question -

S&P and relax vs. S&P, sell Aug 31, buy it back Oct 1.

That would be a fools errand. The stock market doesn’t know calendar date returns. Especially for one that hasn’t happened yet.

Cheers

Reread my post. I ask about back testing, not forecasting.

No need to reread. Backtesting provides no insight into the next cycle.

Cheers

[OKOKOK]

Curious if anyone has seen an on-line back tester than could answer the question -

S&P and relax vs. S&P, sell Aug 31, buy it back Oct 1.

That would be a fools errand. The stock market doesn’t know calendar date returns. Especially for one that hasn’t happened yet.

Cheers

Reread my post. I ask about back testing, not forecasting.

No need to reread. Backtesting provides no insight into the next cycle.

Cheers

So if you burned your hand on a hot stove 97 out of 100 times you touched it, over your lifetime, you would not take that knowledge and apply it going forward?

[Silk McCue]

Curious if anyone has seen an on-line back tester than could answer the question -

S&P and relax vs. S&P, sell Aug 31, buy it back Oct 1.

That would be a fools errand. The stock market doesn’t know calendar date returns. Especially for one that hasn’t happened yet.

Cheers

Reread my post. I ask about back testing, not forecasting.

No need to reread. Backtesting provides no insight into the next cycle.

Cheers

So if you burned your hand on a hot stove 97 out of 100 times you touched it, over your lifetime, you would not take that knowledge and apply it going forward?

Market timing is a losers game. You are either a long term low cost diversified investor or you are a trader.

Cheers

[OKOKOK]

Touchy subject here.
Testing a theory with back testing should be encouraged.

[rkhusky]

Curious if anyone has seen an on-line back tester than could answer the question -
S&P and relax vs. S&P, sell Aug 31, buy it back Oct 1.

No tool, but for curiosity I did it via spreadsheet for VTSAX from Jan 2001. August was worse and October best, but with large error bars.

Average StDev
Jan -0.1% 4.6%
Feb 0.6% 4.5%
Mar 2.4% 5.0%
Apr 0.8% 3.7%
May -0.5% 4.1%
Jun 2.1% 4.1%
Jul 0.0% 3.5%
Aug -1.8% 5.2%
Sep 2.0% 6.1%
Oct 2.8% 4.3%
Nov 0.5% 3.8%
Dec 0.6% 4.6%

And here are results for VFINX from Jan 1986 (August still worse, March best)
Average StDev
Jan 0.7% 4.3%
Feb 0.9% 4.1%
Mar 2.1% 4.2%
Apr 1.4% 3.5%
May 0.0% 3.7%
Jun 2.0% 4.0%
Jul -0.1% 4.5%
Aug -1.3% 4.8%
Sep 1.5% 6.3%
Oct 2.0% 4.5%
Nov 0.5% 4.1%
Dec 1.9% 4.8%

The standard deviation of the averages is much less than the standard deviation of the monthly returns, indicating that the noise is quite a bit larger than the signal.

[OKOKOK]

Curious if anyone has seen an on-line back tester than could answer the question -
S&P and relax vs. S&P, sell Aug 31, buy it back Oct 1.

No tool, but for curiosity I did it via spreadsheet for VTSAX from Jan 2001. August was worse and October best, but with large error bars.

Average StDev
Jan -0.1% 4.6%
Feb 0.6% 4.5%
Mar 2.4% 5.0%
Apr 0.8% 3.7%
May -0.5% 4.1%
Jun 2.1% 4.1%
Jul 0.0% 3.5%
Aug -1.8% 5.2%
Sep 2.0% 6.1%
Oct 2.8% 4.3%
Nov 0.5% 3.8%
Dec 0.6% 4.6%

And here are results for VFINX from Jan 1986 (August still worse, March best)
Average StDev
Jan 0.7% 4.3%
Feb 0.9% 4.1%
Mar 2.1% 4.2%
Apr 1.4% 3.5%
May 0.0% 3.7%
Jun 2.0% 4.0%
Jul -0.1% 4.5%
Aug -1.3% 4.8%
Sep 1.5% 6.3%
Oct 2.0% 4.5%
Nov 0.5% 4.1%
Dec 1.9% 4.8%

The standard deviation of the averages is much less than the standard deviation of the monthly returns, indicating that the noise is quite a bit larger than the signal.

Thanks!!!! I'm on cell this weekend with no desktop. Could you throw VTSAX into a chart ?

[blortchplop]

With standard deviation multiple times the average, this is essentially noise.

Greenday has a song titled Wake Me Up When September Ends, although I could have sworn it wasn't about investing.

[avalpert1]

Curious if anyone has seen an on-line back tester than could answer the question -

S&P and relax vs. S&P, sell Aug 31, buy it back Oct 1.

That would be a fools errand. The stock market doesn’t know calendar date returns. Especially for one that hasn’t happened yet.

Cheers

Reread my post. I ask about back testing, not forecasting.

No need to reread. Backtesting provides no insight into the next cycle.

Cheers

So if you burned your hand on a hot stove 97 out of 100 times you touched it, over your lifetime, you would not take that knowledge and apply it going forward?

The stock market doesn't operate under laws like physics and thermodynamics...

[avalpert1]

Touchy subject here.
Testing a theory with back testing should be encouraged.

Not really no. First, you haven't articulated a 'theory' of any sorts and fishing for correlations without a solid theory behind what you are looking for is a recipe for disaster. Second, backtesting alone doesn't really offer you assurance of the validity of what you tested particularly when the variance will dwarf the signal. Third, encouraging random goose chasing isn't productive for investors and is very likely to lead to poor decisions, it may be producing for PhD candidates who can't find a topic but even then it rarely ends in something successful let alone something that informs investing.

[blortchplop]

Touchy subject here.
Testing a theory with back testing should be encouraged.

Not really no. First, you haven't articulated a 'theory' of any sorts and fishing for correlations without a solid theory behind what you are looking for is a recipe for disaster. Second, backtesting alone doesn't really offer you assurance of the validity of what you tested particularly when the variance will dwarf the signal. Third, encouraging random goose chasing isn't productive for investors and is very likely to lead to poor decisions, it may be producing for PhD candidates who can't find a topic but even then it rarely ends in something successful let alone something that informs investing.

It would have been jaw dropping if the backtest showed a strong, durable effect for the month of September only. I would love to read a short fiction story about that happening and how bogleheads might respond.

Regarding the bolded part of what you wrote, it is entirely possible for investing theories themselves to be poorly founded. I suppose it's less likely for a baseless theory to correspond to a strong signal in a backtest compared to just fishing, but those stars could align...

You did say solid theory, so perhaps it is a moot point. But one could still go fishing with bad theories and may even catch something. I will bet this has happened at least once in history.

[OKOKOK]

Touchy subject here.
Testing a theory with back testing should be encouraged.

Not really no. First, you haven't articulated a 'theory' of any sorts and fishing for correlations without a solid theory behind what you are looking for is a recipe for disaster. Second, backtesting alone doesn't really offer you assurance of the validity of what you tested particularly when the variance will dwarf the signal. Third, encouraging random goose chasing isn't productive for investors and is very likely to lead to poor decisions, it may be producing for PhD candidates who can't find a topic but even then it rarely ends in something successful let alone something that informs investing.

It would have been jaw dropping if the backtest showed a strong, durable effect for the month of September only. I would love to read a short fiction story about that happening and how bogleheads might respond.

Regarding the bolded part of what you wrote, it is entirely possible for investing theories themselves to be poorly founded. I suppose it's less likely for a baseless theory to correspond to a strong signal in a backtest compared to just fishing, but those stars could align...

You did say solid theory, so perhaps it is a moot point. But one could still go fishing with bad theories and may even catch something. I will bet this has happened at least once in history.

Well said.
Based on the backtesting, the theory doesn't stand-up. But to say don't backrest the theory, well that's just blindly following a methodology. Sorry Bogleheads ....

[avalpert1]

Touchy subject here.
Testing a theory with back testing should be encouraged.

Not really no. First, you haven't articulated a 'theory' of any sorts and fishing for correlations without a solid theory behind what you are looking for is a recipe for disaster. Second, backtesting alone doesn't really offer you assurance of the validity of what you tested particularly when the variance will dwarf the signal. Third, encouraging random goose chasing isn't productive for investors and is very likely to lead to poor decisions, it may be producing for PhD candidates who can't find a topic but even then it rarely ends in something successful let alone something that informs investing.

It would have been jaw dropping if the backtest showed a strong, durable effect for the month of September only. I would love to read a short fiction story about that happening and how bogleheads might respond.

Regarding the bolded part of what you wrote, it is entirely possible for investing theories themselves to be poorly founded. I suppose it's less likely for a baseless theory to correspond to a strong signal in a backtest compared to just fishing, but those stars could align...

You did say solid theory, so perhaps it is a moot point. But one could still go fishing with bad theories and may even catch something. I will bet this has happened at least once in history.

Well said.
Based on the backtesting, the theory doesn't stand-up. But to say don't backrest the theory, well that's just blindly following a methodology. Sorry Bogleheads ....

Actually, using backtesting to find what months perform better/worse would be blindly following a methodology... asking for actually formed theories that include the mechanism of affect before fishing for patterns is just solid analysis.

[OKOKOK]

Touchy subject here.
Testing a theory with back testing should be encouraged.

Not really no. First, you haven't articulated a 'theory' of any sorts and fishing for correlations without a solid theory behind what you are looking for is a recipe for disaster. Second, backtesting alone doesn't really offer you assurance of the validity of what you tested particularly when the variance will dwarf the signal. Third, encouraging random goose chasing isn't productive for investors and is very likely to lead to poor decisions, it may be producing for PhD candidates who can't find a topic but even then it rarely ends in something successful let alone something that informs investing.

It would have been jaw dropping if the backtest showed a strong, durable effect for the month of September only. I would love to read a short fiction story about that happening and how bogleheads might respond.

Regarding the bolded part of what you wrote, it is entirely possible for investing theories themselves to be poorly founded. I suppose it's less likely for a baseless theory to correspond to a strong signal in a backtest compared to just fishing, but those stars could align...

You did say solid theory, so perhaps it is a moot point. But one could still go fishing with bad theories and may even catch something. I will bet this has happened at least once in history.

Well said.
Based on the backtesting, the theory doesn't stand-up. But to say don't backrest the theory, well that's just blindly following a methodology. Sorry Bogleheads ....

Actually, using backtesting to find what months perform better/worse would be blindly following a methodology... asking for actually formed theories that include the mechanism of affect before fishing for patterns is just solid analysis.

Backtesting to find what months perform better/worse would be blindly following a methodology? Perhaps the methodology of questioning things. I am proud to reside in that camp.

[avalpert1]

Not really no. First, you haven't articulated a 'theory' of any sorts and fishing for correlations without a solid theory behind what you are looking for is a recipe for disaster. Second, backtesting alone doesn't really offer you assurance of the validity of what you tested particularly when the variance will dwarf the signal. Third, encouraging random goose chasing isn't productive for investors and is very likely to lead to poor decisions, it may be producing for PhD candidates who can't find a topic but even then it rarely ends in something successful let alone something that informs investing.

It would have been jaw dropping if the backtest showed a strong, durable effect for the month of September only. I would love to read a short fiction story about that happening and how bogleheads might respond.

Regarding the bolded part of what you wrote, it is entirely possible for investing theories themselves to be poorly founded. I suppose it's less likely for a baseless theory to correspond to a strong signal in a backtest compared to just fishing, but those stars could align...

You did say solid theory, so perhaps it is a moot point. But one could still go fishing with bad theories and may even catch something. I will bet this has happened at least once in history.

Well said.
Based on the backtesting, the theory doesn't stand-up. But to say don't backrest the theory, well that's just blindly following a methodology. Sorry Bogleheads ....

Actually, using backtesting to find what months perform better/worse would be blindly following a methodology... asking for actually formed theories that include the mechanism of affect before fishing for patterns is just solid analysis.

Backtesting to find what months perform better/worse would be blindly following a methodology? Perhaps the methodology of questioning things. I am proud to reside in that camp.

No, the methodology of looking for random patterns in backtests and for no discernible reason assuming they have meaning.

Why would you think any month should consistently perform better or worse?

[DesertGator]

Don’t you think it’s unlikely that a bunch of yahoos like ourselves on a public form like this are going to discover something this simple that all of the full time analysts and data scientists on Wall Street, with vast computer modeling skills, over many years, haven’t figured it out and exploited it already?

And you know once the bozos find out about it, it’s too late [Paraphrasing Bernstein quoting Rekenthaler]

[rkhusky]

Curious if anyone has seen an on-line back tester than could answer the question -
S&P and relax vs. S&P, sell Aug 31, buy it back Oct 1.

No tool, but for curiosity I did it via spreadsheet for VTSAX from Jan 2001. August was worse and October best, but with large error bars.

Average StDev
Jan -0.1% 4.6%
Feb 0.6% 4.5%
Mar 2.4% 5.0%
Apr 0.8% 3.7%
May -0.5% 4.1%
Jun 2.1% 4.1%
Jul 0.0% 3.5%
Aug -1.8% 5.2%
Sep 2.0% 6.1%
Oct 2.8% 4.3%
Nov 0.5% 3.8%
Dec 0.6% 4.6%

And here are results for VFINX from Jan 1986 (August still worse, March best)
Average StDev
Jan 0.7% 4.3%
Feb 0.9% 4.1%
Mar 2.1% 4.2%
Apr 1.4% 3.5%
May 0.0% 3.7%
Jun 2.0% 4.0%
Jul -0.1% 4.5%
Aug -1.3% 4.8%
Sep 1.5% 6.3%
Oct 2.0% 4.5%
Nov 0.5% 4.1%
Dec 1.9% 4.8%

The standard deviation of the averages is much less than the standard deviation of the monthly returns, indicating that the noise is quite a bit larger than the signal.

Thanks!!!! I'm on cell this weekend with no desktop. Could you throw VTSAX into a chart ?

Here you go. The error bars are 1 standard deviation. For 95% confidence, double the error bars, assuming the monthly returns follow a normal distribution. If the null hypothesis is that monthly returns are all the same, the data does not disprove it.

[muffins14]

Touchy subject here.
Testing a theory with back testing should be encouraged.

The theory is nonsensical. It’s like saying stocks always go up on Tuesdays.

[muffins14]

Touchy subject here.
Testing a theory with back testing should be encouraged.

Not really no. First, you haven't articulated a 'theory' of any sorts and fishing for correlations without a solid theory behind what you are looking for is a recipe for disaster. Second, backtesting alone doesn't really offer you assurance of the validity of what you tested particularly when the variance will dwarf the signal. Third, encouraging random goose chasing isn't productive for investors and is very likely to lead to poor decisions, it may be producing for PhD candidates who can't find a topic but even then it rarely ends in something successful let alone something that informs investing.

It would have been jaw dropping if the backtest showed a strong, durable effect for the month of September only. I would love to read a short fiction story about that happening and how bogleheads might respond.

Regarding the bolded part of what you wrote, it is entirely possible for investing theories themselves to be poorly founded. I suppose it's less likely for a baseless theory to correspond to a strong signal in a backtest compared to just fishing, but those stars could align...

You did say solid theory, so perhaps it is a moot point. But one could still go fishing with bad theories and may even catch something. I will bet this has happened at least once in history.

Well said.
Based on the backtesting, the theory doesn't stand-up. But to say don't backrest the theory, well that's just blindly following a methodology. Sorry Bogleheads ....

We don’t go test whether gravity still exists every day

[OKOKOK]

Touchy subject here.
Testing a theory with back testing should be encouraged.

Not really no. First, you haven't articulated a 'theory' of any sorts and fishing for correlations without a solid theory behind what you are looking for is a recipe for disaster. Second, backtesting alone doesn't really offer you assurance of the validity of what you tested particularly when the variance will dwarf the signal. Third, encouraging random goose chasing isn't productive for investors and is very likely to lead to poor decisions, it may be producing for PhD candidates who can't find a topic but even then it rarely ends in something successful let alone something that informs investing.

It would have been jaw dropping if the backtest showed a strong, durable effect for the month of September only. I would love to read a short fiction story about that happening and how bogleheads might respond.

Regarding the bolded part of what you wrote, it is entirely possible for investing theories themselves to be poorly founded. I suppose it's less likely for a baseless theory to correspond to a strong signal in a backtest compared to just fishing, but those stars could align...

You did say solid theory, so perhaps it is a moot point. But one could still go fishing with bad theories and may even catch something. I will bet this has happened at least once in history.

Well said.
Based on the backtesting, the theory doesn't stand-up. But to say don't backrest the theory, well that's just blindly following a methodology. Sorry Bogleheads ....

We don’t go test whether gravity still exists every day

No one said to do this daily.

[retireIn2020]

Just looking at past history of September for the last 4 years

Has the Fed cut interest rates in September of any of those years?

If I were a betting man, I'd bet that in September 2024 equities will have a positive return due to the fact that the Fed has already signaled that there will be a rate cut in September.

Since I'm not a betting man, I'll just stay the course.

[blortchplop]

Touchy subject here.
Testing a theory with back testing should be encouraged.

Not really no. First, you haven't articulated a 'theory' of any sorts and fishing for correlations without a solid theory behind what you are looking for is a recipe for disaster. Second, backtesting alone doesn't really offer you assurance of the validity of what you tested particularly when the variance will dwarf the signal. Third, encouraging random goose chasing isn't productive for investors and is very likely to lead to poor decisions, it may be producing for PhD candidates who can't find a topic but even then it rarely ends in something successful let alone something that informs investing.

It would have been jaw dropping if the backtest showed a strong, durable effect for the month of September only. I would love to read a short fiction story about that happening and how bogleheads might respond.

Regarding the bolded part of what you wrote, it is entirely possible for investing theories themselves to be poorly founded. I suppose it's less likely for a baseless theory to correspond to a strong signal in a backtest compared to just fishing, but those stars could align...

You did say solid theory, so perhaps it is a moot point. But one could still go fishing with bad theories and may even catch something. I will bet this has happened at least once in history.

Well said.
Based on the backtesting, the theory doesn't stand-up. But to say don't backrest the theory, well that's just blindly following a methodology. Sorry Bogleheads ....

Just to be clear, I think that looking for various kinds of correlation after forming a theory is still a little dangerous if the theory itself is not valid or rigorous, since the backtest data might match the theory on chance. Acting on the result would not be wise. I'm agreeing with avalpert1.

Imagine September actually showed statistically significant lower returns. I'd still want a theory about it that's based on some kind of reason. Maybe a historical correlation such as:

Just looking at past history of September for the last 4 years

Has the Fed cut interest rates in September of any of those years?

Imagine the data showed statistically significant higher returns in September, maybe you could say it was back to school shopping, lol. But I'd be wary of relying on a backtest result without knowing why the result is that way.

I'm not against doing the backtest though! This one came out pure noise, which is what you'd expect. Any other result would have been astounding and would have been neat to look into more to figure out why.

[OKOKOK]

Just looking at past history of September for the last 4 years

Has the Fed cut interest rates in September of any of those years?

If I were a betting man, I'd bet that in September 2024 equities will have a positive return due to the fact that the Fed has already signaled that there will be a rate cut in September.

Since I'm not a betting man, I'll just stay the course.

Agreed. I was curious about historicals (hence backtesting).

[02nz]

Just looking at past history of September for the last 4 years

Has the Fed cut interest rates in September of any of those years?

If I were a betting man, I'd bet that in September 2024 equities will have a positive return due to the fact that the Fed has already signaled that there will be a rate cut in September.

Since I'm not a betting man, I'll just stay the course.

The fact that the Fed has already clearly signaled the coming rate cut in September means that's already priced in. Stock markets move in reaction to changes to what's already expected.

[avalpert1]

Just looking at past history of September for the last 4 years

Has the Fed cut interest rates in September of any of those years?

If I were a betting man, I'd bet that in September 2024 equities will have a positive return due to the fact that the Fed has already signaled that there will be a rate cut in September.

Since I'm not a betting man, I'll just stay the course.

The fact that the Fed has already clearly signaled the coming rate cut in September means that's already priced in. Stock markets move in reaction to changes to what's already expected.

Right now futures markets are fully pricing in a 25bp cut and putting the odds of a 50bp cut at ~33%, so if anything I think there is probably more downside potential on stocks from the Fed meeting than upside (not that I would try to time the market around that nor suggest that the Fed rate decision will be the biggest driver of equity prices in September).

[retireIn2020]

Just looking at past history of September for the last 4 years

Has the Fed cut interest rates in September of any of those years?

If I were a betting man, I'd bet that in September 2024 equities will have a positive return due to the fact that the Fed has already signaled that there will be a rate cut in September.

Since I'm not a betting man, I'll just stay the course.

The fact that the Fed has already clearly signaled the coming rate cut in September means that's already priced in. Stock markets move in reaction to changes to what's already expected.

Right now futures markets are fully pricing in a 25bp cut and putting the odds of a 50bp cut at ~33%, so if anything I think there is probably more downside potential on stocks from the Fed meeting than upside (not that I would try to time the market around that nor suggest that the Fed rate decision will be the biggest driver of equity prices in September).

In September 2019 the fed cut the rate by 25bp and the S&P 500 did show a positive return for the month.
Just Sayin .

[retireIn2020]

Just looking at past history of September for the last 4 years

Has the Fed cut interest rates in September of any of those years?

If I were a betting man, I'd bet that in September 2024 equities will have a positive return due to the fact that the Fed has already signaled that there will be a rate cut in September.

Since I'm not a betting man, I'll just stay the course.

The fact that the Fed has already clearly signaled the coming rate cut in September means that's already priced in. Stock markets move in reaction to changes to what's already expected.

Right now futures markets are fully pricing in a 25bp cut and putting the odds of a 50bp cut at ~33%, so if anything I think there is probably more downside potential on stocks from the Fed meeting than upside (not that I would try to time the market around that nor suggest that the Fed rate decision will be the biggest driver of equity prices in September).

In September 2019 the fed cut the rate by 25bp and the S&P 500 did show a positive return for the month.
Just Sayin .

September is in the books, S&P 500 up almost 2%.

[retireIn2020]

Has the Fed cut interest rates in September of any of those years?

If I were a betting man, I'd bet that in September 2024 equities will have a positive return due to the fact that the Fed has already signaled that there will be a rate cut in September.

Since I'm not a betting man, I'll just stay the course.

The fact that the Fed has already clearly signaled the coming rate cut in September means that's already priced in. Stock markets move in reaction to changes to what's already expected.

Right now futures markets are fully pricing in a 25bp cut and putting the odds of a 50bp cut at ~33%, so if anything I think there is probably more downside potential on stocks from the Fed meeting than upside (not that I would try to time the market around that nor suggest that the Fed rate decision will be the biggest driver of equity prices in September).

In September 2019 the fed cut the rate by 25bp and the S&P 500 did show a positive return for the month.
Just Sayin .

September is in the books, S&P 500 up almost 2%.

Actually, the S&P 500 was up a full 2% in September. I didn't realize my chart didn't include Monday Sept 30.

[rkhusky]

Actually, the S&P 500 was up a full 2% in September. I didn't realize my chart didn't include Monday Oct 30.

Close to the average for the last 40 years and the last 20 years. And I think you meant Sept 30.

[Tamalak]

Sept is also a month hampered by quarterly dividend distributions which typically knock down the price. This Sept held up very well

[retireIn2020]

Actually, the S&P 500 was up a full 2% in September. I didn't realize my chart didn't include Monday Oct 30.

Close to the average for the last 40 years and the last 20 years. And I think you meant Sept 30.

Yes, I meant Sept 30, my main point is that you can't go by Hyperlative scenarios. Sorry, had to use a slang word for meaning.
https://www.urbandictionary.com/define. ... yperlative

Not to insult anyone's intelligence, but basically trying to identify the fact that some often believe in these fallacies. What's next? "Sell in May"?