Historical (1995-2020) trend analysis of S&P 500 Peaks & Valleys
Spent some time this weekend looking at all of the historical data from the S&P 500 back to 1995. I was looking for some basic trends that could help potentially predict what the current market is going to do. My general trend analysis shows that the market likes to move up with a long up-trend followed by short, steep down-trend. The pattern is very consistent across the entire 25 years at both the micro and macro scale. I estimated the trends to pick out a ratio of uptrend, peak, down trend, bottom. I then took a look at how each of these trends compared to the one following it (rebound). I did the analysis in excel (pasted below, sorry for the formatting).
What does all of this say? Well, the mean ratio of days up vs days down is 30%. This is a very consistent value with a standard deviation of 8%. The important of the ratio is that it does not matter how long the trend lasts. What matters is that once you find a peak, then a bottom is expected to come 30% * Days_Up. The drop for COVID-19 was not a "market" driven crash, so the peak and drop do not fit the model. However, the subsequent trend we are in right now is! That means once we can establish the rebound peak (next few days or weeks), then we can predict with some level of certainty when the "correction" or next bottom will occur. For example, if the peak happens tomorrow, then the rebound low will be sometime between 5/22 and 5/30.
In addition, we can also take a look at the peak to bottom % change. That value is much harder to nail down. If varies wildly over each of the trends for the past 25 years. The current drop was 3386 to 2241 or 34%. All of the drops after 2007 have only been a max of 20%. In all of those cases the rebound peak was higher. However, the rebound low following that rebound peak was lower 80% of the time!!! The only drops of the same scale are the dot com and housing market bubbles, but they were much more severe at 49% and 57%. Their rebound peaks where lower by around 23% and the rebound low was around 30% lower!
What does that mean? Well, excluding the 2 times where the rebound low was higher, then we will see a rebound low any where from -14% to -26% lower than the peak. If the peak happens on 5/12, then we could see the S&P 500 have a rebound peak around 2957 (about what it is now) and low between 2710-2518 around the end of May.
Well that is my $0.02 any way. It is probably a better guess than what you get from those so called experts that "have called 10 of the last 5 recessions" that I keep seeing in the news without any numbers to back up their predictions. My confidence is moderate on my analysis, but we all know the market can be very irrational. So, take my analysis with a block of salt :) Hope it helps in some useful way.
I AM SORRY FOR THE FORMATING. TradingView just mashed all of the numbers together without any useful formatting
Trading days not calendar days!
Bottom Cause Peak Date Days Up, Days Down, Down:Up Peak Bottom, Pk-Btm Pk : Btm Rebound Peak RePk - Pk RePk : Pk Rebound Bottom ReBtm - Pk ReBtm : Pk
1 .com Bubble 1/3/2001 1404 532 38% 1520 779 741 49% 1160 -360 -24% Lower 1061 -459 -30% Lower
2 Housing Market 10/15/2007 1260 342 27% 1567 681 886 57% 1216 -351 -22% Lower 1016 -551 -35% Lower
3 4/20/2010 272 62 23% 1216 1016 200 16% 1366 150 12% Higher 1099 -117 -10% Lower
4 6/7/2011 230 88 38% 1366 1099 267 20% 1418 52 4% Higher 1278 -88 -6% Lower
5 5/29/2012 128 49 38% 1418 1278 140 10% 1468 50 4% Higher 1350 -68 -5% Lower
6 10/10/2012 92 22 24% 1468 1350 118 8% 2135 667 45% Higher 1830 362 25% Higher
7 6/4/2015 656 152 23% 2135 1830 305 14% 2193 58 3% Higher 2085 -50 -2% Lower
8 8/23/2016 134 52 39% 2193 2085 108 5% 2871 678 31% Higher 2583 390 18% Higher
9 1/31/2018 114 41 36% 2871 2583 288 10% 2929 58 2% Higher 2351 -520 -18% Lower
10 China Trade War 9/28/2018 125 37 30% 2929 2351 578 20% 2941 12 0% Higher 2745 -184 -6% Lower
11 China Trade War 9/28/2018 477 61 13% 2929 2351 578 20% 3390 461 16% Higher 2241 -688 -23% Lower
12 4/30/2019 84 23 27% 2941 2745 196 7% 3026 85 3% Higher 2838 -103 -4% Lower
AVG 30% 20% 6% -8%
STD 8% 16% 19% 18%
-14% Mean ignore higher
12% Std ignore higher
Number of rebound peak/bottom pairs that where higher/lower with respect to previous peak
Total 12 Total 12
Higher 10 83% Higher 2 17%
Lower 2 17% Lower 10 83%
Most likely value at bottom following peak (COVID-19)
3386 2241 1145 34% Mean+Std 4246 860 25% Higher Mean+Std 3704 318 9%
3386 2241 1145 34% Mean 3594 208 6% Higher Mean 3111 -275 -8%
Most realistic -> Hit 2957 2X and it is 62% Fib. Level 3386 2241 1145 34% Mean-Std 2942 -444 -13% Lower Mean-Std 2518 -868 -26% Lower
3386 2241 1145 Mean 2912 -474 -14%
Next most likely -> 2730 is major support/resistance line 3386 2241 1145 Split diff 2710
3386 2241 1145 Mean-Std 2508 -878 -26%
Most likely date of bottom following peak (COVID-19)
Business days Calendar weeks Calendar days Date
13.3 2.66 18.6 5/30/2020 Mean+Std Most realistic to occur at end month
5/12/2020 35 10.4 2.08 14.5 5/26/2020 Mean
7.5 1.50 10.5 5/22/2020 Mean-Std
What does all of this say? Well, the mean ratio of days up vs days down is 30%. This is a very consistent value with a standard deviation of 8%. The important of the ratio is that it does not matter how long the trend lasts. What matters is that once you find a peak, then a bottom is expected to come 30% * Days_Up. The drop for COVID-19 was not a "market" driven crash, so the peak and drop do not fit the model. However, the subsequent trend we are in right now is! That means once we can establish the rebound peak (next few days or weeks), then we can predict with some level of certainty when the "correction" or next bottom will occur. For example, if the peak happens tomorrow, then the rebound low will be sometime between 5/22 and 5/30.
In addition, we can also take a look at the peak to bottom % change. That value is much harder to nail down. If varies wildly over each of the trends for the past 25 years. The current drop was 3386 to 2241 or 34%. All of the drops after 2007 have only been a max of 20%. In all of those cases the rebound peak was higher. However, the rebound low following that rebound peak was lower 80% of the time!!! The only drops of the same scale are the dot com and housing market bubbles, but they were much more severe at 49% and 57%. Their rebound peaks where lower by around 23% and the rebound low was around 30% lower!
What does that mean? Well, excluding the 2 times where the rebound low was higher, then we will see a rebound low any where from -14% to -26% lower than the peak. If the peak happens on 5/12, then we could see the S&P 500 have a rebound peak around 2957 (about what it is now) and low between 2710-2518 around the end of May.
Well that is my $0.02 any way. It is probably a better guess than what you get from those so called experts that "have called 10 of the last 5 recessions" that I keep seeing in the news without any numbers to back up their predictions. My confidence is moderate on my analysis, but we all know the market can be very irrational. So, take my analysis with a block of salt :) Hope it helps in some useful way.
I AM SORRY FOR THE FORMATING. TradingView just mashed all of the numbers together without any useful formatting
Trading days not calendar days!
Bottom Cause Peak Date Days Up, Days Down, Down:Up Peak Bottom, Pk-Btm Pk : Btm Rebound Peak RePk - Pk RePk : Pk Rebound Bottom ReBtm - Pk ReBtm : Pk
1 .com Bubble 1/3/2001 1404 532 38% 1520 779 741 49% 1160 -360 -24% Lower 1061 -459 -30% Lower
2 Housing Market 10/15/2007 1260 342 27% 1567 681 886 57% 1216 -351 -22% Lower 1016 -551 -35% Lower
3 4/20/2010 272 62 23% 1216 1016 200 16% 1366 150 12% Higher 1099 -117 -10% Lower
4 6/7/2011 230 88 38% 1366 1099 267 20% 1418 52 4% Higher 1278 -88 -6% Lower
5 5/29/2012 128 49 38% 1418 1278 140 10% 1468 50 4% Higher 1350 -68 -5% Lower
6 10/10/2012 92 22 24% 1468 1350 118 8% 2135 667 45% Higher 1830 362 25% Higher
7 6/4/2015 656 152 23% 2135 1830 305 14% 2193 58 3% Higher 2085 -50 -2% Lower
8 8/23/2016 134 52 39% 2193 2085 108 5% 2871 678 31% Higher 2583 390 18% Higher
9 1/31/2018 114 41 36% 2871 2583 288 10% 2929 58 2% Higher 2351 -520 -18% Lower
10 China Trade War 9/28/2018 125 37 30% 2929 2351 578 20% 2941 12 0% Higher 2745 -184 -6% Lower
11 China Trade War 9/28/2018 477 61 13% 2929 2351 578 20% 3390 461 16% Higher 2241 -688 -23% Lower
12 4/30/2019 84 23 27% 2941 2745 196 7% 3026 85 3% Higher 2838 -103 -4% Lower
AVG 30% 20% 6% -8%
STD 8% 16% 19% 18%
-14% Mean ignore higher
12% Std ignore higher
Number of rebound peak/bottom pairs that where higher/lower with respect to previous peak
Total 12 Total 12
Higher 10 83% Higher 2 17%
Lower 2 17% Lower 10 83%
Most likely value at bottom following peak (COVID-19)
3386 2241 1145 34% Mean+Std 4246 860 25% Higher Mean+Std 3704 318 9%
3386 2241 1145 34% Mean 3594 208 6% Higher Mean 3111 -275 -8%
Most realistic -> Hit 2957 2X and it is 62% Fib. Level 3386 2241 1145 34% Mean-Std 2942 -444 -13% Lower Mean-Std 2518 -868 -26% Lower
3386 2241 1145 Mean 2912 -474 -14%
Next most likely -> 2730 is major support/resistance line 3386 2241 1145 Split diff 2710
3386 2241 1145 Mean-Std 2508 -878 -26%
Most likely date of bottom following peak (COVID-19)
Business days Calendar weeks Calendar days Date
13.3 2.66 18.6 5/30/2020 Mean+Std Most realistic to occur at end month
5/12/2020 35 10.4 2.08 14.5 5/26/2020 Mean
7.5 1.50 10.5 5/22/2020 Mean-Std
Disclaimer
The information and publications are not meant to be, and do not constitute, financial, investment, trading, or other types of advice or recommendations supplied or endorsed by TradingView. Read more in the Terms of Use.