Beat the Market with Stochastic Momentum Index

 

Beat the Market with Stochastic Momentum Index


Why is it important to beat S&P500 (the Market)? 

"Beating the Market" means to earn an investment return that exceeds the performance of the Standard & Poor's 500 index which measures the stock performance of 500 large American companies. What is important to understand here is the fact that S&P500 is actually a managed portfolio that frequently add newer, better companies and discard laggards. The rebalancing of S&P500 is done four times in a year by a committee and the included companies are weighted by market capitalization. The total current market cap of S&P500 is about $30 trillion. Many institutions (banks, insurance companies, pension funds, credit unions, hedge funds etc. called together institutional investors) invest primarily in stocks that are in the S&P500 and try to overweight or underweight their positions in certain sectors. This practice ensures they’ll never do much better or much worse than the S&P 500.    

The following computer tests were performed for the time period between 01/01/1985 and 30/11/2020. The price/volume data was read from Yahoo Finance website. Sixteen major declines of about 10% or more of the Market (=S&P500) were detected during that period of time. The purpose was to beat the Market, notoriously difficult task, holding the same stocks as S&P500 ( not different stocks, which is entirely different task!) but attempting to time the market using Stochastic Momentum Index of William Blau. All computer tests were done on weekly charts to smooth out the noise the Market is experiencing on daily basis.


Test Setup

We assume that the value of the S&P500 index is a fictitious price for which you can buy it. On 01.01.1985 you can buy 6.1095 shares of the index for the total price 1000$. The question is simple:

If you held these shares for 35 years, could SMI‑based timing rules beat the S&P 500’s buy‑and‑hold return?

All tests use:


Weekly data (to reduce noise)


Price/volume from Yahoo Finance


Sixteen major S&P 500 declines (>10%) identified between 1985–2020


Identical buy/sell dates across all tests


No leverage, no shorting, no dividends (pure price return


Test1 

sell rule:

we sell our shares at date i if:

Price(i) < SMA20

SMI(i,10,3,3) < 42.55 and SMI(i-2,10,3,3)>74.69  - the fast stochastic line has declined from its average value 74.69 at the sixteen price peaks to its average value 42.55 two weeks later. 

​

buy rule: we buy again shares of S&P500 if: 

SMI(i,9,3,3)>SMI(i-1,9,3,3)>SMI(i-2,9,3,3)

​

Result: return =  5.4467/6.1095= 89% of S&P500 return

​

Oops! Its not easy to beat the Market!

Lets change our selling rules and require that the current price should be below 26 periods (=weeks) simple moving average SMA.


Test 1A — Longer Moving Average

sell rule:

Price(i) < SMA26

SMI(i,10,3,3) < 42.55 and SMI(i-2,10,3,3)>74.69

​

buy rule: SMI(i,9,3,3)>SMI(i-1,9,3,3)>SMI(i-2,9,3,3)

​

Result: 

  You have bought shares  = 6,109482177213925 for 1000 dolars at 1984-12-31

  You should sell at 1988-11-07 sell price = 267,920013

  You should buy at  2018-11-12 buy price = 2736,27002

   ........
  Now you have shares = 5,841293072842256 but your cash is = 0

  Now you have shares = 5,841293072842256
  Your total cash is = 0
  The current value of SP500 index is 3638,350098
  Time elapsed: 00:19:10.0342743

​

  (that's correct - this test took me 19 minutes and 10 seconds on my new Lenovo Yoga Slim 7 laptop!)

​

 Return: 5.8413/6.1095 = 95.6% of S&P500 return - still not good enough.  


Test1B

sell rule: 

Price(i) < SMA26

SMI(i-2,10,3,3) - SMI(i,10,3,3)>74.69-42.55=32.14 - the two week drop of the fast stochastic line K should be greater than the difference between the average values

SMI(i-2,10,3,3)>55 - the Stochastic Momentum Index must have been in overbought position

​

buy rule - the same as test1A

​

Results: 

You have bought shares  = 6,109482177213925 for 1000 dolars at 1984-12-31

You should sell at 1988-11-07 sell price = 267,920013
Now you have shares = 0 but your cash is = 1636,852544342423

You should buy at  1988-12-12 buy price = 276,290009
Now you have shares = 5,924400054373385 but your cash is = 0

....
Now you have shares = 6,128554462417772
Your total cash is = 0
The current value of SP500 index is 3638,350098
Time elapsed: 00:20:35.0398696

​

​

Return=6.1286/6.1095=100.3% of S&P500 return  Voila ! We beat the Market! Hmmm - by very small amount. Let's smooth the K-line more. 

​

Test 2A

sell rule: 

Price(i) < SMA26

SMI(i,12,4,4) < 55.09  and  SMI(i-2,12,4,4)>70.38 - the fast stochastic K-line now with settings (12,4,4) has declined from its average value 70.38 at the sixteen price peaks to its average value 55.09 two weeks later. 

​

buy rule: SMI(i,9,3,3)>SMI(i-1,9,3,3)>SMI(i-2,9,3,3)

​

Results: 

You have bought shares  = 6,109482177213925 for 1000 dolars at 1984-12-31

You should sell at 2007-02-26 sell price = 1387,170044

......

You should buy at  2020-03-30 buy price = 2488,649902
Now you have shares = 7,160320087635551 but your cash is = 0

Now you have shares = 7,160320087635551
Your total cash is = 0
The current value of SP500 index is 3638,350098
Time elapsed: 00:19:00.7235007

​

Return = 7.1603/6.1095 = 117.2% of S&P500 return. Here we are - we beat the market by a substantial amount! 

​

Test2C 

sell rule: if ClosePrice[i] < SMA27[i] &&

StocMomIndex(12, 4, 4)[i - 2]- StocMomIndex(12, 4, 4)[i]> 20.0

&& StocMomIndex(10, 3, 3)[i - 2]>70.38

this test was a mistake - it should have been SMI(12,4,4)>70.38

​

buy rule: SMI(i,9,3,3)>SMI(i-1,9,3,3)>SMI(i-2,9,3,3)

Result: 

You have bought shares  = 6.109482177213925 for 1000 dolars at 1984-12-31

You should sell at 2000-01-24 sell price = 1360.160034
Now you have shares = 0 but your cash is = 8309.873485881686

You should buy at  2000-03-06 buy price = 1395.069946
Now you have shares = 5.956599889280165 but your cash is = 0

You should sell at 2007-02-26 sell price = 1387.170044
Now you have shares = 0 but your cash is = 8262.81693050316

You should buy at  2007-03-26 buy price = 1420.859985
Now you have shares = 5.815363243200322 but your cash is = 0

You should sell at 2010-05-03 sell price = 1110.880005
Now you have shares = 0 but your cash is = 6460.170748683189

You should buy at  2010-07-12 buy price = 1064.880005
Now you have shares = 6.066571555809417 but your cash is = 0

You should sell at 2018-10-08 sell price = 2767.129883
Now you have shares = 0 but your cash is = 16786.99143943804

You should buy at  2018-11-12 buy price = 2736.27002
Now you have shares = 6.134990814772746 but your cash is = 0

You should sell at 2020-02-24 sell price = 2954.219971
Now you have shares = 0 but your cash is = 18124.11238690321

You should buy at  2020-03-30 buy price = 2488.649902
Now you have shares = 7.282708737913594 but your cash is = 0

Now you have shares = 7.282708737913594
Your total cash is = 0
The current value of SP500 index is 3638.350098
Time elapsed: 00:19:19.8304324

​

​

Although a mistake the result is: return= 7.2827/6.1095 = 119.2% of S&P500 return.

​

Test2D

Correction of the mistake in test2C: 

sell rule: if ClosePrice[i] < SMA27[i] &&

StocMomIndex(12, 4, 4)[i - 2]- StocMomIndex(12, 4, 4)[i]> 20.0

&& StocMomIndex(12, 4, 4)[i - 2]>70.38

​

buy rule: SMI(i,9,3,3)>SMI(i-1,9,3,3)>SMI(i-2,9,3,3)

result = the same as test2C

​

Test2E
                     

sell rule: if ClosePrice[i] < SMA27[i] &&

StocMomIndex(12, 4, 4)[i - 2]- StocMomIndex(12, 4, 4)[i]> 16.31+13.35=29.66 && StocMomIndex(12, 4, 4)[i - 2]>70.38

16.31 is the average 2-week drop of fast SMI line

13.35 is the standard deviation of that drop

​

buy rule: SMI(i,9,3,3)>SMI(i-1,9,3,3)>SMI(i-2,9,3,3)

Result: 

You have bought shares  = 6.109482177213925 for 1000 dolars at 1984-12-31

You should sell at 2018-10-08 sell price = 2767.129883
Now you have shares = 0 but your cash is = 16905.730702224555

You should buy at  2018-11-12 buy price = 2736.27002
Now you have shares = 6.178385385454231 but your cash is = 0

You should sell at 2020-02-24 sell price = 2954.219971
Now you have shares = 0 but your cash is = 18252.30949424342

You should buy at  2020-03-30 buy price = 2488.649902
Now you have shares = 7.3342214505845025 but your cash is = 0

Now you have shares = 7.3342214505845025
Your total cash is = 0
The current value of SP500 index is 3638.350098
Time elapsed: 00:19:21.8258729

​

​

Return = 7.3342/6.1095 = 120.05% of S&P500 return.

​

Test2F

sell rule: if ClosePrice[i] < SMA27[i] &&

StocMomIndex(12, 4, 4)[i - 2] - StocMomIndex(12, 4, 4)[i]> 16.31 

&& StocMomIndex(12, 4, 4)[i - 2]>70.38

​

buy rule: SMI(i,9,3,3)>SMI(i-1,9,3,3)>SMI(i-2,9,3,3)

result = the same as test2C

​

Test3

sell rule: if ClosePrice[i] < SMA27[i] && StocMomIndex(12, 4, 3)[i - 2]- StocMomIndex(12, 4, 3)[i]> 23.0
                        && StocMomIndex(12, 4, 3)[i - 2]>73.85

​

buy rule: SMI(i,9,3,3)>SMI(i-1,9,3,3)>SMI(i-2,9,3,3)

​

Result: 

You have bought shares  = 6.109482177213925 for 1000 dolars at 1984-12-31

You should sell at 2000-01-24 sell price = 1360.160034
Now you have shares = 0 but your cash is = 8309.873485881686

You should buy at  2000-03-06 buy price = 1395.069946
Now you have shares = 5.956599889280165 but your cash is = 0

You should sell at 2007-02-26 sell price = 1387.170044
Now you have shares = 0 but your cash is = 8262.81693050316

You should buy at  2007-03-26 buy price = 1420.859985
Now you have shares = 5.815363243200322 but your cash is = 0

You should sell at 2010-05-03 sell price = 1110.880005
Now you have shares = 0 but your cash is = 6460.170748683189

You should buy at  2010-07-12 buy price = 1064.880005
Now you have shares = 6.066571555809417 but your cash is = 0

You should sell at 2018-10-08 sell price = 2767.129883
Now you have shares = 0 but your cash is = 16786.99143943804

You should buy at  2018-11-12 buy price = 2736.27002
Now you have shares = 6.134990814772746 but your cash is = 0

You should sell at 2020-02-24 sell price = 2954.219971
Now you have shares = 0 but your cash is = 18124.11238690321

You should buy at  2020-03-30 buy price = 2488.649902
Now you have shares = 7.282708737913594 but your cash is = 0

Now you have shares = 7.282708737913594
Your total cash is = 0
The current value of SP500 index is 3638.350098
Time elapsed: 00:19:49.9682156

​

Return:7.2827/6.1095=119.2% of S&P500 return


Conclusion

Across all tests, one pattern becomes clear: the Stochastic Momentum Index can outperform the S&P 500, but only under specific conditions that capture genuine market exhaustion rather than ordinary pullbacks.

The early tests (1, 1A, 1B) show that simple SMI thresholds or basic moving‑average filters are not enough to beat the market. These rules either trigger too often or fail to distinguish between shallow corrections and true trend reversals. As a result, they underperform or only match buy‑and‑hold.

The breakthrough comes when the strategy focuses on the magnitude of the two‑week SMI decline, combined with:

  • a longer smoothing period (SMI 12‑4‑4),

  • a price filter (Price < SMA26/27), and

  • an overbought requirement (SMI > ~70 before the drop).

These conditions isolate the major turning points that preceded the 16 significant declines between 1985 and 2020. When the SMI drops sharply from an overbought level while price falls below a medium‑term moving average, the probability of a deeper correction increases. Avoiding these periods — and re‑entering only when SMI turns upward for three consecutive weeks — produces a measurable improvement.

The best-performing variations (Tests 2A, 2C, 2D, 2E, 3) deliver 117%–120% of the S&P 500’s return over 35 years. This is a substantial outperformance, especially considering:

  • no leverage

  • no alternative assets

  • identical underlying exposure as the index

  • only weekly data

  • only a handful of trades per decade

In other words:

The SMI does not beat the market by predicting tops — it beats the market by avoiding the worst declines.

The strategy works because it steps aside during the deepest drawdowns (2000, 2008, 2020), then re‑enters once momentum stabilizes. Over decades, avoiding even a few major crashes compounds into meaningful outperformance.

The final takeaway is simple:

Yes, it is possible to beat the S&P 500 using the Stochastic Momentum Index — but only with carefully tuned rules that capture true momentum reversals, not noise.  

The SMI is not a magic indicator, but when combined with trend filters and statistical thresholds, it becomes a powerful tool for long‑term market timing.                                                    

Comments

Post a Comment

Popular posts from this blog

Insider Trading, Institutional Ownership, and Trend‑Cycle Dynamics 2010–2025

 1. Introduction Understanding how insiders (corporate executives, directors, major shareholders) and institutional investors (mutual funds, hedge funds, pension funds) behave across market cycles is central to explaining the anatomy of long‑term price trends. This report synthesizes the most relevant academic research from 1986–2024, focusing on: • insider buying/selling patterns • institutional accumulation/distribution • herding and feedback trading • trend‑phase behavior (early, mid, late) • predictive power for future returns 2. Insider Trading Behavior Across Market Cycles 2.1 Foundational Research Seyhun (1986–2014) H. Nejat Seyhun is the leading authority on insider trading research. Key findings across his body of work: • Insiders are contrarian. • They buy aggressively after large declines. • They sell heavily after strong rallies. • Net insider buying predicts positive 12‑month returns. • Net insider selling predicts negative 6–12 month returns. 2.2 I...
Read more

Micron Technology (MU) — Investment Thesis (April 2026)

Image
Core Thesis: Micron is no longer a cyclical DRAM manufacturer. It has become a structural bottleneck supplier to the global AI compute stack, with multi‑year visibility, unprecedented pricing power, and a technology roadmap that positions it as a critical enabler of trillion‑parameter AI models . The market is still valuing Micron like a commodity memory vendor, creating a significant valuation disconnect. 1. Structural Demand Shift: The AI Memory Supercycle HBM as the New Compute Bottleneck AI training and inference workloads have shifted the bottleneck from GPU cores to memory bandwidth and capacity . High‑Bandwidth Memory (HBM) is now the most constrained component in the AI supply chain. Micron is one of only three global suppliers capable of producing HBM at scale, and the only US‑based one. Full HBM Allocation Through 2026 Micron has publicly confirmed that its entire HBM output is sold out through calendar 2026 , driven by hyperscaler and accelerator OEM demand. This provid...
Read more

Is There a Limit to Bitcoin's Price?

Image
                                          Why $500,000 Might Be the Ceiling Bitcoin has often been the subject of incredible price predictions. Some crypto enthusiasts argue it could reach $1 million or more per coin. But is that really realistic? In this post, we take a sober look at the fundamentals behind Bitcoin's value and explain why prices beyond $500,000 might be out of reach. 1. Market Cap Math: The Trillion-Dollar Threshold If Bitcoin were to hit $500,000 per coin, its total market capitalization (21 million coins) would be around $10.5 trillion . To put that into perspective: The entire gold market is valued at approximately $13 trillion . Global real estate is worth over $300 trillion . The U.S. stock market alone exceeds $40 trillion . For Bitcoin to reach $500,000, it would need to capture a large chunk of global wealth—not just from gold, but from e...
Read more