Moderator: Alex Mint Ventures Research Partner
Guest: Colin, free trader, on-chain data researcher
Recording time: 2025.2.15
Hello everyone, welcome to WEB3 Mint To Be initiated by Mint Ventures. Here, we continue to ask questions and think deeply, clarify facts, explore reality, and find consensus in the WEB3 world. We clarify the logic behind hot topics, provide insights that penetrate the events themselves, and introduce multiple perspectives.
Disclaimer: The content discussed in this podcast does not represent the views of the guests’ institutions, and the projects mentioned do not constitute any investment advice.
Alex: This episode is a little special because we have discussed many topics about specific tracks or projects before, and also exchanged some cyclical narratives, such as memes. But today we are going to discuss on-chain data analysis, especially on-chain data analysis of BTC. We will take a close look at its working principles, key indicators, and learn its methodology. In today's program, we will mention many concepts about indicators, and list these concepts at the beginning of the text version for your convenience.
Some data metrics and concepts mentioned in this podcast:
Glassnode: A commonly used on-chain data analysis platform that requires payment.
Realized Price: Calculated based on the price weighted at the time of Bitcoin’s last on-chain movement. It reflects the historical on-chain cost of Bitcoin and is suitable for assessing the overall profit/loss status of the market.
URPD: Realized price distribution. Used to observe the price distribution of BTC chips.
RUP (Relative unrealized profit): Relative unrealized profit. It is used to measure the ratio of the unrealized profits of all holders in the Bitcoin market to the total market value.
Cointime True Market Mean Price: An on-chain average price indicator based on the Cointime Economics system. It aims to more accurately evaluate the long-term value of BTC by introducing Bitcoin's "time weight". Compared with BTC's current market price and realized market price (Realized Price), the True Market Mean Price under the Cointime system also comprehensively considers the impact of time and is suitable for BTC prices in large cycles.
Shiller ECY: A valuation indicator proposed by Nobel Prize winner in Economics Robert Shiller, it is used to evaluate the long-term return potential of the stock market and measure the attractiveness of stocks relative to other assets. It is improved from the Shiller Price-to-Earnings Ratio (CAPE) and mainly considers the impact of the interest rate environment.
An opportunity to learn on-chain data analysis
Alex: Our guest today is Colin, a free trader and on-chain data researcher. Please let Colin say hello to our listeners first.
Colin: Hello everyone, first of all, thank you Alex for the invitation. I was a little surprised when I received this invitation, because I am an unknown retail investor, and I don’t have any special titles. I just quietly do my own trading. My name is Colin. I run an account on Twitter called Mr. Berg . I usually share some teaching on chain data, analysis of the current market conditions, and some trading concepts. I have three positionings for myself: the first is an event-driven trader. I usually think about event-driven trading strategies; the second is an analyst of chain data, which is also the content I usually share on Twitter; the third is more conservative. I call myself an index investor. I will choose to allocate part of my funds to the US stock market. Through this part of the funds to invest in Beta, I will reduce the overall volatility of my asset curve, while maintaining a certain defensiveness of the overall position. The above is roughly how I position myself.
Alex: Thank you Colin for introducing yourself. I invited Colin to participate in the program because I saw his on-chain data analysis of Bitcoin on Twitter, which was very inspiring. This is a topic that we have rarely talked about before, and it is also a relatively lacking part in my own section. I read the series of articles he wrote and felt that the logic was clear and meaningful, so I invited him. I would like to remind everyone that today, both my and the guests’ opinions are highly subjective in the program, and the information and opinions may change in the future. Different people may have different interpretations of the same data and indicators. This content is not intended as any investment advice. This program will mention some data analysis platforms, which are only for personal sharing and examples, not as commercial recommendations. This program has not received commercial sponsorship from any platform. Let's get to the point and talk about on-chain data analysis of crypto assets. I just mentioned that Colin is a trader, so under what circumstances did you start to contact and learn on-chain data analysis of crypto assets?
Colin: I think this question should be divided into two parts. First of all, I think that no matter who is around me, as long as they want to enter or have entered the financial market, including myself, the main goal should be to make money and use the profits to improve their quality of life. So my philosophy has always been consistent, that is, I will learn whatever can help me make money. In this way, I can improve the expected value of my overall trading system. In simple terms, I will learn whatever can make money. The second part is that I was exposed to on-chain data by accident at the beginning. About six or seven years ago, I didn’t understand it at all. I looked at this and that. When exploring various fields, I saw very interesting research theories and wanted to learn them. At that time, I accidentally saw that Bitcoin had a so-called on-chain data analysis field, so I started to learn and study it. In the later stage of learning, I will combine the knowledge I have learned in other fields, mainly the part of quantitative trading development, and combine it with on-chain data, and then develop some trading models, and finally integrate these models into my own trading system.
Alex: So, since you officially started to get involved in on-chain data analysis, how many years have you been studying and researching it systematically?
Colin: I think this is hard to define. In fact, I have never really studied it systematically. Because from the past to now, I have encountered a problem, that is, I have not seen any systematic teaching at all. When I first saw this field, it was probably several years ago. I discovered it at that time, but I didn’t study it in depth. I just read two or three articles and knew about it. After a while, I came back and saw some more in-depth content. At that time, I was focusing on studying other things. I came back here and saw that this was quite interesting, so I continued to study it. There was no time for systematic learning. I just pieced it together.
Alex: I see. How long did it take you to learn about on-chain data and apply it to your actual investment practice?
Colin: This boundary is difficult to define, but I think it is close to two rounds of Bitcoin cycles... It can't be considered two rounds, it depends on whether you start from the bull market or the bear market. I started to get in touch with it around 2020 or 2019, but there was no practical application at that time because I didn't dare to. At that time, I was not very familiar with this thing, but I had already started to learn.
The value and principles of on-chain data analysis
Alex: I see. We will talk about many specific concepts about on-chain data analysis, including some indexes. What on-chain data observation platforms do you usually use?
Colin: I mainly use one website now, which is Glassnode. Let me briefly explain that it is paid. There are two paid levels. One is the professional version which is more expensive. I remember it cost more than 800 US dollars a month. I forgot the second one, which is about 30 to 40 U a month. It also has a free version, but the free version actually provides very little information. Of course, there are many other websites besides Glassnode. I finally chose it because this website was the most suitable for me when I was screening and researching.
Alex: I see. After reading a lot of information from Colin, I also registered for Glassnode and became their paid member. I feel that their data is indeed very rich and timely. So let's talk about the second question. You just mentioned that you are a trader, and you value its help in investment practice. So what is the core value of on-chain data analysis in your investment? What is the principle behind it? Please introduce it to us.
Colin: Okay. First, let me talk about the value and principle of on-chain data analysis. I plan to talk about these two together because they are actually quite simple. Our traditional financial markets, whether trading stocks, futures, bond options, or even real estate, or some commodities, Bitcoin has a fundamental difference from them, which is that it uses blockchain technology. The most important and most often mentioned value of this technology is its transparency. All of this Bitcoin transfer information is open and transparent, so you can directly see on the chain that, for example, 300 Bitcoins are transferred from one address to another, which can be checked on the blockchain browser. Although I have no way of knowing who is behind this string of addresses, it is not important because no single individual can actually affect the price trend and trend of the entire Bitcoin. So normally, when we study on-chain data, we look at the overall market, its trends, and the consensus and behavior of the group. Even if I don’t know who is behind this or that address, I can analyze the flow of their chips by aggregating all the addresses to see whether they have taken profits or stopped losses, their profit and loss situation, at which price they prefer to buy a large amount of Bitcoin or at which price they don’t like to buy Bitcoin. These data are actually visible. I think this is the greatest value of Bitcoin chain data analysis compared to other financial markets, because other markets cannot do this.
Alex: This is indeed very important. When we invest in cryptocurrencies, we need to analyze fundamentals just like we do when we look at stocks or other products. As you just said, the on-chain data is transparent and everyone can observe it. If other professional investors look at the on-chain data and you don't, then you are missing an important weapon in your investment.
Difficulties in on-chain data analysis
Alex: When you are actually doing on-chain data analysis, what do you think are the main difficulties and challenges?
Colin: I think this question is very well asked, and I plan to answer it in two parts. First of all, the first part is relatively easy to solve. There is a more difficult point in learning, which is the basic knowledge. For most people, including me at that time, as I mentioned before, it is difficult to find a truly systematic teaching. Of course, I did not ask offline if there are any paid courses of this kind, but if there are, I would not dare to buy them, because I have been trading myself until now, and I don’t really pay to buy some courses. I have not come into contact with any systematic teaching courses, so in fact, all the content must be explored and explored by myself. There are many types of on-chain data. In the process of research, my own idea is to figure out the calculation method and principle behind each indicator I have seen. This is actually a very time-consuming process, because when you only see a certain indicator, it will give you a calculation formula. My idea is to figure out what is behind this calculation formula and why it is designed in this way. After I figure out these indicators, the second thing I have to do is called screening. If you have experience in developing quantitative strategies or have studied indicators, you will know that the correlation between many indicators is very high. A high correlation will cause a problem, that is, it is easy to generate noise in your interpretation, or you will over-interpret. For example, suppose I have a system for escaping the top today. This system may have 10 signals from 1 to 10. If the correlation between 1 to 4 is too high, it will cause a problem. For example, if the price of Bitcoin has a certain behavior or change today, it may directly make the lights from 1 to 4 light up at the same time, which is actually very troublesome. Because if their correlation is too high, this is an inevitable phenomenon. If 4 out of 10 lights are on today, you say it is very dangerous, but in fact it is not reasonable, because they will light up anyway. If you don't cut them according to the correlation, this phenomenon is very easy to happen. After I have studied the principles of each indicator and data, I can actually know whether their correlation is high or not by looking at the calculation formula directly, and I cut it according to the correlation. For example, if these 5 are highly correlated, I will cut and screen them slightly, and finally select one or two.
The first part is actually easy to solve and is not the main difficulty. The second part is the real challenge, which is about the on-chain data. How do you prove your point of view to the people around you or to yourself? I may have to give a more vulgar example here, but it is easy to understand. I have written in a tweet before that in fact, the quantitative field will tell you that trading is not easy to stick to the old ways. I have given an example before. Suppose there is a very strange trading strategy today. Its entry standard is that if my dog at home barks twice and it rains outside, I will enter the market and go long. As a result, I backtested this strategy 1,000 times and found that the winning rate was 95%, which far beat the market. Does anyone dare to use this strategy? It is actually quite strange. If the dog barks for no reason and it rains outside, you can go long, and the winning rate is still so high. There is actually a term called survivor bias. If you can't give it any logical support today, even if the number of samples is sufficient, this strategy cannot be used. Some people will argue that it has been backtested 1,000 times, and the winning rate is 95%. The backtest results support that this strategy can be used. I just mentioned the so-called survivor bias. Simply put, if I toss a coin 10 times, the probability of getting heads 10 times is actually 1/1024. In other words, on average, when 1,024 people do this, one person will succeed. The situation of throwing heads 4 times in a row is actually the so-called survivor. The other 1,023 people failed when doing this, but we don’t actually see it. We always see those successful cases. Back to Alex’s question just now, that is, where is the so-called main difficulty. Because we mainly look at large-scale consensus and trends, looking back at the history of Bitcoin, the three most obvious cycle tops are the two tops in 2013, 2017 and 2021. This is only 4 samples, which is definitely not enough. Since the number of samples is not enough, if we go back to the old ways today and see where a certain indicator has been in 2013, where a certain indicator has been in 2017, so we have to go there this year, this is unreasonable. Because the number of samples is completely insufficient, if we don't give it logic to do research at this time, your theory is very easy to make mistakes. One of the main problems is that in the face of such a small number of historical samples, I must use the deductive method instead of simply using the inductive method to study. After I finish my research, I draw a conclusion based on the deductive method, and I need time to prove whether my view is right or wrong. If it is right, it means that my previous deductive reasoning process may be reasonable. If it is wrong, then I need to continue to correct the previous deductive logic. But if we just rely on induction today, in fact, most retail investors like to do this the most, thinking that the previous trend is very similar to the current trend, so there should be a surge or plunge in the future, which is actually unreasonable. Back to the first sentence I said at the beginning, I think the biggest difficulty is that I have to prove to others or myself that my inference is correct, so I have to correct my logic and assumptions all the time, and then check whether there are any flaws. Because Bitcoin is too young, there will always be a problem of insufficient samples in the on-chain data analysis. At this time, you actually have to use a simple deductive method in research, and use a logical way to infer it, and then wait for time to prove your judgment. This is the biggest difficulty I have encountered so far.
Key on-chain metrics to watch
Alex: I see. I think it is very inspiring. The question I asked you just now was also some confusion when I started to look at various indicators on Glassnode. It has so many indicators. Which indicator should I use as my trading reference? Because many indicators have various calculation logics. Later, I tend to choose the logic of those indicators, which is quite similar to the logic you just mentioned. That is, first of all, I have to look at the calculation logic behind this indicator, and I have to think that this logic makes sense, rather than backtesting and pulling it out and saying that it seems that this indicator is very accurate, and then use this accurate indicator to predict the future. As you said, the reference of the deductive method needs to be greater before it can be used as the main indicator we adopt. So after the experience you just talked about, in your current daily analysis of Bitcoin, which on-chain indicators have you been paying attention to for a long time or do you think are more important?
Colin: I have actually talked about this question before. I will try to filter based on correlation. I usually look at a lot of on-chain data indicators, so today I will introduce them from different dimensions, that is, try to divide them into three levels from the part with low correlation.
The first indicator that I will pay attention to for a long time and focus on must be the URPD indicator. It is a chart, which is presented in a row of bar charts. The horizontal axis is the price of Bitcoin and the vertical axis is the number of Bitcoins. Suppose we see a very high and large column at the position of 90,000 today, then we will know that a very large number of Bitcoins are built at this position, that is, the cost of their purchase. The bar chart will show how many Bitcoins they bought at this price. So in fact, based on this matter, we can see at a glance that if there are a lot of accumulations above 100,000, then we can know that many people buy above 100,000. There are two main points of observation in this URPD chart. The first is the simplest chip structure. Suppose today I see that the current market situation is around 87,000, and a very large number of chips have been accumulated above 87,000. According to last week's data, it should be 4.4 million. Then we know that there is a very large turnover in this range, or someone has bought here. Since someone has bought, it is very likely that a certain consensus has been formed. In this range of large accumulation, it is easy to form an attractive effect on the price, that is, the price is likely to fluctuate in this range, and it is easy to repair after a period of time if it falls, and then rise back. If it rises, the chips below have all become floating profits, then they are easy to sell, sell, sell, do short-term transactions, and then sell the price back. So it is easy to fluctuate in this range. This is the first observation point. The second observation point is that we can observe the process of Bitcoin distribution through URPD. The so-called distribution is that in the early bear market, those chips that bought Bitcoin at a low price, and then they sold the cheap chips in their hands upwards, so I define this process as distribution. Suppose today there are 300,000 more chips at the price of 100,000, and the chips with a cost of 20,000, assuming it is 20,000, just reduced by 300,000, then we can actually see that people with a cost of 20,000 sold 300,000 today, and their average selling price is about 100,000. We can see whether those low-cost chips usually have some drastic changes. Of course, the current price is 100,000 or 90,000, so if they change dramatically, it must be a decrease, not an increase, because the current price range is more than 90,000, not more than 20,000, so it will only decrease, not increase. So we can observe the distribution rate based on this matter, which is roughly what I mean. This is the first indicator that I will pay attention to for a long time.
The second indicator I want to introduce is called RUP, which means Relative Unprofitable Status in Chinese. This indicator actually has only one purpose, which is to help us measure the profitability of the overall market, that is, the profitability of the entire market corresponding to the current price of Bitcoin. For example, how much did you earn, or not much, or a lot, roughly this concept. The principle of this indicator is actually very simple, because through the so-called transparent mechanism of blockchain, we can track the purchase price of most chips. We can compare the purchase price of these chips with the current price. Suppose he bought it at 50,000, and the current price is 100,000, we will know that this Bitcoin is currently profitable, so we can calculate how much money it has earned. For example, if there are 10 Bitcoins bought at 50,000, and now it is 100,000, 1 will earn 50,000, and 10 will earn 500,000. We add up all these floating profits and losses, and then standardize this number according to the current market value, then we can get a number between 0 and 1. The range between 0 and 1 is easy to observe. If today's RUP is very high, such as 0.7, 0.68, or 0.75, we know that the overall profitability of the market is very high, which may make more people want to take profits. Therefore, a high RUP is usually regarded as a relative warning.
The third dimension I want to talk about is a fair valuation model of the market. There are actually many different Bitcoin valuation models on the market, and each model actually uses a different method to evaluate the fair value of Bitcoin. The so-called fair value is actually how much a Bitcoin is worth. After looking at so many models, I think the Cointime Price model is the most proven. I have not seen the Chinese translation of this term anywhere else. Simply put, we often hear a name called Cathie Wood, her ARK Invest, and the chain data website, which is the Glassnode I mentioned just now. This concept is mentioned in a document produced by the two parties in cooperation. The biggest feature of this model is that it introduces the concept of time weighting and then calculates the fair value of Bitcoin. The calculated number has two main uses. The first is very simple, which is to buy at the bottom. Suppose today in the bear market, it falls and falls, and finally falls below the valuation given by Cointime Price. As I said just now, this number is actually how much a Bitcoin should be worth. If it falls below this position today, it is equivalent to buying at a very cost-effective position. According to historical backtesting and its logic, we can actually see that whenever the price falls below the Cointime Price, it is actually a very good position to buy at the bottom. The second application is to escape the top. We can monitor the current price and the Cointime Price to see how far it is away. If it deviates too much from the Coin Time Price, we can evaluate whether the market may be close to the top if the deviation is too large. The above three dimensions are the chip structure, profit status and fair valuation model, which are the three indicators and aspects I want to share.
How to view data conflicts
Alex: Okay, I have made it very clear just now. Many users may ask a question. The three indicators you just listed may represent different aspects, which is consistent with what you just said that the correlation between them is not that high, so they can be put together as a reference indicator. Then suppose that such indicators have divergent situations in actual application. For example, indicator one may feel that it is currently in a distribution situation, while indicators two and three may show that the current distance to the top does not seem to be that high from a cycle perspective. In this case, how would you deal with the conflicting data?
Colin: I think this is not just in the field of on-chain data analysis, but also in other fields such as technical analysis or macroeconomics, where there may be so-called fighting. In the on-chain field, my personal approach is very simple. I will give different weights to different levels. What I value most is actually the chip structure, that is, the progress of distribution. Because in fact, in terms of profitability, it also helps me observe the low-cost chips in the market. During the bear market, for example, the Bitcoin chips bought at 15,000 or 16,000, have they been distributed? There is a very special phenomenon that in every cycle of Bitcoin over the years, there are actually two very obvious large-scale distributions. For example, in 2024, the most obvious case is from March to April last year. In fact, in terms of profitability, you can definitely see large-scale distributions at that time. But if I only see large-scale distributions today, then my next question is to think about whether they have been distributed? All the criteria for judgment start from this question. If they have a large-scale distribution, but have not yet completed it, then I can tell myself with peace of mind that the bull market has not ended. For example, from March to April last year, Bitcoin rushed to more than 70,000. I was actually quite excited because the bull market finally came and set a new high. As a result, it began to fluctuate for more than half a year. At that time, I could not draw the conclusion that the bottom had been reached by observing these data. At best, it was the first distribution. And a lot of data is also the same. For example, I have published some mid-term analysis and chip structure analysis before. At that time, according to the average cost of short-term holders, his situation was actually different from the end of the real bull market. So I was actually very at ease at that time. Then you say that the data is conflicting, and now he says that he has distributed, so should I flee the top? In fact, no, because the main issue is still the one I just talked about: whether the distribution has ended. Using this issue as the standard for screening each indicator and as the basis for judgment, it is actually very easy to draw this conclusion, that is, even if the distribution has occurred and is still large-scale, I just need to judge whether it has ended. Using this as a criterion can effectively deal with the so-called data conflict problem.
Alex: Let's make a scenario. For example, let's look at URPD. Assume that this indicator has already had two distributions, which is more like what you just said, one in March and April last year, and then there was a peak distribution from December to January at the end of the year. Assuming that it has such a distribution, but the other two valuation indicators may not be so high, when this happens, you just said that you will give them different weights. Then, will you reduce part of the position according to the proportion of the weight, or will you consider the three indicators in a unified way and not adjust the position according to the weight, but make one or two important decisions at critical times?
Colin: My own approach is the former, because in fact no one can know whether it is the real top now, and no one can escape at the highest position. If there is, it would be too great, and I would definitely want to know it. My personal interpretation of the top is that it is a slow process. Although you feel it is very fast if you look at the daily chart, in fact, if you are in the present, for example, if you are at 69,000, the top of the previous cycle, you will not feel that it is the top now. We can only make a judgment based on the data and say that it is possible that the conditions for the formation of the top are now in place. So based on this premise, I will actually take a segmented position. For example, when I think the conditions for the top have gradually matured, once I see a certain indicator giving me a warning during this period, such as a RUP divergence I shared on Twitter before, I will do a corresponding reduction in positions. Of course, the extent of this reduction in positions must be determined in advance from the beginning. It is impossible to say that there is a divergence now and I don’t know how much to reduce. It won’t be like this. I will first make a rough plan, for example, I divide my position into 4 parts, and then once a certain type of warning signal appears, I will reduce one of them first, and then reduce the second warning signal. At the same time, I will plan that the last part of the funds must be exited no matter what. For example, if the bear market is definitely over, but other warning signals have not yet appeared, we need to formulate an extreme, final escape strategy to do the screening.
Alex: I understand. We will gradually exit the market and reduce our positions based on different warning signals.
Colin: Yes.
Judgment and basis of BTC's position in this cycle
Alex: I see. I have been following your Twitter account recently. You also practice your trading according to the indicators just mentioned, including the concepts behind these indicators. Now let's look at Bitcoin. It has been fluctuating in the range of 91,000 to 109,000 for almost three months. There are quite a lot of differences in the market about this price range. Unlike in December and January, everyone thinks that this bull market is far from over and will reach 150,000, 200,000 or even 300,000. There are many positive views. There are big differences in the market at present. Some people think that the top of BTC in this round is around 100,000, but some people think that BTC has not reached its peak in this cycle, and there will still be a main uptrend in 2025. So based on your current comprehensive judgment, what is your opinion? Where is BTC in this round of our big cycle? And what are the data sources that support your judgment?
Colin: Before answering this question, I may need to give a precautionary shot. I am actually very bearish on 2025. I think BTC is currently in a state where the top is formed. In fact, I know that many people, including some participants around me, did not have good returns during the so-called special bull market in 2024, because the overall market performance in 2024 is different from every previous cycle. The most obvious point is that there is no altcoin season. This hurts many people, including some non-professional traders around me. They also came to participate in this market. In fact, they suffered a lot of losses on altcoins. Why is this the case? Let's take a look back at 2024. There was an altcoin market at the beginning of the year, and the second one was in November last year, when Trump was elected as the US president. Compared with our previous cycles, these two altcoin markets actually have a big and obvious point, that is, their sustainability is actually not very good. Even in the market in November and December last year, altcoins did not rise across the board at all. It was a very obvious sector rotation. At that time, there was a Defi sector, and after the rise, it switched to old coins, such as XRP, and then Litecoin, etc. The sector rotation was very obvious. From this incident, we can see that this round of bull market in 2024, if everyone thinks it is a bull market, this round of cycle is actually very different from the previous ones. There is also a theory that there must be a so-called cottage season before the end of the bull market. In fact, I personally think that you can't say that the bull market will end only when the cottage season appears. This is obviously not strongly related. We can't use this as a judgment on whether the bull market is over. As mentioned earlier, there is a shortcoming in on-chain data analysis, that is, the number of samples is never enough. If we simply use historical conditions to extrapolate today's market, it is actually a practice of carving a boat to find a sword, which is not very good. If you want to carve a boat to find a sword, the tops of 13, 17, and 21 should appear around the end of the year, according to the time.
I personally think that the conditions for the formation of the so-called top are already in place. The reasons are very complicated, and I use a lot of indicators and data to make judgments. Let me briefly talk about a few of the more core ones. The first one is the chip structure we just mentioned, which is the URPD chart. We can see one thing, in 2022 and 2023, those low-cost chips accumulated, at that time they bought a lot of BTC at a low price, and so far a lot of chips have been distributed. To put it bluntly, they have sold them, they are not playing anymore. Some listeners may have a question, that is, what does it have to do with me if they sell? There is a concept that I may need to explain to you. At the end of each round of bull market, almost every time it is because the distribution of those low-cost chips has ended, and then the bull market ends. There is a relatively unintuitive point here. It is not because they smashed the market that the bull market ended, but because the price has been rising, they sold all the way, and they sold all the way, and then the price stopped, and the bull market will end. This is not just my head saying that it must be like this, there is a logic in it. Assume that every BTC chip involved in the market today is a high-cost chip, for example, those bought at more than 90,000, and those bought at 50,000, 20,000, and 30,000 have already run away. At this time, as long as the price does not show a very obvious or strong main rising wave, even if it is simply a so-called wide range of fluctuations, such as the fluctuation range between 70,000 and 50,000 last year, or the current fluctuation range of about 90,000 to 109,000, it will put these high-cost chips under great pressure to hold positions. High pressure on holding positions will lead to a problem. The price is now about 95,000 or 96,000. Suppose it falls to 89,000 today, which is actually less than 10%, but these chips are under great pressure. Many of them are even short-term traders. Once the pressure is high, they may choose to sell, and then the selling will lead to a further drop in prices. If the price falls, other high-cost chips will not be able to withstand the pressure, and they will sell again, which will cause a chain reaction. This is what I think I can see from the URPD chart, that is, many low-cost chips have been distributed.
The second indicator I just mentioned is called RUP, which is an indicator used to measure the profitability of the market. If you are interested in this indicator, you can check it out. It is very interesting. If you put its line together with the price line, their correlation is very, very high, and they almost move together. This is actually a very reasonable thing, because the higher the price, the higher the holding cost profitability will be, and the shapes of the two lines will be almost exactly the same. So the higher the price, the higher the RUP will be; the lower the price, the lower the RUP will be. This is very simple. But once the RUP shows a so-called divergence, it actually means that the market situation has changed. What is divergence? For example, Bitcoin rose to 90,000, and then fell back to 100,000, creating a higher high, but when the RUP was 100,000, it was not as high as when it was 90,000, but went down. This is the so-called RUP lowered, but the price higher. It is very strange why this situation occurred? The only logical explanation for this is that, as we just said, RUP is calculated using unrealized profits. The majority of unrealized profits in the market are actually contributed by low-cost chips. For example, if you buy a bitcoin at 16,000 today and it is now 96,000, the floating profit of this bitcoin alone is 80,000. But if you buy a bitcoin at 86,000 today and it is now 96,000, this one is only 10,000, so the main contribution in terms of proportion is made by low-cost chips. So once your price is higher, but RUP is lower, it means that a part or even a large amount of low-cost chips must have been sold in advance, resulting in your subsequent higher price. Because these low-cost chips have left the market, they have turned part of the unrealized profits into realized profits, so they cannot be seen in RUP, which will lead to a lower RUP and a deviation. This can help me get a verification in interpreting RUP, that is, there are indeed low-cost chips leaving the market.
As for the third aspect, there is actually a lot to talk about on-chain data, but I personally share another unique point of view, which is the US stock market. If someone has studied the stock market, they will know that there is a so-called valuation concept in the stock market, that is, the price-earnings ratio, or the price-earnings ratio. There are many different variations of this valuation method. The indicator I personally refer to is called Shiller ECY. This indicator comes from Professor Shiller of Yale University. He measures the yield of stock targets relative to bond targets. This indicator was mentioned in a paper he published after the outbreak of the epidemic in 2020. Because he believes that his previous model or data is called Shiller PE, called Shiller Price-Earnings Ratio. He believes that after the epidemic, because of the changes in the structure of the global market, many conditions are actually different from before, so he invented a new indicator called Shiller ECY to measure this market, and then found that this indicator has a better prediction effect. In a nutshell, this indicator currently shows that the valuation of the US stock market is already a little too high. One thing needs to be clarified here. A high valuation does not mean that it must fall. After a high valuation, it can be higher, and even higher. But it measures a concept similar to a spectrum, that is, it is getting closer and closer to the danger zone. In fact, the position it is approaching now is relatively dangerous in my opinion. The valuation of the stock market is currently mainly contributed by the hottest topic, that is, AI. Some time ago, there was a DeepSeek, which suddenly appeared unexpectedly and caused a sudden downward adjustment in the valuation of the US stock market. But in fact, on this point, I am pessimistic in the short and medium term. Because although DeepSeek is a chip decline in the long run, it is definitely good for the AI industry, but in the short term, I don’t think this valuation effect will end so quickly, so I think there is still room for downward adjustment in valuation. If the US stock market is not good, then Bitcoin, as a younger brother, will naturally not look too good. But these are my personal biases, personal biases, for your reference.
Alex: Okay, Colin just spoke in great detail, let's briefly review his views. He believes that the current price range has met many of the previous valuation peaks or price peaks, including the chip distribution he just mentioned, the unrealized profit ratio, and he also cited Professor Shiller's ECY indicator in the traditional financial market. He believes that the current price range meets many signs of peaking.
How to get started with on-chain data analysis
Alex: Today we have talked a lot about the analysis principles of on-chain data, including how to observe some common data and how to use these data in practice. Many of our listeners may not have studied this concept or system in depth before. So suppose a beginner asks you for advice and says, "Colin, I think what you said today is very attractive to me. I also want to learn this knowledge from scratch and guide myself to make some BTC investments." What kind of learning advice would you give them to help them start this period of learning?
Colin: Okay, actually I have received dozens of private messages asking similar questions so far. My personal advice has always been the same. First of all, I have two main strengths. The first strength is on-chain data, and the second strength I think is in the field of technical analysis. In fact, when most people come to ask me, they usually hold a line chart, draw some morphology or draw an indicator, MACD, RSI, and they use these things to ask me if there is a way to match this thing with the on-chain data point of view. In fact, I must give a suggestion here first. I personally do not recommend that novices start learning from the field of technical analysis. The main reason is very simple, because there are too many schools of thought, and some of the views in many schools cannot withstand the test of science. Because they are simply inductive methods, there is no logic behind them, and it is easy to go back to the example of the barking of dogs and heavy rain that I just talked about. In fact, it is entirely possible that it is a survivor bias, but the average novice has no ability to distinguish whether this is really useful or just a survivor bias. My personal suggestion is that on-chain data is a very suitable field for novices, and I will mention the way to learn it later. I think the reason why it is suitable for beginners is very simple. First, most of the retail investors around us, or our traders, are not full-time traders. Most of them may be high school students, college students or office workers. They actually have their own jobs. If you can't spend a lot of time on the so-called market monitoring, the role of on-chain data trading is very suitable for you. Because we mentioned earlier that the level of on-chain data observation is very high, at least it starts at the daily level. Since you observe the daily level, it means that you make operations based on on-chain signals. For example, the frequency of buying or selling is actually very low. You don't need to make 5 or 10 transactions a day. You may only make four or five transactions a year at most. So I think this is very consistent with the daily routine of students or office workers in terms of observation. You don't need to spend too much time. You may take out half an hour to an hour every day to observe the set alarms. You observe whether there are any different changes in these data. The second part is how to learn. I mentioned earlier that in my own learning process, I have not seen any free and systematic teaching to this day. There are many teaching methods, but they are not systematic. They may give you a long article that introduces one or two indicators in great detail. I think these articles are great, but the problem is that you still don't have a framework from 0 to 1, so it's actually quite painful to learn. This indicator looks great, so should I learn it or study it in depth? The next indicator also looks great, so which one should I start learning? My own approach is to make steel by primitive methods. I am more direct, because I don't know which one is good or bad at the beginning, so I learn all of them. I will look at the principles of each one, and see what the calculation principle is, why the author wants to design such a formula, what does he want to see, and can this formula really help him see what he wants to see? This takes a lot of time. After reading all these indicators, you have to screen them. But for novices, this process requires a lot of patience, and you have to read them one by one slowly. Because trading is not an easy thing. From what I have seen so far, whether it is simplified or traditional Chinese, the resources available in the Chinese area are quite few. So my suggestion is that if you want to study a certain indicator, it would be best if you can find the original author's article. Try not to read other people's articles. The original author himself is definitely the person who understands the indicator best. If you really can't find it, at least read his formula. There is a column called Weekly onchain on the website of Glassnode mentioned just now. They will publish a weekly report every week based on some different indicators, not fixed indicators, to share the current market conditions and why they think the current market conditions are like this. Then you can see a variety of indicators from it, you can grab each indicator and study it, and you will have a large library of learning materials. There are some teachings on my Twitter, which are not systematic. If you are interested, you can also take a look.
Alex: It's quite systematic. I have been following your updates. It seems that you have written more than ten articles. Basically, each issue talks about an indicator concept. You can also take a look. I have another question. You just mentioned that your first identity is a trader. Today we spent a lot of time talking about the help of on-chain data for trading. But in fact, when you are trading, in addition to the analysis of on-chain data indicators, do you refer to other factors? For example, macroeconomics, for example, some fundamental events of Bitcoin, such as the US state finances or even the national finances are promoting the reserve of Bitcoin. In addition to on-chain data analysis, other indicators are used as references for your transactions. What is the approximate weight of each indicator in your entire trading decision?
Colin: Okay, I think this question is very profound. First of all, in terms of my system, the on-chain data part can be thought of as an independent system for my position configuration. I will have a relatively large long-tail so-called spot configuration, and even at the bottom of the bear market, I will slightly leverage it, for example, 1.5 times or 1.3 times. This is a system, and the main trading decision-making basis of this system is based on on-chain data. The on-chain data will provide me with a general direction framework. I will know whether it is the early, middle or late stage of the market, whether it is a bull market or a bear market. It provides a general direction guidance benefit. As for other parts, I mentioned earlier that my other strength is the technical analysis part. In fact, there is no way to talk too much about this part because it is too complicated. Many schools of thought and some premise assumptions must be explained first. If they are not explained clearly, it will easily mislead others. For the technical analysis part, I will use it for short-term to medium-term trading orders. The main role of technical analysis in my own trading system is to refine the final entry point. That is, if I have confirmed that I want to take a certain opportunity today, where will I enter the market in the end for this trading opportunity? I will try to use technical analysis to refine my entry point. I will give an example casually. This is not a financial suggestion. Suppose that Ethereum can enter the market between 2000 and 2600. I think it will definitely rise later. Then suppose I am God, I know it will rise, so of course I will buy it. But because I am not God, I will try to get a more satisfactory entry point in this area through technical analysis. As for what this number is, I have to make an assessment every time, so there is no way to get a precise data, but I will have a set of measurement benchmarks. Next is the macro level. I am more concerned about the supply chain of the global market and the decision of the US Federal Reserve, because in fact, the United States still has a relatively large influence in the financial market. Their expectations of raising and lowering interest rates will have a very serious impact on the risk market. For example, if the CPI data is not very good recently, the risk market will make a corresponding pricing, because the market is priced in advance. They are trading expectations. It is impossible to wait until the interest rate is really cut before rising, and it is impossible to wait until the interest rate is really raised before falling. There will be an advance expectation. Those futures traders or option traders will make a pricing based on the overall judgment of the market. So this part is also what I will pay more attention to, but my macro is not as deep as my technical analysis or on-chain data, which is my relative weakness. Finally, there is the news or fundamentals mentioned by Alex just now, the so-called strategic reserve news. This part actually goes back to what I said at the beginning that I like to do more, that is, I will design some event-driven trading strategies. This is to make some trading opportunities with higher certainty for specific events. Let me give you an example. In late May last year, there was a senior ETF analyst named Eric at Bloomberg. The market paid close attention to his posts. He suddenly posted a post at 3:00 a.m. Eastern Time 8 saying that the probability of Ethereum ETF passing was adjusted to 75%. At that time, the entire market expected that Ethereum ETF would not pass. As soon as the news came out, Ethereum rose by 20% within 24 hours, and the increase in value directly exceeded Solana, which was very impressive. After such news came out, the first thing I thought of was to start preparing to find time to cut in and do an event-driven transaction, that is, to prepare to go long Solana and go short ETH at the same time. The background is actually very simple, because the whole world knows that the ETF is going to pass, which is a very big positive, so Ethereum will immediately pull up the market, which is very simple. The real thing to think about is who will be the next one? In terms of the market environment at that time, the support or popularity of Litecoin and Dogecoin was not as high as Solana. At that time, the first thing I targeted was Solana, and then about a week after that period, I began to lay out the trading opportunities of Solana's long and short strategies for ETH. Simply put, it is to use contracts to go long Solana, and then short ETH, to take advantage of the price increase of the exchange rate between the two. I think the next expected hype is Solana, because Ethereum is already a confirmed fact. Assuming that Ethereum really passes, Solana will inevitably receive a wave of related increases. Some people may ask, can your idea stand the test? I dare not say 100%, but there is a most obvious example. In January 2024, I don’t know how many people found that the day when the Bitcoin ETF was passed, Ethereum skyrocketed, and the exchange rate also skyrocketed directly at that time. If I remember correctly, the exchange rate of ETH to BTC rose by about 30 percentage points within 24 hours. Many people have doubts, Bitcoin ETF has passed, what does it have to do with Ethereum? The next hype is Ethereum. So this is one of the so-called event-driven transactions. Back to Alex’s question, I think it is too difficult to quantify the part of paying attention to news or fundamentals, so I personally prefer to design some event-driven strategies to deal with these opportunities where there may be operating space for inefficient market pricing.
Alex: I understand. Thanks to Colin for his very logical and organized explanation. He explained the thinking behind each operation strategy very clearly, including what scenarios it might be applicable to. It can be seen that he has a very rich toolbox and knows what tools to use in what scenarios, rather than making a vague decision based on his feelings.
Daily life of an on-chain data researcher
Alex: So, the last question is, as a trader and an on-chain data analyst, what is your typical working day like? In addition to paying attention to on-chain data, what other information might you look at or what tools might you use?
Colin: Okay, this is an interesting question, because my daily life is quite boring. My work and rest schedule is not very normal, but I will try to stay awake when the US stock market opens. The reason is very simple, because the US stock market opens when the liquidity of the Crypto market is usually the best. If my physical strength allows, I will look for some short-term trading opportunities during this period. This is actually a habit I have developed for several years. If I am really tired during the day, I will take a short nap to make up for the lack of sleep, because the chance of missing the market during the day is relatively low, and the chance of missing the market at night is relatively high, and it is more valuable to watch the market. In fact, you can find that every weekend or weekday daytime, during the daytime of Asian time, the market is actually quite boring in most cases, that is, it is sideways, there is no trading volume, and the liquidity is very poor, so this is why I try to stay awake in the middle of the night. After I get up normally, I will get up in the morning. In addition to observing, as Alex said, whether there are any changes in the on-chain data, I will observe and record some additional data that I want to see. In addition to the candlestick chart, I will regularly scan all the coins that I usually pay attention to. I will also manually record the net inflow and outflow of the US Bitcoin and Ethereum ETFs, as well as the market volatility. I will look at the fear and greed index because it is another quantified indicator to measure market sentiment. There is also the open interest in the contract market. If there is an extreme surge or plunge today, I may also look at the liquidation. I will record all these data. I am quite sensitive to these data. The remaining data is to see if there are any additional events. Once they happen, I want to see if there are any changes in this data. Normally, the fixed ones are the ones I just mentioned, the open interest in the contract market, market volatility, fear and greed index, and the net inflow and outflow of ETFs. There is another data that I like to look at, which is whether Coinbase’s contract quotes are at a premium or discount compared to mainstream exchanges, such as Binance and OKX. This is also a sentiment indicator that I personally think can be quantified. Sentiment refers to the sentiment of US funds, that is, the sentiment of people in the United States. For example, if the premium of Coinbase is obvious, it means that their buying may be relatively strong. This happened very obviously when Trump was elected. If there is any abnormal movement in these numbers, I actually observe it every day. I have to maintain this sensitivity. Once I find it, I will start to think about whether it is groundless or there are some trading opportunities. In addition to the time I record these data, I will also watch the market at other times, because as I mentioned earlier, technical analysis is one of my few strengths that I can brag about. I will spend a short period of time, such as a few hours, to watch the market, and then observe whether my daily planned and revised trading plan has reached the position I expected. If it is close to or has reached it, I will immediately concentrate on the market and look at the data I want to see. Or whether the trading plan has deviated and needs to be corrected. I have two screens. On the other screen, I open Twitter and manage my own Mr. Berg account on Twitter. The rest of the day is pretty boring. I go out for a run occasionally, but not very often. The purpose is to keep myself moving and not be without exercise all day. The rest of the day is mainly spent with my family. So my day is pretty boring, there is nothing particularly eye-catching, because trading is actually my job, so there is not much difference between me and ordinary office workers or students. I mainly work, then go home, eat, and sleep, that's about it.
Alex: I see. Colin just talked about his work for a day. The amount of information and his mental workload are quite large, but he may have fixed and modularized it, so the brain does not need to be activated every day to do a series of important tasks, including data follow-up, etc. He has a habit of doing what in each period, and has a very clear arrangement, so that he can enter a state faster. We can also observe that Colin himself is very curious about trading, investment, and the business world. He gets more than money from it. I feel that he has a lot of fun. I think such a state is an important talent for a good trader and a good investor. Thank you Colin for coming to the show today to share with us so many thoughts and systematic explanations on chain data analysis, investment, and trading. I hope that in future shows, we can invite Colin to tell us more knowledge in other aspects. Thank you Colin.
Colin: Alex, you are very kind. I am just sharing my personal opinion. Thank you.
