Guest of this issue: Murphy, on-chain data researcher, Twitter @Murphychen888
*All text is for sharing only and does not constitute any investment advice.
TL;DR
About Trader Murphy
The core of trading is "people". A person's experience, background, personality, and capital allocation determine the formation of their trading strategy.
What is Murphy's trading strategy?
1) Capital volume and allocation ratio: Own funds, 90% allocated to mainstream coins (BTC, ETH, BNB), 10% allocated to altcoins
2) Expected return: 200%-300%
3) Tolerable drawdown:
For mainstream coins (BTC, ETH, BNB), only consider the cycle, no stop loss, a 20%-30% pullback is normal
For altcoins, stop loss when it falls below 20%
4) Trading logic:
Timing in large cycles, i.e., buying in a price range near the relative bottom and selling in a price range near the relative peak within a large cycle, no scalping, contracts, or high-frequency trading
Guiding trading through on-chain data analysis, i.e., conclusions with relatively high certainty derived from various data indicators
Why did Murphy form such a trading strategy?
1) Trading experience: Murphy did not have a stable trading strategy in the previous cycle. When the market first rose to $63,000, he gradually sold chips. Later, when the price broke through a new high to $69,000, he felt very regretful and anxious due to various FOMO emotions in the market. Therefore, in 2022, he decided to find a trading strategy within his cognitive range, based on evidence, which can be effectively executed, replicated, and closed-loop, to avoid being disturbed by emotions.
2) Professional background: Before entering the cryptocurrency industry, Murphy worked in the traditional financial industry for many years in marketing, where he needed to use data to track and adjust various marketing investments. This experience made Murphy sensitive to data, so when looking for a "evidence-based" trading strategy, he naturally chose data as the basis.
3) Personal character: Murphy is very cautious and seeks high certainty. At the same time, he hopes that his trading strategy can be effectively executed, replicated, and closed-loop. Therefore, he chose "timing in large cycles" and "using data analysis to guide trading".
4) Capital attributes: The attributes of capital determine the capital cycle, and the capital cycle determines the investment method. Murphy's capital is his own funds, without leverage, and comes from profits in the previous cycle, with low cost.
What kind of person is Murphy's trading strategy suitable for?
Meeting two characteristics:
1) Conservative, seeking high certainty, and can hold assets relatively long-term
2) Trend trader, with the vast majority of positions in mainstream coins
Murphy's Trading Story
Real knowledge is verified through practice. Reviewing and reflecting on specific trades can provide a more intuitive understanding and learning of trading strategy application.
In this cycle, what indicators/data did Murphy use to judge the timing of bottom fishing?
There are mainly the following 5 data points:
1) CVDD, Cumulative Value-Days Destroyed, CVDD (USD) = ∑(CDD × price) / (days × 6,000,000)
When BTC is transferred from one investor to another, the transaction not only has a dollar value, but also destroys the time value associated with the time the original investor holds the tokens. The CVDD index is a numerical value calculated by a mathematical model. Its numerator is the coin days destroyed index, i.e., CDD. For example, if you hold two BTC for three days, and then move them after three days, once you move, the value of your coin days destroyed is reset to zero. The most significant feature of the CVDD index is that it never retraces, i.e., the CVDD value for each day will be higher than the previous day's value. Since historically, the price of BTC has never effectively fallen below the CVDD value, it can effectively evaluate the market's bottom price through CVDD.
In October 2022, when the price of $BTC was around $19,000, the CVDD was roughly $14,800. Murphy estimated that if bottom fishing at this time, the cost of buying BTC would be around $20,000, and at the same time, the conservative expectation for the peak price of the bull market is $50,000, which is within the expected range, so he decided to bottom fish.
2) RP (Realized Price) and BP (Balance Price)
RP does not include exchanges and represents the average cost of turnover on the chain; BP is the BTC cost that removes the time value of funds (income from other uses of funds such as deposits), reflecting the fair market price. In a bear market, when the price of BTC falls below RP, and when it fluctuates around the middle channel of BP and RP, it is a very good buying point.
3) PSIP (Person Supply in Profit)
The essence of PSIP is the proportion of profitable chips in the circulating chips. When it is less than 50%, it means that more than half of the people are at a loss, which is usually an extremely extreme bottom in a bear market.
4) LTH (Long Time Holder) NUPL and MVPV
NUPL uses different colors to represent the level of profit/loss on the Bitcoin chain. When it turns red, it means that long-term holders have all surrendered (i.e., sold chips at a loss), which is a market low point and can be bought.
MVRV considers the ratio of BTC's circulating market value to the realized market value. The calculation method of the realized market value is to sum the prices of all moved BTC from the last time. The difference between the circulating market value and the realized market value is the unrealized market value. The larger the ratio of the two, the larger the market bubble, and the easier it is to cause profit selling. The smaller the ratio, the more undervalued it is. The generally referenced numbers now are, sell when MVRV is greater than 3, buy when it is less than 1. At the same time, in the rising trend of a bull market, look at the MVRV of short-term holders, and when looking for the bottom in a bear market, look at the MVRV of long-term holders. The logic here is that in a bull market, the price of BTC is determined by short-term holders, while in a bear market, it is determined by long-term holders.
5) Miner's perspective, with two data points as examples
Mining cost: The mining cost of miners can be calculated using different models. The closer the market price is to the cost, the closer it is to a phase bottom, and it can be bought. The specific calculation method can be found in Murphy's previous tweets:
Mining pulse index: The 14-day average time interval of block generation and the deviation from the target (the reason for using 14 days is that it is a difficulty adjustment period). If the deviation is a positive value, it indicates that some miners have shut down, and this is also a good time to buy.
Murphy has also organized and explained these indicators. You can refer to his explanation here: https://x.com/Murphychen888/status/1814284223597187467
How did Murphy accumulate these indicators?
In three steps:
First, find a teacher. Learn from the content shared by on-chain data analysts in the market, clarify the logic, and experiment and calculate according to their algorithms. If problems are found, adjust the parameters.
Second, accumulate and categorize your own indicator library. Record and track indicators in an Excel spreadsheet according to name, source, algorithm, function, category (emotional, trend, top judgment, bottom judgment, etc.), continuously optimize and iterate, and then use different indicators according to different scenarios.
Third, backtesting. Continuously improve the effectiveness of the indicators through backtesting.
Murphy has a classic summary: Macro expectations lead to changes in emotions, and emotions affect supply and demand. Supply and demand ultimately determine the price of BTC. By observing and analyzing on-chain data, finding the factors and logic behind the data that affect supply and demand, it is possible to judge the possible trends that may form next, or the transition of this cycle, and then combine other methods, such as technical analysis of candlesticks, to draw more certain conclusions to guide trading decisions. However, it should be noted that even the perfect indicator cannot provide an exact entry point. It can only give a signal in a certain price range. When the signal appears, it should be executed decisively and resolutely based on one's risk tolerance and with proper position management.
In this bull market, how will Murphy judge the "peak" time?
1) Logically:
In the entire BTC cycle, the supply side is long-term holders (LTH), and the demand side is short-term holders (STH). From the bear market to the bull market, it is a process of long-term holders continuously distributing chips to short-term holders. Then, from the peak of the bull market to the bear market, it is a period of short-term holders continuously selling off chips and returning them. When this cycle has not yet seen LTH profit-taking or exiting behavior, it can be considered that this cycle has not yet been completed. Conversely, when LTH has started to sell, or has sold almost all, it is the time to "escape the peak".
In each bull-bear cycle transition, short-term holders will have two large-scale receptions. The first time often occurs before the halving, mainly for three reasons: first, halving expectations; second, the emotional release accumulated during the long bear market; third, as the market warms up, some medium and long-term chips will be distributed to STH.
The first reception of this cycle occurred when BTC exceeded $70,000. The timing is different from the previous two cycles mainly because the ETF approval changed the rhythm of the entire cycle. However, from a data perspective, the first reception chips were not many, so there will be a larger-scale reception next. The occurrence of the reception is also when long-term holders distribute chips, and these two times are the relative peak range of the bull market cycle. However, this peak range may be a double top or a triple top, and the price of the latter top is not necessarily higher than the previous one.
2) From the perspective of indicators, the main ones to consider are:
- MVRV
When MVRV is greater than 3, it is time to start selling according to the trading strategy. Selling can be completed in the process of MVRV gradually increasing from 3 and then falling back to 3, forming a smile curve in reverse.
- URPD
URPD reflects the on-chain chip structure data, which can tell us how many chips have changed hands in a certain price range. The more chips accumulate, the stronger the consensus, and a consensus range will be formed in this price range, creating resistance and stickiness effects. Resistance means that when breaking through from below, it is not easy for selling to allow the price to break through this range. Stickiness means that when falling through from above, even if the price falls below the range in a short time, without forming a large consensus and turnover, the price will quickly rise back up. However, URPD alone cannot be used to judge the peak. When other multiple data points show peak signals, URPD can be used to verify the effectiveness of the data.
For a detailed explanation of URPD data, you can refer to Murphy's tweet here: https://x.com/Murphychen888/status/1753331946439217362
Therefore, the conclusion is: Judging the peak will be more complex than judging the bottom, and it requires comprehensive analysis using multiple data (rather than a single indicator). However, based on the "three-line convergence" indicator, the relative peak range will roughly occur from March to April 2025, but this is not objectively supported and needs to be combined with macro data to see step by step. Referring to the previous three cycles, a phenomenon will occur in the peak area, which is that MVRV is higher than 3.
When will Murphy's indicators fail? What is his Stop Doing List?
First, it is not a "precise" indicator and cannot judge the specific long and short positions or the specific price movements in a very short time.
Second, it cannot be used to analyze altcoins. This is because BTC's characteristic is the UTXO structure, which has a timestamp function to record the time point when unspent transaction outputs are generated and the USD value of BTC at that time. Many on-chain data cannot be calculated for altcoins that do not have the UTXO structure.
The Stop Doing List consists of 5 points:
1) Stop relying on luck. You can admire others' wealth, but do not try others' track with a mentality of "if he can, I can too".
2) Stop being emotional. There are many opinions and noises in the cryptocurrency industry. Think calmly, face the data objectively, and avoid being disturbed by noise.
3) Stop frequent trading. Because this is beyond one's own capabilities.
4) Stop relying on a single source of information. Do not use a single indicator as the conclusion of a complete logic.
5) Stop making decisions without research. Others' opinions can only be used as a reference. Decisions should always be based on one's own research.
Murphy's "Must Read"
The growth of excellent traders depends on continuous external input, learning from other excellent people, and exploring content that is meaningful for reference. Through other people's "Must Read" lists, one can continue to accumulate and grow.
Murphy likes bloggers who have originality and can continuously output content.
Chinese:
@Phyrex_Ni, Ni Da, interprets the impact of macro data and events on cycles
@Trader_S18, macro + candlestick analysis of the market
@qinbafrank, macroeconomic observations, research on global capital liquidity
Foreign:
On-chain data researchers @CryptoCon @_Checkmatey @therationalroot @TechDev_52
No one can predict the market with 100% accuracy, but through the accumulation and adjustment of indicators, one can increase the probability of winning in trading. Hopefully, we can all find methods that bring us closer to certainty. See you next time.
Conversation Record
FC
Let's get started. The best time to invite you should have been when the market was down, but we also discussed last week, and you thought it was close to a bottom last week. Could you briefly introduce yourself first?
Murphy
My major in college was advertising. After graduation, I worked for a short period of time and then transitioned to the traditional finance industry. In 2005, a few friends and I started a traditional financial service company. I was mainly responsible for marketing, so my professional background and work experience made me sensitive to data because I had to deal with marketing data and user profiles or needs in my previous work. I entered the cryptocurrency industry in 2017 by chance. I consider myself an old hand in the industry, except for the first cycle when I was a novice and didn't have the experience. I've experienced the ups and downs of the following three cycles and all the hardships that come with them.
In 2022, I started researching on-chain data and found that I had a deep understanding or sensitivity to this data. The original intention was to help myself, as a retail investor, find an effective, replicable, and closed-loop investment methodology. When I started writing tweets, I didn't expect to have so many followers like now. My original intention was to use Twitter as a diary of my growth, recording every bit of research and insights during each cycle. When the next cycle comes, I can look back at my research, insights, and experiences at that time, which I thought would be wonderful. But now, with many followers, I not only have to be responsible for myself but also for the followers who read my tweets. That's a simple background introduction.
FC
You have a marketing background, and we actually have some similarities. I used to be responsible for marketing and sales. I met BMAN because before he entered the cryptocurrency industry, he ran a public account called "Professor Li," mainly about marketing. I used to joke that I was his outstanding student at the time. I even visited his office. At that time, no one had heard of the cryptocurrency industry. Later, I joined 36KR. So, I want to know, since you mentioned that marketing is related to data, what specific area of marketing were you involved in? Was it more related to the middle office?
Murphy
Yes. Because marketing serves sales and customers, areas such as SEO (search engine optimization) and purchasing ad space were mainly handled by our department. I would check the backend every day, such as the click-through rate and cost of conversion for click-type ads, as well as the user conversion channels (paths) and the user's journey from one webpage to another before converting. Finally, we had to calculate how much to invest in advertising, how much manpower to invest, and what level of bidding or optimization ranking to achieve in order to bring actual user conversions to the company. All of our work was related to data.
On the other hand, marketing also involves promotion. We had to communicate with advertising or design companies on how to promote our brand or introduce the high-quality services our company provides. We had to make sure these suppliers understood our key points and our industry and the characteristics behind our company. So, everything I'm doing now, including what I'm doing on Twitter, is quite similar to my previous work. At least, I can apply the experience and skills I accumulated from my previous work to what I'm doing now.
FC
I understand. It's quite interesting. I like to start with a background chat because many of your decisions today are related to your past experiences. You mentioned effective execution, closed-loop, and replicability. I understand that these should be the basic principles of your entire trading strategy. How did you come up with these three? Why these three?
Murphy
First of all, effective execution is very important. I found that many trading systems are subjective. Some people may find a particular trading system very suitable, but I may find it difficult to execute because I may be influenced by emotions or the source of the data. I just can't execute it.
As for replicability, it means that when I explain my entire trading strategy to my colleagues or friends, they can continue my entire trading mindset or some investment methodology and benefit from it. They can replicate my entire trading model and form a closed loop. It's quite simple because all of our theories must be based on cause and effect. You can't just say you think it should be this way, so it is. So, I think these are the three most important points in my investment methodology.
FC
I understand. You mentioned that some strategies within effective execution are not suitable for you. You must have tried things like contracts before. In turn, what kind of person do you think you are? What kind of trading strategy are you pursuing that suits you?
Murphy
I think I'm a cautious and introverted person. As I retweeted your tweet last time, I'm not someone who likes to be in the spotlight. In my direct messages, I often receive invitations from project teams or friends to participate in spaces, but I politely decline. So, for myself, I don't excel at or suit high-frequency trading, contract trading, or long and short trading. I prefer to use time to exchange for space, or to use data that I am relatively certain about to make a decision to buy. In simple terms, when I make a purchase, even if the price drops or my expectations are not met, such as a 20% or 30% drop, I can still sleep at night and go out with friends in the morning without having to watch the market every day. I prefer this kind of trading method, which I believe I can execute effectively. When I see certain signals or entry points based on a few data or indicators that I think are worth considering, I will buy a portion of my position. I usually manage my positions and the cycle allocation of my funds in a way that I find comfortable. I think this is effective execution.
FC
Let's talk about trading strategy. With your current trading strategy, how much capital can you accommodate? What is your expected return? What is the cycle of your funds? What is the maximum drawdown you can accept? For example, if it drops by 20%, what will you do? Will you add to your position or reduce it?
Murphy
First of all, I want to say that for each of us, the cost and source of our funds, as well as our investment needs, are different. Personally, because I made some profits in the last investment cycle, and some of my friends who know me well know that we also did some mining in the last cycle, so the cost of obtaining BTC chips for me is relatively low, probably just a few thousand dollars. After the peak of the last cycle, I withdrew a portion of my funds. At that time, there was news that Binance was going to close its OTC channel and might have to connect to a third-party channel in the future. I thought it might be difficult to withdraw funds if Binance closed, so I withdrew a portion. The funds I currently have left are actually very low-cost for me, or even zero-cost, as they are entirely profit. To put it in the most extreme terms, even if all of this capital were to go to zero, it wouldn't affect my life. But for some of my friends who have chatted with me privately or sent me direct messages, I understand that some of them may have funds that are scraped together from their daily lives, or even borrowed funds, and these funds may not have much use for them now, but they will need them for marriage or buying a house in a few days or in the near future, so the nature of the funds is different. The nature of your funds determines your investment method, your expected return, and your fund cycle. So, from my personal perspective, my current situation, combined with position management and a cyclical trading strategy, makes me feel very comfortable. But for other friends, you need to consider your own situation, your risk tolerance, and the cost of your funds to develop a method that makes you feel comfortable.
As for the expected return, it's okay for me. When I built my position in this bear market cycle, I set a target for myself of 300%, between 200% and 300%. In fact, I have already far exceeded 200% and should have reached 300%. So, for me, in the next cycle, even if this bull market doesn't reach the heights we expected, such as 150,000, 120,000, or 100,000, if it doesn't reach that high, I would still be very satisfied with 80,000 to 90,000. I would have achieved my expected return. Other friends will have to look at the specific entry cost.
As for the acceptable risk, my followers who have been reading my tweets should know that I rarely talk about or almost never talk about altcoins. It doesn't mean I don't buy altcoins, but the majority of my positions are in mainstream coins, such as BTC, ETH, and BNB. For BTC and ETH, I only look at the cycle and don't set stop-loss orders. In a bull market cycle, in my opinion, a 20%-30% pullback is very normal. Like the pullback a few days ago, it probably hasn't reached 30%, just over 20%. In the previous cycle, I remember there were probably 6 or 8 instances where the pullback exceeded 30%. So, if you set a very small stop-loss, you will keep getting stopped out. Once you've been stopped out, based on my experience, you probably won't buy back in because you won't be able to buy at the same price. For altcoins, I buy relatively small amounts. I might use about 10% of my total position to buy some altcoins, but my skill in buying altcoins is not great. Sometimes, when I make a profit, it's only a small amount, so it doesn't matter if I lose. So, I feel like even if I make a profit of one or two times, it hasn't met my expectations, and then it's a roller coaster ride. For altcoins, I usually set a stop-loss of 20%. If it falls below 20%, regardless of whether there might still be a bull market or a possibility of a rebound, I will close the position first and then see. But I probably won't buy back in. I acknowledge this 20% loss because it's a problem with my own perception, and I have to take responsibility for my perception.
As for the fund cycle, because my funds are all self-owned, there is no borrowing or leverage involved. So, I usually invest a portion of my funds at a relatively low point (not the absolute bottom) in a bear market cycle. I usually buy 30% first and then look for opportunities to buy more. When I think that based on the overall on-chain data, including various macro aspects, the bull market cycle is almost over, or it's at a relatively high level, I will gradually sell off my positions. That's roughly the situation.
FC
Understood. So, you don't do short-term trading.
Murphy
Right, I don't do short-term trading. I'm not good at it.
FC
We've been talking for the past two months. I think from a data perspective, we actually couldn't see the peak at 78,000 the first time, but you should have seen the peak during the second phase. In fact, I also belong to the buy and hold camp. I feel like I can't sleep after selling, and I don't know when to buy back, so I don't really care about making that profit.
We usually discuss four parts. We just talked about background and the entire trading growth experience. Next, we might delve into investment strategies, and finally, talk about personal growth. So, regarding the trading growth experience and the last question, you said you were in marketing and probably dealt with data more, so how did you transition from marketing to data analysis? Because you have many options, you can play with candlestick charts, or you can look at altcoins, as there are many tools to analyze data. Why did you choose on-chain data for macro analysis? What was the process like?
Murphy
Actually, in the last cycle, there were two peaks. The first peak was at 63,000, and after that, I couldn't resist and thought that if I didn't sell at that point, it might be close to the end. So, when it dropped from 63,000 to 52,000, I started selling, and I kept selling until it reached 48,000. That was the first time. After it rose to 69,000, I started to regret and wondered why I sold at 48,000 and didn't buy back when it dropped to 40,000. When it rose to 69,000, many people thought that after such a long period of washing out, the main force finally made BTC reach a new high again. The last time was 63,000, so how could it have peaked at 69,000? This cycle was bound to exceed 100,000. I remember there were many such opinions, so I was particularly anxious at the time and felt foolish for selling at an average cost of over 50,000 and not buying back at around 40,000.
Afterward, I analyzed it myself and realized that at the time, I was just following the crowd and listening to hearsay, which was not the most correct investment path. I felt the need to find a trading strategy within my own cognitive range that was based on evidence. When making any decision, whether it's buying or selling, there must be a basis. This basis may be something you believe to be correct or something you think is correct. Maybe it's actually incorrect, but it doesn't matter. You can review, summarize, and improve later. The point is, you need to find a basis. Only with a basis can you execute effectively. That's what I believe. So, later on, I also felt that I needed to try to avoid emotional interference or market noise as much as possible. Of course, emotional fluctuations are human nature, and the seven emotions and six desires are not something ordinary people can restrain. But can emotions be quantified with data? By observing changes in data, can we truly do the opposite of what others do, such as being fearful when others are greedy and being greedy when others are fearful?
In fact, I look at much more data every day than what I share. I can't possibly put all of these indicators on Twitter, I can't write about them one by one, but I will put up the most important or what I consider to be the most critical ones, such as emotional indicators, and explain to everyone why this indicator is a quantified data of emotions, where its principles lie, and how we use these indicators to judge the market's sentiment. When this sentiment is very close to the historical market sentiment ice point, or a limit value, it often cannot continue to influence our emotions. We have to do the opposite, we have to go against human nature.
Actually, many times, some friends may feel that analyzing on-chain data is a futile effort. What is the essence of a futile effort? This is how I understand it: the essence of a futile effort is to ignore change. You fail to realize the changes in the environment and conditions, and you rigidly adhere to established rules when interpreting data. You only focus on superficial markers and signals, treating historical events as equivalent to the present. This is the essence of a futile effort, ignoring the logic behind the data and its actual situation.
Let me give an example. A few days ago, when I tweeted, including when we communicated, I mentioned a three-line convergence indicator. I always thought this indicator was a very standard mystical indicator, or a futile data type. This is because it does not consider the external environment of each cycle, the specific changes that have occurred, such as the approval of an ETF in this cycle, and the macro conditions in the previous cycle may be different from the current macro conditions. The current conditions may be macro-tightening, while the previous cycle may have been macro-loose, but in the previous cycle, an ETF for BTC futures was also approved. It doesn't consider the changes in the environment, it simply overlaps the MVRV data after each halving, and looks for a trend of resonance. However, if we look at it the other way around, the three-line convergence is a futile data, but its composition consists of 3 or 4 MVRV data. MVRV data is not futile, it has objective logic. For example, it considers the ratio of BTC's circulating market value to its realized market value. Everyone knows what circulating market value is, I don't need to explain it further. What about realized market value? When BTC moved last time, the prices it generated for all the BTC moved at that time are added together, and that is the realized market value. What does this ratio mean? If the current circulating market value is calculated at 60,000, but if I have a BTC that moved last time at 50,000, its realized market value is 50,000. The person who took over at that time took it at 50,000, so their cost is 50,000, and now it's at 60,000, so the difference is 10,000, which is an unrealized market value. So, when this MVRV value is larger, it indicates a larger bubble and is more likely to trigger profit selling. So, many professional institutions say that if MVRV reaches 3, you can sell it, and if it's below 1, you can sell your kidney. Why? Because the smaller the ratio, the more undervalued it is, and it has a lot of investment value. What's the logic behind being below 1? Being below 1 means that the average cost of BTC obtained by most people is a loss, so when most people are at a loss, do you think this is a good thing to buy? I think it is.
Once you understand this principle, on this basis, we can add some algorithm variables and evolve it into various different usage methods, flexibly changing it, and this is no longer a futile effort. Our data is actually diverse, including macro data, candlestick data, and on-chain data. The key is how you use it, how you interpret it, and whether you combine it based on understanding the principles. Some time ago, I saw many people on Twitter opposing the logic of using macro data to guide BTC prices. They thought it was impossible to follow macro trends, and studying macro trends was unnecessary and a waste of time. But I think it's a matter of integration. How do you integrate? Even if you have a strong theoretical foundation in candlestick analysis or practical experience, if you combine macro data, it will be very helpful for your trading, because only by considering multiple factors can you arrive at an objectively higher certainty conclusion. So, in my previous tweets, I summarized my own insights into a sentence: I believe that currently, macro factors lead expectations, expectations change emotions, emotions affect supply and demand, and supply and demand ultimately determine the price of BTC. So, whether it's through macro information or technical theories on candlesticks, most of the time, what we analyze is an expectation. Through observing some on-chain data, what we are actually observing is the change in chip supply and demand, which ultimately reflects in the price. In the end, it's a matter of probability. No one can predict the market 100% because you can't travel back from the future. You can only say that under certain objective conditions, the probability of arriving at a conclusion is higher than the probability of not encountering such a situation. So, through observing and analyzing on-chain data, I mainly try to find the factors and logic that can affect supply and demand, in order to judge the trends that may form next, or the potential transition of these cycles. Lastly, it's worth mentioning that everything I've mentioned above is only related to Bitcoin and does not involve any altcoins.
FC
Understood. Can you describe how you currently use on-chain data through your own story, such as how you bottomed out at the time, and which indicators you used?
Murphy
My real experience is like this. I started studying on-chain data in 2022, just when BTC dropped from 69,000 to 50,000, then paused for a while, and then dropped to 40,000. It was at that time that I started studying on-chain data. At that time, I judged the bottom like this: there weren't as many in-depth studies as there are now, and the indicators I used were very simple. This is also why I wrote some tweets a while ago with the title "Simple Indicators Have Wonderful Uses." Those who have seen my tweets should know that sometimes you don't need to pay attention to how complex the algorithm of an indicator is, or how impressive its logic sounds. In my actual experience, many free indicators that you can find are very useful and effective, as long as you combine them and don't use a single indicator.
For example, when I bottomed out last time, the first indicator I looked at was CVDD. It's a consumption index algorithm, with the numerator being the data of coin days destroyed. For example, if you hold two BTC and move them after three days, your coin days destroyed value is reset after three days. How is it reset? If you hold two for three days, 2 x 3 is 6, and this 6 is reset. Its denominator is multiplied by the price of BTC, and the numerator is the number of days you held it, multiplied by a calibration value, which is 6 million. I don't actually know why it's 6 million, but it's said to be calibrated at 6 million, and its price trend forms a very interesting trend. It never retraces. For example, today's CVDD price is always higher than yesterday's, and tomorrow's will be higher than today's. What's the benefit of this? For the overall market price decline, it constructs the market's bottom in a bear market, and it gradually rises. This lower limit is very useful. So, when I looked at CVDD at the time, BTC should have been at 19,000, and CVDD was over 14,000. So, many friends at the time said it should drop to 12,000 or 8,000 in this cycle. If you've experienced the previous cycle, you should know that what I said at the time was indeed the case. I used CVDD to calculate the profit and loss of the data, because in BTC's entire history, it makes sense to use 6 million as the calibration value. Once you use 6 million as the calibration value, the data will show that every time in history, the deepest drop never fell below CVDD, and CVDD has always been rising, it doesn't retrace. So, when I saw that CVDD was over 14,000, if I thought this data was reliable, it was impossible for BTC to drop to 14,000 in this bear market cycle, assuming it was at 19,000 at the time. The difference between 19,000 and 14,000 is 5,000 dollars, but from the depth of 19,000 at the time, I thought that in the next bull market cycle for BTC, although it might not exceed the previous peak of 69,000, it should be no problem to reach 40,000 or 50,000. If my average cost was around 19,000, doubling to 40,000 would be acceptable in terms of profit and loss ratio, so at that time, I felt that I could buy.
However, whether to buy or not, I still need to look at it from other aspects, such as the RP and BP data. RP stands for Realized Price, which I often mention in my tweets. We can simply consider it as the average cost of turnover on the chain, excluding exchange turnover. Balance Price, on the other hand, is a cost that removes the time value. For example, when you buy BTC for 100,000 RMB, does this 100,000 RMB have a cost? Definitely, because if you put 100,000 RMB in the bank, you have interest, and if you borrow 100,000 RMB, your loan also has a cost. So, every fund has a time cost. Balance Price actually removes this time cost and represents a fair market price. In a bear market, when BTC's price falls below RP and consolidates between BP and RP, I believe this is a very good buying point. This period can be long, and you don't need to make this decision in one or two days; it may give you several months.
Another indicator is PSIP (Person Supply in Profit), which I often mention on Twitter. Its essence is the proportion of profitable chips in the circulating chips of BTC. If the proportion of profitable chips is high, everyone will be happy and FOMO. If the proportion is low, everyone will be fearful. You can look at historical data; when PSIP is less than 50%, it's mostly an extreme bottom in a bear market. Why do I say this? Because when 50% of all chips on the chain are at a loss, and you also need to exclude lost chips and long-term holders who haven't moved their chips for seven or eight years, the actual proportion of chips at a loss is far more than 50%. When the majority are at a loss, is there anything to fear about buying at this time?
There's also the NUPL data of LTH (Long Time Holder), which is very effective and simple. It uses different colors - red, yellow, blue, green. When it turns red, it means long-term holders have already given up. At this point, what is there to worry about when buying? When it comes to long-term holders, STH (Short Time Holder) MVRV is also a very important data, for example, when looking for opportunities to get on board during the pullback in a bull market. I usually advise everyone to look at STH MVRV during the rhythm of the rising market, because the entire price of BTC in a bull market is determined by short-term holders. When long-term holders are not selling much or are not very sensitive to profit-taking and price, it's the short-term demanders who are determining the price. So, at this time, we need to look at STH MVRV. But when looking for the bottom of a bear market, we do the opposite and don't look at STH MVRV, because it's not very meaningful for price reference. Instead, we look at the MVRV of long-term holders. If the MVRV of long-term holders is already less than 1, it means that long-term holders, those who have held for more than 155 days, are all diamond hands in the market. Even if they are whales or some investment teams, their research capabilities far exceed those of us retail investors. If even they are in an average loss state, do you think this is not a market bottom? I think it's highly probable.
In addition, because I used to mine for several years, my understanding of mining costs may be different from those who have not mined before. I can deeply feel that when the price drops to a certain level, miners are in agony every day. So, I generally use a relatively meaningful price model to calculate (mining costs) from the perspective of mining costs. Another method is to estimate the 14-day average block generation time interval and the deviation from the target using the pulse index of mining. Why 14 days? Because 14 days is a difficulty adjustment period. For example, if the target for BTC is to produce a block every 600 seconds, you can calculate the average block generation rate over 14 days and compare it to 600 seconds. If it's a positive value, it means the actual block generation time is slower than the target time, and the reason for the delay is that some miners have shut down and turned off their computing power. I've done this before; the amount of BTC mined every day wasn't enough to cover my electricity bill. Why would I keep the machine running to mine? It would be better to turn off the machine and buy from the secondary market. So, in this situation, the machine will be shut down, leading to longer block generation times. Of course, there are many more examples, but I'm just giving one.
In summary, using these five approaches to find severely undervalued market ranges, if friends are interested, you can try it in the next cycle, maybe in 2025 or 2026, using this method to find and test. I believe that the method I just mentioned has a probability of 80-90% to help you buy at a relatively low point in a bear market, although it's only a relative bottom.
Now, talking about how to escape the peak. Actually, in 2021, when the price was in the range of 52,000 to 48,000, my escape from the peak was mostly based on luck. In the previous cycle, I made money based on luck, so this time I didn't want to spend it all based on luck. Because I had mined a lot of chips, my cost was relatively low, so whether I sold at 60,000, 55,000, or 48,000, it didn't have much impact on me. I didn't feel much psychological pressure, and I didn't have a high requirement to absolutely chase the peak. It was just about making a little less. At that time, I also read a lot of news, both domestic and foreign, and then looked at other people's analyses, and even joined some paid groups of teachers at the time. I didn't speak in the group, I just watched the teacher talk about the point every day, and when it was almost there, I started selling. It was just hearsay. But I definitely won't do that this time, because I will completely base my decisions on my own set of data analysis theories.
FC
Speaking of escaping the peak, you definitely think that we are currently in the pullback stage of this cycle, because the bull market hasn't ended. So, how can we use three indicators to judge the high probability of being in the peak range?
Murphy
When escaping the peak, relying solely on three indicators is far from enough, because these three indicators cannot cover all the attributes of the data. Our data attributes are diverse, including emotional, capital, supply and demand, and mystical aspects, and you can't cover all of this with just three indicators. If I had to summarize, I think there are three that are relatively important to me.
The first is the MVRV I mentioned earlier. Some free indicators you see are just a general MVRV, but there are two subcategories within it - STH, which is the MVRV of short-term holders, and LTH, which is the MVRV of long-term holders. These two are used differently. But if you must look at the free MVRV, the simplest method is that when MVRV is above 3, you can sell. You don't have to sell everything; you need to have a well-defined trading strategy. When it reaches 3, 3.2, 3.5, and then drops back to 3.2, 3.1, 3, you should sell everything. In terms of the inverted smile curve, it's an escape from the peak. When we build positions, we want to create a smile curve that goes up, and when escaping the peak, it's an inverted curve, this MVRV. I explained the logic of MVRV earlier; it measures the issue of unrealized profits, the extent of the bubble. If the bubble reaches 300%, and you still don't sell, how much higher do you think it can go? According to my observations, except for the first bull market cycle, when MVRV reached 5 or 6, I forget, that was a very extreme time, because at that time, BTC's market value was small, and there were fewer participants, mostly geeks playing, not capital. Currently, at the trillion-dollar level, MVRV is basically between 2-3, like it is now at 2. Many friends say they expect it to reach 150,000 this time, but now they're hesitant even at 70,000, because it's too big, it has reached the level of an aircraft carrier, and this is an important reference.
The second thing to consider is the URPD data, which is the structural data of on-chain chips. I want to emphasize the controversy surrounding URPD data. For example, both I and Ni Da often use this data for analysis, but many friends feel that this data is useless because it cannot determine long or short positions, whether it will fall below, or whether it will break through. This data itself is not used to determine long or short positions. What is it used for? Let me give an example. When we look at the exchange's candlestick data, you can only see the volume at a certain price level or on a certain day, but you cannot see how many chips are in a certain price range. However, URPD can show us this. It can tell us how many chips have changed hands in each price range. We consider this turnover range as a consensus range. The more chips accumulate, the stronger the consensus becomes. After the consensus strengthens, it will produce an effect. Ni Da believes it's a support effect, but in my opinion, it's a resistance and stickiness effect. I won't go into resistance, but when trying to break through from below, there will be selling pressure, making it not easy for the price to break through this range. When falling through from above, there will be stickiness, which doesn't mean it can't fall through; it can, for example, it fell below 50,000 to 54,000, but if you look at the 63,000-64,000 and 68,000-69,000 ranges, the chips did not decrease significantly. This means that at over 50,000, it did not fully form a consensus in terms of turnover to digest the high-level chips above 64,000. If it had been digested, the previous stickiness would have disappeared, and a new chip range would have started. But if it hasn't been digested, what does it mean? It means that it won't stay there for long and will soon return. If you don't believe in this data, you can read my tweet titled "Viewing the Evolution of BTC Cycles from God's Perspective." I compiled the URPD data changes for each month of the previous cycle, and it does have such a condition.
The third data to pay attention to is the supply and demand relationship. Throughout the BTC cycle, its supply side is actually LTH, long-term holders, and the demand side is STH, short-term holders. From the bear market to the bull market, long-term holders continuously distribute their chips to short-term holders. During the period from the peak of the bull market to the bear market, short-term holders continuously sell off and return their chips. Each cycle on the chain can clearly show this kind of cycle, repeating itself. So, when we haven't seen LTH making a profit-taking or exiting behavior in this cycle, I believe it's not yet a bull market. But conversely, when long-term holders have started selling, or have sold almost all, as was very obvious during the 63,000-69,000 period, LTH had already sold almost all. From the drop from 63,000, it had already started selling, and the same was true in the previous cycle, from the bottom to the top, LTH also started selling. So, when the supply and demand relationship changes, we need to pay attention.
FC
Actually, I particularly like the third indicator because it can show the difference in trading styles between retail investors and institutional investors from a data perspective. What does "chasing highs and killing lows" mean? I think this is particularly evident. I remember it very well, at 20,000, some people within our group also said it would drop to 12,000. I think there's a particularly dangerous mindset called the "magical number." You must not be fixated on numbers, for example, some people are fixated on round numbers. I think this is a big problem, and everyone must pay attention to it.
Murphy
Yes.
FC
I actually want to ask a question, how did you learn? What was your learning approach and path?
Murphy
First, you need to gather all the scattered information, including some foreign on-chain data analysts. There may not be many in China who specialize in this area, such as me and Pionex. But there are many on-chain data analysts abroad. At the beginning, I actually used some of their analyses, but they don't explain things in a systematic way. They talk about one thing today and another thing tomorrow. I would take notes on these data and then analyze the phenomena they talked about, understand their logic, create algorithms based on theirs, and make adjustments if I felt something was not right. It's a process of gradually accumulating from less to more. Later, I found that my brain couldn't handle all the information, and I couldn't remember so many complex indicators, some for tops, some for bottoms, some for trend cycles. What to do? I created an Excel sheet, listing the names, sources, algorithms, and purposes of the data, and added a column for the category of the indicator. This is an emotional indicator, this is for trend cycles, this is for tops, this is for bottoms. Later, it's very simple; my spreadsheet should have over 300 rows, which means there are over 300 data indicators as reference. Excel has filtering functions, so now we can filter out the bottom indicators, leaving the top and trend or emotional indicators. Of course, we don't need to look at the top now because we're not there yet. We're basically looking at how long the trend cycle will last and where we are now. It may still take a while, which is a question many people are concerned about. The second thing is the emotional feedback data. What kind of state is it reflecting? Are we really entering a period of transitioning from a bull market to a bear market? Has the bull market already ended? As you accumulate, it's actually like this. If you want to learn this data, it's also very simple. Just write down all the indicators I share on Twitter, put them in Excel, list them, and then go through them one by one, mark them, and classify them. Over time, it will definitely be effective.
FC
I think this is quite practical. First, find a teacher, then categorize. Will you backtest? For example, to see if a certain indicator is accurate, will you combine it with candlesticks or cycles to see if the indicator is correct?
Murphy
Backtesting is a must because you can't guarantee that you're 100% correct every time you make a judgment. Some of the conclusions I drew in the past were also incorrect. For example, after the ETF was approved, there was a pullback. I wrote a tweet called "Will We Still See BTC Starting with 3," and it did drop to the 30,000s, but at the time, I felt that 30,000 was not enough because I believed that in every bull market cycle, there must be a deep pullback, at least 30%. But it hadn't happened before; it was just these shallow, unsatisfying 10% or so pullbacks. So, I thought this was a great opportunity, a one-time pullback in place, and the next time would be a high magnitude, but I was wrong at the time. Some friends may have seen my tweet at the time and missed a great opportunity to get on board.
FC
Why do you think you were wrong at the time? Was it because the data already told you that it might not be like that, but you were fixated on it?
Murphy
No, it's because the data at the time was very difficult to judge. When is on-chain data not useful? It's when you're trying to determine whether it's a long or short position in a very short period, or whether it can drop to a certain level, or whether it can rise to a certain level. The precision of the indicators is not very good, but if you use this data to look at trends and cycles, it's very effective. So, you can't use a cannon to kill a mosquito, that's the general idea.
FC
That's great, you've already answered our next question, which is when will your trading strategy become ineffective? I understand it should be on a daily, monthly, or weekly basis, which is more likely to become ineffective.
Murphy
Yes, judging short-term ups and downs is ineffective, using this for contracts is also ineffective, or predicting the future top and bottom prices, all of these are ineffective.
FC
Understood. Why can't on-chain data analysis be used for altcoins?
Murphy
It's like this, BTC's characteristic is its UTXO structure, which stands for Unspent Transaction Output. This structure has a feature, which is its timestamp function. You can use the timestamp when the UTXO generates unspent transaction outputs and record the USD value of BTC at that time. With these two pieces of data, many variations can be derived, such as categorizing holders as short-term holders if they have held for less than 155 days, and long-term holders if they have held for more than 155 days. Various analyses, including URPD, are derived from this. However, other altcoins do not have UTXO, and ETH also does not have this structure, which means that many on-chain data cannot be analyzed. Therefore, indicators cannot be used to analyze small coins.
FC
Understood. Where do you think the current cycle is at, and approximately when or in what range do you think the top interval might occur? Again, I emphasize that this is not investment advice.
Murphy
Regarding the current stage of the cycle, in my opinion, the transition of each bull-bear cycle is marked by two large-scale acquisitions by short-term holders, which is related to their emotions. The first large-scale acquisition often occurs before the halving, when everyone expects the halving to happen soon after enduring a long period of bear market torture, and many high-level chips were sold at the bottom of the bear market. Some short-term holders built positions at the bottom, triggering a significant emotional response. In this cycle, the first acquisition occurred when BTC exceeded 70,000. So, this cycle is actually different from the previous two cycles, mainly because the ETF was approved, changing the rhythm of the entire cycle. However, from the overall data perspective, the first acquisition is not considered significant, so I believe there will be another large-scale acquisition by STH, which will coincide with the distribution by LTH, long-term holders. These two points in time will mark the relative top interval of the next bull market cycle.
But in this top interval, it may be a double top or even a triple top. Will the third top be accompanied by a third acquisition and distribution? I don't know, but there should be at least two. In terms of the time period, as I mentioned earlier, it's not possible to completely predict the timing of this cycle from on-chain data. However, coincidentally, from the three-line integrated indicator I shared a few days ago, it seems that the high MVRV values usually occur in March to April of the following year. In that time period, for all previous cycles, the MVRV value was higher than 3, and in the first cycle, it was even higher than 5. In the second and third cycles, it was above 3, sometimes reaching 3.5 or 3.2. As I mentioned before, if this situation occurs, we need to consider it as a signal to exit the market. So, roughly speaking, it should be around March or April next year, but this is not certain.
FC
Got it. I think everyone can pay attention to Murphy because he really shares unconditionally, and you can see when he suggests exiting the market. We will include two questions in the written version: the first is about the three people Murphy usually follows, and the second is about which traders Murphy likes and which content is better. Everyone can follow Murphy and my Twitter.
I think that's about it for today. Actually, there are four dimensions we didn't delve into: emotions, funds, supply and demand, and metaphysics. Expanding on these four dimensions would take a long time, so I think we should find an opportunity to chat again when we think we're close to the peak of the bull market.
Murphy
Actually, I have a plan. In the future, when I think the top interval has already appeared or is about to appear, I want to write a series of tweets, such as "Judging the Top Interval of the Bull Market," and we will verify it step by step. We will decide to exit at this position and then at the next position, and when we look back during the bear market, we will see if our exit position was relatively high.
FC
Interesting. I think your and Ni Da's tweets are a record of personal growth, a way to practice your own trading logic. It's really interesting. I'll grow and learn with you guys. That's it for our conversation today.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
