Dollar cost averaging Bitcoin - can we do better?

[virginorange]

Can we time the Bitcoin cycle?

The case for Dollar Cost Averaging (DCA)
Most people discover Bitcoin during a bull market, invest money, get discouraged during the bear market and sell. As a result, they can lose in USD terms even though they have invested in an asset with a huge average trend growth. Investing the same amount in Bitcoin every month is an easy rule to follow. It saves you the mental load of thinking about when and how much to invest and prevents you from buying at the euphoric top and selling at the depressed bottom.

The case for countercyclical buying
However, if we could be reasonably sure that Bitcoin were near the top or near the bottom, we could invest accordingly. For example, we could stop DCA during the euphoria and double DCA during the bear market. Or we could sell some of our bitcoin close to the top and buy back close to the bottom. This would be a modified DCA strategy.

In this article, I attempt to develop a reasonable strategy for the timing of bitcoin cycles.

TL;DR: A smart and disciplined investor can outperform a simple dollar cost average strategy, by buying Bitcoin depending to its price cycle. If Bitcoin is cheap (40% below trend), I would take a full position. If Bitcoin's price is expensive (more than 750 days ahead of trend), I would not add to my position and possibly even trim my position. If Bitcoin is neither cheap nor expensive I would dollar cost average into Bitcoin.


Seperating Trend and cycle
My Bitcoin price history dates from 16.07.2010 (0,07€) to now 20.03.2024 (64.000€) and price increase of six orders of magnitude 10^6. There seems to be an uptrand of higher highs and higer lows.



There seems to be an upward trend with some movement around the trend. Which function could describe this trend? I perform a logarithmic transformation on both axes of my chart. The new y-axis is ln(Bitcoinprice_in_EUR) and the x-axis is the ln number of days since the Bitcoin genesis block ln(Date - 03 Jan 2009).



The in ln-space the Bitcoin price follows the trend:
y = 5,8323x - 39,296

This translates into (=black line):
Bitcoin_price_in_EUR = e^[ 5,8323 * ln(Date - 03.Jan.2009) - 39,296 ]

Making the graphs easier to follow
A Bitcoin price in ln(EUR) and a Bitcoin age in ln(days) is easy to draw, but difficult to understand. Therefore I added the Bitcoin price in EUR and the calendar year.



Interestingly, the Bitcoin price does not seem to fall more than 1 below the trend (=red line).
Bitcoin_price_in_EUR_low = e^[ 5,8323 * ln(Date - 03.Jan.2009) - 39,296 -1]



This chart is similar to the bitcoin rainbow charts produced by others*. We can now subtract the observed price from the trend to separate the cyclical component of the bitcoin price from the trend component.
*https://www.blockchaincenter.net/en/bitcoin-rainbow-chart/



A graph of -1 on the y-axis means the Bitcoinprice was 63% below trend.
If the trend was 10.000€, then the Price at -1 would be at 3679, since exp(-1) = 0.36 .


Can we use the cycle information to buy at a low price and sell at a high price?

A quick back-of-the-envelope chart shows us that there may be value in knowing whether bitcoin is currently cheap or expensive. If we buy bitcoin when it is expensive, we can expect a worse return than if we buy it when it is cheap.



Even in recent history, the rule remains that "the cheaper you buy your bitcoin, the better":



Can we find some trading rules when to buy Bitcoin and when to sell Bitcoin to outperform DCA?


Finding the bottoms of the cycle

We can see that all the bottoms (except the first one on 04 Feb 2011) are between 0.5 and 1.0 below the trend:



To see the exact bottoms, I have zoomed into this chart at 0.5 to 1.0 below trend. Especially between -0.7 and -1.0 there are many bottoms. Only one bottom is lower than -1.0.



We now have a pretty good understanding of where the bottoms are. However, this assumes that the shape of the function and the parameters are stable going forward. Later I will show that the parameters are quite stable.

When we are between 0.5 and 1.0 below the trend, it has always been a good time to invest in bitcoin, because bitcoin has managed to come back to the trend every time until today. Your return has been quite nice as you have received bitcoin trend growth + catching up to the trend line.


Overvaluation alone is not enough to find cycle tops

Finding the top of the cycle is a bit more complicated for three reasons.
1. Bitcoin tops fall into a wider range than bottoms, as can be seen in the green box.



While the bottoms range from -0.5 to -1, the tops range from 0.88 to 2.7.
If you think you bought the bottom at -0.5 and the price goes to -1, you lose 39%.
If you think you bought the top at +0.88 and the price goes to 2.7, you miss out on a return of 517%.

2. Bitcoin's tops are getting smaller and smaller, which means you might miss the next top.
For the bottoms we don't see a trend, the bottoms fall randomly between -0.5 and -1, so we could expect the next bottom to fall between -0.5 and -1 again.
For the tops, we can see a declining trend. This means that the next top could be even lower than 0.88. Maybe only 0.5.


3. Selling expensive bitcoin to buy cheap bitcoin could backfire in fiat terms

Selling expensive bitcoin to buy cheap bitcoin later seems like a good idea. However, waiting in cash or gold for bitcoin to become cheap again takes time. The longer you wait, the higher the trend will be. If Bitcoin does not become cheap soon enough, the trend growth will exceed the cycle contraction. So you sell expensive bitcoin at a low USD price only to buy cheap bitcoin at a much higher USD price later.



I think Bitcoin bubbles (How much new retail will join?) are less predictable than bitcoin bear markets (The stable core of bitcoiners keeps buying bitcoin). Bubbles are also more difficult to trade, because you are trading against the trend, whereas in bear markets you have the trend on your side. Selling in bubbles gives you a short window before the trend catches up. Buying in bear markets has the trend on your side.


Because one of the main risks of selling bitcoin during a bubble is running out of time, therefore I ask: "How much time do we have before the trend catches up with us?".

So instead of saying "the current bitcoin price is e^1 too expensive", we could say "The current bitcoin price will only be justified in 500 days.". Also, bitcoin bubbles are not dominated by growing bitcoin network effects, but by human behavior. So we work with simple days (e.g. 500 days) and not the growth of days relative to the bitcoin genesis block ln(Date - 03 Jan 2009).




During peaks, the price of bitcoin is 500-1200 ahead of its trend. This means that the trend would have to continue for 500-1200 days to catch up with the current price. I like this chart because the peaks fall into a narrower band and since the peaks don't show a trend, you have to worry less about the next peak being outside of this band.


How many days ahead are too many?

Since we have found our peak indicator "days ahead of price", we can now think about a cut-off point. If the current bitcoin price is 1000 or 1500 days ahead, we will probably be able to buy bitcoin cheaper if we wait a bit. However, if the current price of bitcoin is only 10 days ahead, we are unlikely to get a cheaper price until the trend catches up with the current price.



Green box: Since the bitcoin price has always returned to the trend, it was always good to buy bitcoin below the trend, as you got the increase in the trend price of bitcoin plus the cyclical price increase to return to the trend.
Red box: If bitcoin has become so expensive that the trend would have to continue for more than 700 days to justify the current price, then it was always better to wait to buy bitcoin until the current price is supported by the trend again.
Yellow box: When bitcoin was moderately expensive (0 - 700 days ahead of the trend), the picture is not clear. Sometimes waiting gave you cheaper prices in the future, sometimes not.

If you decided to wait until you could buy bitcoin without a premium to the trend, you would have been able to get cheaper bitcoin in more than half of the cases:



Does this mean that we should never buy expensive bitcoin and instead always wait until we can buy bitcoin at the trend price?
No!

First of all, when bitcoin is moderately expensive (e.g. trend + 10 days), in most cases you can successfully wait for the trend to return and buy bitcoin slightly cheaper (e.g. for 5% less EUR). However, sometimes you will have to wait a long time to buy bitcoin back at fair value, which will cost you much more in fiat terms (e.g. for 200% more EUR).

Second, if bitcoin is cyclically expensive, we could always wait for bitcoin to become even more expensive before selling.


Putting our strategy to the test.

We invest 1000€ each month. We can either DCA or try to buy low and sell high.

If bitcoin is too expensive, we sell our bitcoin for gold.
If bitcoin is too cheap, we sell our gold for bitcoin.
If bitcoin is neither cheap nor expensive, we DCA into bitcoin and maintain our gold position.

How much do we outperform a DCA strategy depending on our definition of cheap and expensive?

Backtest 1:
Cheap Bitcoin means Bitcoin is 18% (e^-0.2) to 59% (e^-0.9) below trend.
Expensive Bitcoin means that Bitcoin is 500 days ahead of the trend.



What do we learn?
1.) In most scenarios we can significantly outperform DCA by up to 11076%.
2.) If we are too greedy and wait for Bitcoin 0.9 or 59% below trend, we will miss the re-entry into Bitcoin and thus underperform.



3.) It is worse to be a little too greedy than a little too cautious. Waiting to buy bitcoin 18% below trend instead of 39% below trend will cost you 2x performance. Waiting to buy bitcoin 59% below trend instead of 55% below trend will cost you 12x performance.

Sorry for the typo, the chart title should be: Sell Gold, if Bitcoin is 39% below trend. Buy Gold if Bitcoin is 500 days ahead of trend. Otherwise keep your Gold position and DCA into Bitcoin.


-> A re-entry into bitcoin at 0.5% or 39% below the trend looks best to me.


Backtest 2:
Ceap Bitcoin means Bitcoin is 39% (e^-0.5) below trend.
Expensive Bitcoin means Bitcoin is between 200 and 1000 days above trend.



What do we learn?
1.) Again, we can significantly outperform DCA by up to 11076%.
2.) We have two peaks at 4834% outperformance and 11076% outperformance. In between we have lower outperformance. There seem to be several local maxima that we can only hit with the benefit of hindsight. As my default position is in bitcoin and not gold, I would only sell my bitcoin when it is 700-800 days expensive.



Backtest 3:
To determine whether bitcoin is cheap relative to our trend function, we must first estimate the function. Therefore, we could not have started measuring Bitcoin vs. Trend from day 1. Currently we are about 5600 days from the Genesis block.
Therefore, I assume that we can only start selling expensive bitcoin and buying cheap bitcoin 3000 days after the genesis block, as only then will we have a stable trend function. Will we still outperform from day 3001 to day 5550?

Cheap bitcoin means bitcoin is 18% (e^-0.2) to 59% (e^-0.9) below trend.
Expensive Bitcoin means that Bitcoin is 750 days ahead of the trend.





What do we learn?
We can confirm past results. In recent history, the strategy still offers significant outperformance. We should be cautious when trading bitcoin and not wait for too big a drawdown.

Selling late (750 days high) is also better than selling early (500 days high).




Crating a simple decision chart for day to day use

The analysis in this post helped us to understand the Bitcoin trend, the Bitcoin cycle and optimum entry and exit points within the cycle. For daily usage I will visualize the relevant information in an simple chart.







Checking the model for weaknesses

Checking the stability of the bitcoin trend line

Risk: We believe that bitcoin is cyclically cheap, but the bitcoin price trend just went down.
We need to measure the bitcoin price trend correctly in order to extract the cyclical price component correctly. Otherwise, we wouldn't know where we are in the cycle and would make inefficient decisions when buying and selling bitcoin. However, the early days of bitcoin's adoption dominate the data set, masking possible recent changes in the trend.

I would like to see a stable trend in the past. A stable trend in the past gives us more confidence in extrapolating that trend than an unstable trend in the past.




Stable: If we divide our data set into the first two major tops (16/10/2010 - 01/05/2016) and the second two major tops (02/05/2016 - 21/03/2024), we get a trend of almost 5.7 in both periods.
More or less stable: If we divide the data set into 3 sets from cycle bottom to cycle bottom, we get a slope of 6.7, 9.6, 6.7. If we divide the data set into 3 sets from cycle top to cycle top, we get a slope of 5.8, 3.5, 5.5. The trend looks reasonably stable at around 6.

One day in the future, the price of Bitcoin could change its behaviour and break our current model (when Bitcion is cheap or expensive). This model will need to be revised if the bitcoin price falls significantly more than 63% (e^-1 = -63%) below trend. This model will also need to be revised if the slope of our function changes significantly. This could be the case due to significant regulatory changes in the US or due to the introduction of Mayor. This could also happen if bitcoin adoption runs out of steam. In the long run, financial assets (cash, bonds, loans, gold, part of the value of real estate, most of the value of art) can't grow faster than the amount of goods and services. The bitcoin trend looks stable so far and we can reasonably expect this trend to continue for some years into the future. However, we should be aware that one day our model will break.

-> The model is stable so far

Introduction of M2 inflation for USD and EUR not necessary at present

We measure the value of bitcoin against fiat. However, money printing devalues fiat and could affect our results. Adjusting for inflation reduces our slope from 5.8 to 5.2, but so far has not had a major impact on the model's results. This would change if the pace of inflation or monetary expansion in the US or the eurozone were to increase significantly. A much stronger monetary expansion in the future would increase the slope of our trend function. This could lead to a premature switch from bitcoin to gold, which would then also rise due to inflation, making this mistake less costly.

-> I use a bitcoin price that is not adjusted for M2 because bitcoin price data is more recent than M2 data and to keep the model simple. However, one should keep an eye on the monetary expansion of the major central banks.

Taxation
Switching between gold and bitcoin has an additional tax burden on realised taxes. As a rule of thumb, it is more efficient to let your wealth compound and pay your legal taxes later than earlier. In my country there is no capital gains tax on gold or bitcoin after 12 months, but your tax situation may be different.

Privacy
Trading in and out of bitcoin leaves data with exchanges, banks, etc. You may value your privacy more than additional financial wealth and therefore refrain from trading.

Even if privacy and tax concerns limit your trading decisions, you can still choose to postpone your regular DCA close to a top. If you are planning to sell some of your bitcoin anyway, you could choose the top. If bitcoin is significantly undervalued, you can buy additional bitcoin.



How do I apply this knowledge to my investment decisions?

In general, I have two constraints on my bitcoin investments:
1. Liquidity constraint: I need a liquidity buffer for (i) expected negative future cash flows and (ii) unexpected liquidity needs, for which I assume 2 annual salaries. This liquidity buffer is invested in cash (for short term needs) and a minimum drawdown portfolio of equities/gold (for medium term needs). I make a cash flow projection for the next 8 years (salary plus other income minus food minus mortgage) and calculate my minimum excess liquidity during these 8 years. This excess liquidity is earmarked for buying bitcoin.
2. Single position risk constant: I'm willing to invest 2/3 of my current and expected future net worth in bitcoin. After 2/3, I can only buy bitcoin in the bear market if I sold bitcoin during the bull market.

I divide my excess liquidity into 3 equal buckets:
1. bucket will be invested when we reach e^-1 undervaluation.
2. bucket will be invested when we reach fair value.
3. bucket will be partially invested each month as long as bitcoin does not exceed 150 days of overvaluation.

Unfortunately the current bitcoin price is lightly above trend. Therefore I can only deploy bucket nr. 3. I should have invested earlier, but I'm still in the process of formulating my strategy.


I hope my post offered some value to some of you. I personally got some insights out of my calculations. However since I'm relatively new to the Bitcoin space, it is possible that my thoughts are not new for some members here. Therefore I would like to get some feedback, if I made some mistakes to how to improve my model.

Do you think, it is possible to outperform investing into Bitcoin taking the cycle into account instead of just dollar cost averaging into Bitcoin?
Did you find any mistakes in my thought process?

[Solosanz]

Long thread, but it's really informative.

The only suggestion I can give is you might need to resize your image a bit smaller, probably like "[img width=700" or less, because it's too big IMO.

Selling Bitcoin for gold or vice versa isn't easy because the seller won't accept Bitcoin or gold as currency, so you're need to sell your Bitcoin to fiat then buy gold, it would cost additional fees.

The thing that I learn, despite Bitcoin keep break new ATH for every four years, but actually the trend tops are getting smaller and smaller, probably in the next two halving events, the tops and bottom are close, which make Bitcoin become more stable...

[thecodebear]

Can we time the Bitcoin cycle?

The case for Dollar Cost Averaging (DCA)
Most people discover Bitcoin during a bull market, invest money, get discouraged during the bear market and sell. As a result, they can lose in USD terms even though they have invested in an asset with a huge average trend growth. Investing the same amount in Bitcoin every month is an easy rule to follow. It saves you the mental load of thinking about when and how much to invest and prevents you from buying at the euphoric top and selling at the depressed bottom.

The case for countercyclical buying
However, if we could be reasonably sure that Bitcoin were near the top or near the bottom, we could invest accordingly. For example, we could stop DCA during the euphoria and double DCA during the bear market. Or we could sell some of our bitcoin close to the top and buy back close to the bottom. This would be a simple DCA strategy.

Yeah since Bitcoin has a very well known 4 year market cycle and clear bull runs and bear markets in the four year cycle, it is easy to DCA extra during the lower part of the cycle and then save up fiat during the high part in order to DCA extra when the low part of the cycle comes back. That's my strategy from here on out. It's still DCA, but using the most basic knowledge about Bitcoin's market cycles to get a much improved DCA over a typical steady buying DCA strategy.

For example, I've been DCA'ing recently in the $60,000s and low $70,000s, but probably won't do any more from this point onward this bull run. I'll wait until prices get back under $80k next bear market to resume DCA around bottom of the next bear market, and then I'll continue doing it until price passes whatever the high ends up being this cycle. There really is no reason to DCA during the main part of the bull markets, you're just buying higher than you would be buying if you waited for the bear market. This better strategic DCA of course only works with Bitcoin (and things that follow Bitcoin), as opposed to the stock market where a consistent DCA is best because there is no way to guess when asset prices will be higher or lower than today. Bitcoin following an regular market cycle allows us to easily beat out the typical consistent DCA with a strategic DCA only buying on the lower half of the market cycle.

In terms of selling at the top, you COULD do that as well, but then you have the risk of missing the top by a lot and maybe after taxes it wouldn't even be worth it, I dunno thats for each person to decide for themselves. I personally would prefer the super low risk strategy of strategically only DCA during low part of the market cycle and save during the high part of the cycle, while holding everything long term.

[Ascended_Intellect]

Fantastic breakdown!

Your analytical approach may prove to be accurate but time will answer this.

Some may complain about the "complicated nature" of the breakdown but when the average person is consumed by short form entertainment it's a given.

Personal standpoint for this current cycle is that we are below the ATH by a good margin, my guess is that BTC will top out at 67-70K GBP.

My theory even though the market is unpredictable is that we cannot possibly have a downtrend in price back to base level due to the recent bear market we just had, I've seen assumptions that this bull run will last into 2025 but November 2024 seems like a more reasonable guess.

[348Judah]

If you think the use of DCA is what you can afford to go with, then the best thing is to embrace using such purchasing pattern for your bitcoin accumulation and you could easily get going well with your investment over a long term, this method has been seen as part of the most simplest means everyone can afford to use in making an investment, you have done well in bringing all these together OP, there is more we can also learn extensively using these same pattern from this link. https://dcabtc.com/

[Turbartuluk]

Very impressive, it is a pleasure to see your brain working.

Indeed it was occasionally hard to understand your process, due to the deepness of analysis. I guess it takes me about 90min to read and understand everything, but it was worth every second.

In the end everything was plausible to me and i was not able to find mistakes in your thought process.


Just one suggestion for improvement: Instead of switching to gold you could switch to a "betting against beta" portfolio. Choosing Altcoins with lower performance than BTC in the bull and betting against them via Altcoin short BTC long could help you to minimize risks and peak out the bull market.
It might also be an approach for entry in the bear market to eliminate the risk of BTC going to zero, but obviously the risk of marketwide altcoinrun is much highter when entering Alts short / BTC long in the Bear market.      

[Queentoshi]

Very long read, but I managed to follow through till I could really make meaning of it.

Do you think, it is possible to outperform investing into Bitcoin taking the cycle into account instead of just dollar cost averaging into Bitcoin?

Your strategy of Dollar cost averaging according to the circles will be effective for a smart investor with a good knowledge of bitcoins and understanding of the market, a newer investor in bitcoins may not be able to monitor the market properly to know when they should DCA more and pause DCA, so we advise them to just first develop the habit of investing by regular DCA, before they are mature enough to now do DCA more smartly like with this your strategy.

[virginorange]

Very long read, but I managed to follow through till I could really make meaning of it.

Thanks for you feedback. I added a TL;DR summary.

TL;DR: A smart and disciplined investor can outperform a simple dollar cost average strategy, by buying Bitcoin depending to its price cycle. If Bitcoin is cheap (40% below trend), I would take a full position. If Bitcoin's price is expensive (more than 750 days ahead of trend), I would not add to my position and possibly even trim my position. If Bitcoin is neither cheap nor expensive I would dollar cost average into Bitcoin.

I also agree with you, that DCA is better for most investors. The bulk of my Bitcoin buys were end 2021, which only recently turned green. So I personally unperformed a simple DCA strategy.

However during the bear market I continued studying Bitcoin, kept my money off changes, didn't sell and continued to add slightly to my position.

For the next cycle I try to improve my setup.


Thanks for your compliments, I always enjoy you high quality feedback.

Indeed it was occasionally hard to understand your process, ... I guess it takes me about 90min to read and understand everything, but it was worth every second.

If I you would like me dig deeper into a particular point or make my explanation easier to follow, let me know.

Choosing Altcoins with lower performance than BTC in the bull and betting against them via Altcoin short BTC long could help you to minimize risks and peak out the bull market.    

How would you short? Via a centralized exchange or on Defi-projects in Etherium with Wrapped BTC. I think this would be quite a challenge managing all the margin requirements. Also shorting is difficult. You can loose 100% going long, but you can loose 500% or 1000% going short. So either you short with a stopp loss, forcing you to buy back even higher or you leave enough liquidity to survive the exuberance. Enough spare liquidity means low position size and low potential return. I tried shorting stocks, which ultimately went to zero, but with many spikes in between.

[coolcoinz]

I to answer the question if we can do better, yes we can, by performing a lump sum buy when bitcoin is in a correction. I used to start buying bitcoin in bear markets which I identified as -50% from the top. Everything below 50% was cheap in my view and worth buying and because of this I've managed to get my hands on $20k bitcoin last year. Sure, I could have done better by waiting and buying the bottom, but at that point nobody knew what the bottom was.

Going to gold and back is only for people who are willing to risk paying exchange fees, network fees, taxes on their gains. If I were to switch back and forth, every $1k transaction would mean paying $10 in exchange fees and another $20 in network fees, holding my bitcoin on exchanges and exposing myself as a holder... I prefer anonymity of a physical bitcoin exchange, where I do no KYC, pay with cash, don't have to report any gains and the only fee I pay is to the network.

[btc78]

Really impressive breakdown. I am sure this took you days or maybe even weeks to have completed.

Many think that those who invest understand all the complexities and technological nature of investing but that is simply not true. Many actually do not know how to analyze the market and only buys when it seems to be going up and sell when it seems to be going down. This really could help a lot of people myself included. Props and keep up the good work!

[Turbartuluk]

Thanks for your compliments, I always enjoy you high quality feedback.

Indeed it was occasionally hard to understand your process, ... I guess it takes me about 90min to read and understand everything, but it was worth every second.

If I you would like me dig deeper into a particular point or make my explanation easier to follow, let me know.

Choosing Altcoins with lower performance than BTC in the bull and betting against them via Altcoin short BTC long could help you to minimize risks and peak out the bull market.    

How would you short? Via a centralized exchange or on Defi-projects in Etherium with Wrapped BTC. I think this would be quite a challenge managing all the margin requirements. Also shorting is difficult. You can loose 100% going long, but you can loose 500% or 1000% going short. So either you short with a stopp loss, forcing you to buy back even higher or you leave enough liquidity to survive the exuberance. Enough spare liquidity means low position size and low potential return. I tried shorting stocks, which ultimately went to zero, but with many spikes in between.

Let's start with some idea's for possible improvements:

1. Formula y = 5,8323x - 39,296
-> I guess you have done some kind of logarithmic Regression like Trolololo did it once
-> therefore i guess the formula will be outdated with the years going and it would be helpful to reference the way of calculations

2. "Cyclical Component of Bitcoin price" graphics
-> due to logarithmic timeline it's somehow difficult to assign the peaks/years
-> maybe vertical halving-lines or labeling the peaks with year count could help as a guideline

3. Entry Price vs. Trend graphic
-> regarding the cloud of blue dots it might be possible to separate dataset into 3-4 parts (coresponding full cycles) and use different colours for the dots and regression lines
-> maybe some trend will show up (regarding hight or incline)  

4. Finding the Tops
-> while i understand the days ahead model itself i find it somehow difficult to switch between two different models for Highs and Lows
-> maybe you could use an approach like: y = (5,8323 -1 )x - 39,296 +9,5 for the Highs (numbers roughly estimated, mistaken thinking is very likely)


Regarding a switch to an Altcoins/BTC short portfolio it is meant to be a kind of hedge.
I guess i would use Kraken CFDs as i am familiar with them. There are about 50+ pairs for margin trading.
I would sort out all those Altcoins with a better performance than BTC during the Bull (regarding sharpe or sortino ratio).
I would also sort out Alcoins with extreme volatility like Memecoins. Lets say there are 40 Altcoins left which build the portfolio.

During the rest of the Bull (and regarding ETFs which only push BTC) it should be likely that these Altcoins do not excessive outperform BTC even in an Altcoin-Run (which could lead to margin call). On the other hand the crash for those Altcoins after the Peak should be much harder in comparison to BTC so you could profit from the crash.

Obviously you need collateral on CEX so you are affectet by Counterparty Risk. Also there might be drawbacks regarding privacy and taxes.

[Husires]

Since Dollar cost averaging is a strategy for beginners, and without going into details, beginners will prefers to follow the approach of buying an amount every week or month, and after gaining sufficient experience, will moves on to more complex strategies, so why not create a simple site called advanced DCA that has a simple recommendation indicator for beginners, where the indicator is suggested to either increase the amount of your investments or reduce them, something like the Fear & Greed Index, but it is for advanced DCA, through which there are points on the index that indicate whether you are investing in large amounts or not. Then the strategy will be better because it is intended for beginners, and beginners may not be able to understand the charts you mentioned above.

[SilverCryptoBullet]

With ROI, very high, it is attractive enough to DCA Bitcoin.
https://casebitcoin.com/

I appreciated your work that is complex and I don't understand all of your charts and math but you did good works to convince newbies that Bitcoin is a good investment choice for them.

https://dcabtc.com/

[d5000]

Very good thread I would have missed if there wasn't a post in a well known German thread. I should frequent this subforum more

In general terms I agree with most of your points, and I think even newbies can come to the simple conclusion that you can almost safely "catch the falling knife" (in an intelligent way, like the described DCA-based strategy, not going all-in at once!) if the price has fallen low enough from the last top. This is always what baffles me: Many people wait way too long to buy when the bear market reaches its end, waiting for even lower and lower lows, but then they return when the bull market is already fully in force (this was the case when $32.000 was passed in late 2023 - those going in at this point missed more than 100% profit!) .

Such strategies becoming more popular would also increase the stability and confidence in Bitcoin in general, the bear markets would be less deep and progressively cause less fear which is a virtuous cycle.


I have only one thing to add:

One could perhaps improve the model trying to spot anomalies caused by external effects. This is based on the theory that the long-term price trend is generally based on an "adoption" curve and has thus "fundamental" reasons, but there can be some temporary effects in place, which seem to change the trend but in reality are only temporary increases of either demand or supply.

The best example was the crash in March 2020 due to the Coronavirus crisis. But there are of course also more gradual and longer-lasting effects, like the influence of the interest rate in major markets like the US and Europe. I believe that for example in 2019-2021 the bull market could have started earlier in full force and have been stronger if the Corona crisis didn't materialize, and the May 2021 top possibly was "lower than normal" for the effect caused by the China mining ban, late Coronavirus crisis effects and the start of the inflation increase in the US and Europe. On the other hand, the El Salvador Bitcoin law could have led to an anomalous second top in late 2021, which was not seen in most other coins. If the market had developed like in previous cycles, without this "positive black swan", then the late 2021 top would perhaps only be at 60% of the May 2021 top.

How to include that into a mathematical model? One could add a kind of "positive/negative general market sentiment bonus" to the figures. This bonus would take into account the situation in high-risk markets (mainly stock markets in US, Europe, India and other relevant countries) but also very relevant Bitcoin- or crypto-specific news and add or subtract this number from the expected tops or bottoms.

[Turbartuluk]

How to include that into a mathematical model? One could add a kind of "positive/negative general market sentiment bonus" to the figures. This bonus would take into account the situation in high-risk markets (mainly stock markets in US, Europe, India and other relevant countries) but also very relevant Bitcoin- or crypto-specific news and add or subtract this number from the expected tops or bottoms.

Interesting approach. What kind of sentiment analysis you have in mind? Do you know some good "pure sentiment" indicators?

The first sentiment indicator i had in mind was the fear and greed Index, but as it has a lot TA in it (volatility, momentum, volume, dominance) i am not sure if it is suitable, it could depend more on chart history than predicting it.

[d5000]

Interesting approach. What kind of sentiment analysis you have in mind? Do you know some good "pure sentiment" indicators?

First, I don't know if "sentiment" is the correct word, because "sentiment"-based indices like Greed & Fear are partly based on the price evolution itself. Perhaps one could call this indicator "general market situation". Or simply "External effects influence".

I'd thought about a simple "bucket list" of items you can assign a value and apply it to the general "trend index".

For example:

- Interest rates in a significant part of the most markets (North America, Europe, India, China) more than 1.5 percentual points above the 20-year average -->  -10%
- " " more than 1.5 percentual points below the 20-year average --> +10%
- same for world economic growth rate (this would explain the Covid anomaly)
- Regulatory/political event affecting demand negatively in a significant part of the Bitcoin market (e.g. trading/mining ban in China) -> -5% to -20% depending on the size of the market affected
and so on.
- Regulatory/polical event of high impact in a smaller market, but first of its kind (example: El Salvador's acceptance as legal tender) +/- 10%

The good thing is that such indicators are hard enough be tested with past market data. I would limit them to important events. For example, El Salvador's Bitcoin law would be an item I'd include, but the Central African Republic's short-lived Bitcoin law not as it wasn't the first of its kind and its impact was also low. It would be of course object of debate which "hard" indicators determine what to include, but that's a challenge for any of such indices.

[Turbartuluk]

~

Hmm. While I agree, that those items mentioned have major impact, i am not a friend of that "bucket list" idea. As it is easy to add items it is likely that you get an overfitted model in the end. And when that model does not fit the future prices anymore it is likewise easy to say, that is because another item that is not on the list and has to be added. Even if you had something different in mind, i dont have a clue how you want to prevent that risk?!

On the other hand you could also argue, that all those items mentioned (an much more major and minor impacts) were already priced in and therefore part of chart history. Additionally you don't have to discuss weighting of the list items, as market has already decided for you. In the end you come back to a (mathematical) model based on chart history as it keeps things more simple.

To my mind it is fine if you keep flexibility and don't get a 100% fit with past price history. Such a high fit is often only possible with a very small number of target points (e.g. only highs or lows instead of 1D closing prices) and would suggest an accuracy which is unrealistic for future "prediction".   

[Etranger]

To my mind it is fine if you keep flexibility and don't get a 100% fit with past price history. Such a high fit is often only possible with a very small number of target points (e.g. only highs or lows instead of 1D closing prices) and would suggest an accuracy which is unrealistic for future "prediction".   

I agree that historical analysis cannot be considered a sufficient basis for building future forecasts, but still it is necessary to rely on it, because retrospective image gives an idea of what was already possible for the crypto market, and therefore - can be possible again. It also makes it real to draw analogies with those new events that have taken place and what their impact on the price may be.

[d5000]

Hmm. While I agree, that those items mentioned have major impact, i am not a friend of that "bucket list" idea. As it is easy to add items it is likely that you get an overfitted model in the end. And when that model does not fit the future prices anymore it is likewise easy to say, that is because another item that is not on the list and has to be added. Even if you had something different in mind, i dont have a clue how you want to prevent that risk?!

Obviously an exact price prediction is impossible. What virginorange instead pretends (if I interpret correctly) is to find general guidelines for retail investors to know when it's a good time to increase the position in Bitcoin, and when it's better to decrease it, not an exact moment when to buy or sell (which would be impossible).

The idea is basically to not ignore these "external" factors. Everybody interested in this model can vary his own bucket list and add or remove items - different traders will always process information differently (at least if they're not using the same trading bot or so ) and thus concede some factors more and others less weight.

It would be interesting to create a software tool to be able to contrast such variables with the past price data, being able to adjust the weights of each factor. This tool could then be used in combination with virginorange's general model.

On the other hand you could also argue, that all those items mentioned (an much more major and minor impacts) were already priced in and therefore part of chart history. Additionally you don't have to discuss weighting of the list items, as market has already decided for you. In the end you come back to a (mathematical) model based on chart history as it keeps things more simple.

The problem is that if you base a model on a chart history which was partly influenced by external factors, you may get a distorted model.

The most significant example would be the "underperforming" 2021 bull run. In my opinion without Covid and the begin of inflation growth later that year, the bull run could have been stronger and already reached or approached $100.000 in that year. That was expected by many Bitcoiners that time. The theory about an "underperformance" of course is only an assumption, like everything we discuss in these sections. But this "underperforming" bull run could lead that in the current bull run (and the next ones, if there are any left ) you may sell too early because the model considered 2021's "underperforming" run part of a general trend where highs tend to be lower. I don't dispute here that the volatility to the upside is probably decreasing, but from my impression particularly the last stretch of 2021 bull (November high was less than 10% above the May high) was quite weak compared with earlier bulls.

To my mind it is fine if you keep flexibility and don't get a 100% fit with past price history. Such a high fit is often only possible with a very small number of target points (e.g. only highs or lows instead of 1D closing prices) and would suggest an accuracy which is unrealistic for future "prediction".    

Fully agree here.

[laijsica]

The case for countercyclical buying
However, if we could be reasonably sure that Bitcoin were near the top or near the bottom, we could invest accordingly. For example, we could stop DCA during the euphoria and double DCA during the bear market. Or we could sell some of our bitcoin close to the top and buy back close to the bottom. This would be a simple DCA strategy.

Your opinion about DCA strategies is very reasonable and a good example of investment. DCAing in the bear market is limited and the tendency to increase in the upper market is promising. Bubbles are constantly entering the market, even in a bull market. I've covered long-term moving averages in the past on DCAing. But I didn't think like you. As this analytical thinking opens new windows for market trending