Dual Momentum Strategy with Niftybees & Goldbees

Algos, strategies, code
1

If you’ve been trading or investing for a while, you know how tricky it can be to balance risk and reward. Equities are great for growth but can hit you hard when there is a panic in the market. Gold, on the other hand, is steady but won’t get your portfolio racing. But what if you could get the best of both worlds by switching between them at the right time?

I recently tested a Dual Momentum Trading Strategy with Niftybees and Goldbees ETFs, using ChatGPT to backtest it over 14 years. The result? A total return of 1221.50% with a max drawdown of just -23.51%. Here’s the breakdown of how this strategy works and why it’s something every investor should consider.

Why Combine Equity and Gold?

Most of us know that equities and gold are pretty much on opposite ends of the risk spectrum. When the stock market tanks, gold usually steps up as the “safe haven” for investors. Equities like Niftybees can skyrocket but are prone to huge drawdowns.

In contrast, Goldbees tends to stay steady, giving you stability when equities are volatile.

So, why not combine the two and rotate between them based on momentum?

Here’s How the Strategy Works:

The rules are super simple, and that’s why I like it.

  1. Every quarter, check the last three months’ returns for Niftybees and Goldbees.
  2. If Niftybees is outperforming, put your money in Niftybees.
  3. If Goldbees has better returns, switch to Goldbees.
  4. Rebalance at the end of every quarter based on which asset is performing better.

This way, you’re always in the asset that’s showing the strongest momentum.

What Did the Backtest Show?

I used ChatGPT to backtest this strategy using EOD data for Niftybees and Goldbees over the last 14 years. Here’s what the data showed:

  • Dual Momentum Strategy (Niftybees & Goldbees Combined):

CAGR: 16.06%

Total Return: 1221.50%

Max Drawdown: -23.51%

  • Niftybees Buy-and-Hold:

CAGR: 11.93%

Total Return: 625.76%

Max Drawdown: -55.15%

  • Goldbees Buy-and-Hold:

CAGR: 11.43%

Total Return: 570.93%

Max Drawdown: -24.38%

The Momentum Strategy clearly outperformed buy-and-hold in terms of both return and risk. What really caught my eye was the much lower drawdown. While holding Niftybees through the 2008 crash would’ve given you a heart attack with a -55% drawdown, this momentum strategy only had a drawdown of around -8%.

Risk-Adjusted Returns: Sharpe Ratios

If you’re the type who likes digging into risk metrics, here are the Sharpe ratios for each strategy:

  • Momentum Strategy (Niftybees & Goldbees Combined): 0.95
  • Niftybees Buy-and-Hold: 0.41
  • Goldbees Buy-and-Hold: 0.47

The higher the Sharpe ratio, the better the return per unit of risk. And guess what? The momentum strategy blows the other two out of the water, almost doubling the Sharpe ratio of Niftybees buy-and-hold.

Here’s the year on year returns, momentum strategy has consistently made better returns. In last 14 year, only once momentum strategy has made negative returns. This clearly shows how we can stabilise the portfolio from extreme drawdown by diversifying with another asset class.

What Happens if We Exclude Goldbees?

I was curious about how the strategy would perform if I excluded Goldbees and just held cash when Niftybees was underperforming. So, instead of switching to gold, I tried sitting in cash (or liquid funds) when gold was doing better. Here’s what I found:

  • Total Return: 409.15%
  • Max Drawdown: -13.78%

While the drawdown was smaller, the total return didn’t even come close to the Dual Momentum Strategy that included both Niftybees and Goldbees. Clearly, gold plays a critical role in smoothing out those rough market periods.

And Here’s the Best Part—No Code Needed

All of this was done without writing a single line of code. I simply used ChatGPT, uploaded the data, and gave it a prompt to perform the backtest. Within seconds, it generated all the necessary results, showing how effective this simple strategy can be.

This proves that you don’t need complicated tools or coding skills to test and implement robust trading strategies. Sometimes, simplicity wins. If you’re interested in diving deeper into this concept, I highly recommend reading Gary Antonacci’s book on Dual Momentum. He explains the logic behind the strategy in detail.

Final Thoughts

Whether you’re an experienced trader or someone who prefers a more hands-off approach, this strategy offers the best of both worlds. You can protect your capital during downturns while still capturing strong upside momentum when the market is in your favor.

4 Likes
2

I have the following question

  1. Is this strategy for traders or for investors
  2. I am a investor, and own nifty 50 etf and started with gold etf .
  3. Your strategy states that every three months if nifty gives higher return for the past three months , put incremental money to nifty 50. If gold gives returns in excess of nifty50, then incremental money should go to gold etf

Is this the right thing or am i missing on anything.

I continue to park incremental money for 3 months, then check which product gives better return and continue this process - is it correct.

Since i am invested in these two, i will give it a try instead of equal allocation of incremental monies to nifty and gold.

The existing portfolio remain unchanged right - no selling. Its only incremental money. As if i sell i will be hit by taxes.

Interesting.

3

Mulitasset momentum has always attracted me, but could never get enough study material to understand its practical impact. Thanks for the post.

Just wanted to understand here … if we are following rule 2 & 3, what is rule 4 doing value add? as rule 3/4 suggest complete shift to single asset class. or have I misunderstood it completely

if we consider rule4 then it doesn’t become dual momentum but its strategic allocation portfolio. I am curious to know… in this case how it is different from say weekendinvestings free smallcase NG5050. WeekendInvesting<!-- --> | <!-- -->smallcases ?

NG5050 also suggests CAGR around 15.20% (it has limited history of 3 years) but switching costs are low, In your strategy what is the CAGR if we consider taxes while switching or have you already considered taxes in your CAGR of 16.06?

2 Likes
4

That’s an awesome analysis. Thanks for sharing here.

Did you upload daily EOD csv files or did you do something else?

What prompt(s) did you use?

5

:man_facepalming: Have you even verified the results? :thinking:

1 Like
6

Good Joke.

7

Have head of dual momentum, good to see that it works here too.
While this looks great, and i will try to use things like this in future for my pledged portfolio - there is an obvious problem.

New things can always happen in future. If we have 100% equity and 3 months timelag before shifting, then a severe quick crash could blow up well past DD.
Just the possibility of this, means i wont use this as it is. Past doesnt predict future that well and in this case we have a very small sample of crashes - so it will be unreliable.

But, can use this as part of portfolio with other ideas for diversification.

8

Why do you say it is a joke, you feel there is something wrong in it. Please explain as I plan to implement it.

@KirubaKaran - can you please clarify the queries raised please.

9

Is he asking for complete shift of existing portfolio from Nifty 50 to Gold every three months?

How is this possible - who will pay taxes.

10

@neha1101 - Please don’t.

Guys, it takes years to develop an algorithm, look for those anomalies, account for those unforeseen black swan events, work out the probabilities, have enough back testing but not curve fitting, have enough randomisation but not future prediction, include the cost of business, taxes, risk free returns, optimise - sharpe, sortino, drawdowns, and what not - To get 1% extra CAGR.

My point is - It is not simple as the strategy is making it look. I just ignore such posts and don’t get into the discussion but I understand that the sheer amount of data, graph, study from different angle shared in this post makes it compelling. So, to add value, I will be the devils advocate and tell exactly how to “train your eye” to look for the red flags.

P.S. Nothing against you @KirubaKaran , I am sure you also fell for it (ChatGPT) or were in a hurry. Hopefully this will help you and others - in the longer run.

How NOT to fall for easy strategies - My 2 cents -

  1. Don’t assume anything.

This is a common misconception. If you actually use excel CORREL function and take daily prices of GOLDBEES and NIFTYBEES for 17 years, you will find that there is a CORREL is of 0.91. That means they work in tandem more often than not. If they were opposite, CORREL would have been negative. A simple chart with both GOLD and NIFTY will show the same except for the small initial divergence in case of a major fall.

  1. Nobody has a crystal ball. Avoid switching with future forecasted returns.

Look at this data set. The returns from NIFTYBEES and GOLDBEES = 625.76% + 570.93% = 1196.69 % which is almost near the Dual Momentum Strategy returns of 1221.50%. This implies that there was a perfect switch of money before the fall as if your strategy had a crystal ball looking into future. This is in fact a very basic mistake that you or ChatGPT is making when it is taking decision before hand rather than after the quarterly returns. You should modify the code to ensure that after quarterly returns data, you make the decision of the switch / stay and you get actual returns based on next quarter returns of the security you are allocated to - for that quarter.

  1. Always account for cost of doing business and taxes.
    It is very common to see a strategy showing a lot of promise to give negative returns just by adding cost of doing business and adding taxes. By any calculation every switch will cost 0.5% of the portfolio. And this is not even considering 20% STCG or 12.5% LTCG.

The red flags are glaring. But just to be sure that I am right, I actually ran your algo and here are the results. This is considering 0.5% (Expense - STT, IGST, etc) switching cost but still NOT considering year end STCG/LTCG. And even then the total returns after 17 years are meagre 369.98% which is worse than GOLDBEES (570.93%) or NIFTYBEES (625.76%) even individually.

            Goldbees_Returns  Niftybees_Returns Current_Holding_In Next_Holding_In  Switch  Portfolio_Value  PortVal_Aft_Exp
Date                                                                                                                          
2007-03-31          0.000000           0.000000               GOLD            GOLD               100.000000         100.000000
2007-06-30         -8.236536          12.998652               GOLD           NIFTY  SWITCH        91.763464          91.304646
2007-09-30          9.321059          16.280712              NIFTY           NIFTY               106.169693         106.169693
2007-12-31         11.789474          22.249993              NIFTY           NIFTY               129.792442         129.792442
2008-03-31         14.595104         -22.873294              NIFTY            GOLD  SWITCH       100.104636          99.604112
2008-06-30          7.806081         -14.657303               GOLD            GOLD               107.379290         107.379290
2008-09-30         -1.905488          -2.953806               GOLD            GOLD               105.333190         105.333190
2008-12-31          2.719503         -24.534581               GOLD            GOLD               108.197729         108.197729
2009-03-31         12.859304           2.088438               GOLD            GOLD               122.111204         122.111204
2009-06-30         -3.217158          42.044721               GOLD           NIFTY  SWITCH       118.182694         117.591780
2009-09-30          7.202216          18.476614              NIFTY           NIFTY               139.318760         139.318760
2009-12-31          6.718346           2.303327              NIFTY            GOLD  SWITCH       142.527727         141.815088
2010-03-31         -2.663438           0.923852               GOLD           NIFTY  SWITCH       138.037931         137.347741
2010-06-30         15.236318           1.207826              NIFTY            GOLD  SWITCH       139.006663         138.311630
2010-09-30          1.511063          13.504941               GOLD           NIFTY  SWITCH       140.401606         139.699598
2010-12-31          6.858054           1.733845              NIFTY            GOLD  SWITCH       142.121773         141.411164
2011-03-31          0.696517          -4.902600               GOLD            GOLD               142.396117         142.396117
2011-06-30          5.434783          -3.194343               GOLD            GOLD               150.135037         150.135037
2011-09-30         18.134958         -12.468570               GOLD            GOLD               177.361962         177.361962
2011-12-31          3.530345          -6.452233               GOLD            GOLD               183.623451         183.623451
2012-03-31          3.486590          14.515710               GOLD           NIFTY  SWITCH       190.025648         189.075520
2012-06-30          3.998519          -0.314415              NIFTY            GOLD  SWITCH       188.481039         187.538633
2012-09-30          6.585974           8.039554               GOLD           NIFTY  SWITCH       199.889878         198.890429
2012-12-31         -3.039412           3.538302              NIFTY           NIFTY               205.927774         205.927774
2013-03-31         -3.306924          -3.768776              NIFTY            GOLD  SWITCH       198.166817         197.175983
2013-06-30        -15.069469           2.809478               GOLD           NIFTY  SWITCH       167.462609         166.625296
2013-09-30         19.211409          -1.829790              NIFTY            GOLD  SWITCH       163.576403         162.758521
2013-12-31         -4.363125           9.915785               GOLD           NIFTY  SWITCH       155.657164         154.878878
2014-03-31         -0.404709           6.348350              NIFTY           NIFTY               164.711131         164.711131
2014-06-30         -3.694126          13.531070              NIFTY           NIFTY               186.998310         186.998310
2014-09-30         -5.255082           4.643723              NIFTY           NIFTY               195.681993         195.681993
2014-12-31         -0.242915           3.991312              NIFTY           NIFTY               203.492272         203.492272
2015-03-31         -2.637987           2.514880              NIFTY           NIFTY               208.609859         208.609859
2015-06-30          0.875365          -1.442704              NIFTY            GOLD  SWITCH       205.600236         204.572235
2015-09-30         -1.363636          -5.014041               GOLD            GOLD               201.782614         201.782614
2015-12-31         -4.524508          -0.032080               GOLD           NIFTY  SWITCH       192.652944         191.689679
2016-03-31         12.505485          -2.616925              NIFTY            GOLD  SWITCH       186.673304         185.739938
2016-06-30          8.307332           7.099013               GOLD            GOLD               201.169971         201.169971
2016-09-30          1.728484           3.902145               GOLD           NIFTY  SWITCH       204.647162         203.623926
2016-12-31         -9.203540          -4.939526              NIFTY           NIFTY               193.565869         193.565869
2017-03-31          1.871345          12.069071              NIFTY           NIFTY               216.927471         216.927471
2017-06-30         -1.377727           3.784167              NIFTY           NIFTY               225.136369         225.136369
2017-09-30          3.453628           2.811709              NIFTY            GOLD  SWITCH       231.466548         230.309215
2017-12-31         -1.462866           7.581268               GOLD           NIFTY  SWITCH       226.940101         225.805400
2018-03-31          4.529882          -3.959851              NIFTY            GOLD  SWITCH       216.863844         215.779524
2018-06-30         -1.201748           5.938479               GOLD           NIFTY  SWITCH       213.186398         212.120466
2018-09-30         -0.368596           2.017397              NIFTY           NIFTY               216.399779         216.399779
2018-12-31          3.662597          -0.621200              NIFTY            GOLD  SWITCH       215.055503         213.980225
2019-03-31          0.214133           7.008944               GOLD           NIFTY  SWITCH       214.438427         213.366235
2019-06-30          6.766382           1.419059              NIFTY            GOLD  SWITCH       216.394027         215.312057
2019-09-30          9.973316          -2.666927               GOLD            GOLD               236.785808         236.785808
2019-12-31          4.488929           6.048220               GOLD           NIFTY  SWITCH       247.414956         246.177881
2020-03-31         10.914369         -29.343918              NIFTY            GOLD  SWITCH       173.939645         173.069947
2020-06-30         11.515310          19.823210               GOLD           NIFTY  SWITCH       192.999488         192.034491
2020-09-30          3.684581           9.177255              NIFTY           NIFTY               209.657986         209.657986
2020-12-31         -0.973291          24.309294              NIFTY           NIFTY               260.624362         260.624362
2021-03-31        -12.754286           5.070538              NIFTY           NIFTY               273.839420         273.839420
2021-06-30          6.156668           7.016684              NIFTY           NIFTY               293.053867         293.053867
2021-09-30         -1.974334          12.064052              NIFTY           NIFTY               328.408039         328.408039
2021-12-31          4.431017          -1.499022              NIFTY            GOLD  SWITCH       323.485130         321.867704
2022-03-31          6.364513           0.637891               GOLD            GOLD               342.353016         342.353016
2022-06-30         -0.815956          -9.645142               GOLD            GOLD               339.559564         339.559564
2022-09-30         -1.371115           8.327498               GOLD           NIFTY  SWITCH       334.903811         333.229292
2022-12-31          8.410565           5.913942              NIFTY            GOLD  SWITCH       352.936281         351.171599
2023-03-31          9.147254          -4.117855               GOLD            GOLD               383.294156         383.294156
2023-06-30         -3.446250          10.537594               GOLD           NIFTY  SWITCH       370.084880         368.234456
2023-09-30         -0.040560           2.341179              NIFTY           NIFTY               376.855484         376.855484
2023-12-31          8.865896          10.658255              NIFTY           NIFTY               417.021700         417.021700
2024-03-31          5.497577           2.740274              NIFTY            GOLD  SWITCH       428.449239         426.306993
2024-06-30          7.101219           7.541128               GOLD           NIFTY  SWITCH       456.579986         454.297086
2024-09-30          4.816098           7.497730              NIFTY           NIFTY               488.359056         488.359056
2024-12-31         -0.015736          -3.280016              NIFTY            GOLD  SWITCH       472.340801         469.979097

--- Portfolio Summary ---
Final Portfolio Value: 469.98
Overall Return: 369.98%
CAGR: 9.11%
Max Drawdown: -29.70%
Sharpe Ratio: 0.60

Hope this helps.

13 Likes
11

Shift happens only when condition changes i think.

But yes, taxes will have to be considered too ( and execution costs as rightly said above) but if risk gets reduced so much that it seems like a no brainer.

Look up on dual momentum, there should be plenty of stuff and there is a book that you should probably read before committing money ( i haven’t read but have heard of this concept in a podcast). And perhaps test this yourself too and look at variations.

Only problem with above is that future doesn’t have to mirror the past, things can shift. And obvious ( but unlikely) way where this might fail in terms of risk is a crazy fast dd within 3 months. So, this probably should be part of a larger portfolio.

I am too lazy too test this myself, hopefully he comes back with his data.

Dunno about investing universe, i havent tested anything like this yet. Will work on this in next few years after working on positional.

But intraday, for people with edges, its not abnormal to beat the market by a large margin - both risk and reward.

1 Like
12

Can you please clarify the following, time permitting @abhiwin123 @SpacemanSpiff . Going by the following concept with incremental money to be distributed between Nifty 50 ETF and Gold ETF.

The logic is.

  1. Every quarter, check the last three months’ returns for Niftybees and Goldbees.
  2. If Niftybees is outperforming, put your money in Niftybees.
    I repeat I have both these ETF. So instead of equal allocation of whatever little savings I intend to invest in these two IN FUTURE which is equal till last month, if I follow the above logic, I may get a incremental increase in return (althought this is not a priority)

I told this concept to a collegue as if this was my idea, she listened to the concept and said the entire opposite.

She says,

  1. Check Nifty 50 for last three months and if it is underperforming that Gold ETF, then put INCREMENTAL money into Nifty 50 for the next three months.

Now this is exactly the opposite, when asked for the logic, she says, when Nifty 50 underperforms its NAV will be lower and this is the time to invest more and not when it is overperforming.

Felt like socking her nose as my ego burst. But can you advise if this is right. If Nifty 50 underperforms on return basis for the last three months, does it goes without saying that its NAV must be lower and good to invest incremental money than when it overperforms?

Listening to her made me feel what she is saying is correct, but can the experts reconfirm.

I am only talking about INCREMENTAL money to be invested in the future with this concept.

13

Ego has to be left at the door. We need to be curious instead. Skeptical but also somewhat optimistic even after multiple failures.

She could be right, both could be wrong. You don’t know. So test it. This doesn’t seem too hard to test even manually.

I will never put money based on what anyone else says. I have made that mistake in past. Never again. Just borrow ideas or form your own and then test/modify.

Also, past data may not predict future. And sample sizes matter too. These are all things that a trader/active investor has to deal with.

That said, i have heard multiple times that gold / gilt 10Y / commodities can provide some amount of diversification against equity crashes. So easiest to do is just manage allocation.

I do not have answer on what you should do, because i have not tested this and @abhiwin123 above has given detailed reply with data and conclusion that this doesnt work. I have not verified that either. You gotta test rather than assume / defer to ‘experts’.

Also, i have respect for KirubaKaran, have gone through some of his videos for timepass and don’t consider his stuff bs ( even if say this one is wrong).

3 Likes
14

correct. This is strategic portfolio allocation and balancing. This reduces volatility for sure and (can) improve return depending on how uncorrelated N50 and Gold at that point.

1 Like
15

Your colleague is probably right. Any highs the market has made, and during downturns, adding up when the markets going down may be a good idea. But this wouldn’t be applicable to a single stock as the company giving the stock may be bad.

Studies shown that it will take time to reach back highs but it might be just a small time or may take years. This is shown in one of the books from Peter Lynch. An example of this is to look at SSE china index: 2007 it was at 5777 and tried reaching at 4651 in 2015 and 3210 this year. So even after 17 years, the market couldn’t reach back the all time highs made in 2007.

But in general momentum investing (aka, punting) is catching the Eel without getting the shock (downturn). If one is a risky investor, they would probably try the momentum investing. This term is super popular in the current bull market, but once market hits the downturn one should be able to pull back their gains. This is pretty much a short term strategy to me than a long term.

There are few small cases run with momentum investing. The folks managing those small cases charge exorbitant amounts. This means that when we invest in these, STCGs gets added and as a result we may get a 2% or so alpha. So Lets’ assume if we invest in 1L in this, making a 30% gain results in 6000 taxes, 9000+ (yes small case yearly fee) fee and results in 15k final profit deposited in bank. This becomes quiet pointless to take so much risk and get actual returns lower than nifty last year. For this to actually make sense, one’s initial capital has to be higher and has higher capability of risk taking. Atleast to me this is not my cup of coffee. :smiley: