Original | Odaily Planet Daily ( @OdailyChina )
Author｜Nan Zhi ( @Assassin_Malvo )
In the previous articles Meme Training Manual: Rebirth of Diamond Hand (I) and Meme Training Manual: Rebirth of Diamond Hand (II) , the author formulated a score evaluation method for smart addresses, aiming to find diamond hand players among these smart addresses and follow their orders, and explore whether following diamond hands can be a way to make stable profits .
Last week, the copy trading account experienced a large drawdown. After evaluation by the author, using trading frequency and overall profit and loss ratio as core evaluation factors is still not comprehensive enough and has too large granularity. Therefore, a set of in-depth evaluation methods has been formulated . This article will disclose and review them, and further share and discuss the copy trading system with readers.
Risk warning: This system has been running for only 21 days since it was launched on October 8. Although it has achieved certain results, the amount of data and control experiments are not sufficient from an objective point of view, and the rigor is limited . It is only for readers reference. In addition, to ensure the objectivity of subsequent tests, this article will not disclose any addresses involved in the operation.
Reflection on the withdrawal: Profits and losses come from the same source after all
Pullback Scenario
Last week, there were four accounts running, i.e., accounts 1 to 4. Among them, accounts 1 and 2 were core accounts, which copied the addresses that clearly met the system requirements. Accounts 3 and 4 adopted some aggressive strategies. During this period, accounts 1 and 2 also experienced significant retracements. The root causes of these retracements are mainly as follows:
High-level, high-frequency heavy purchases: As mentioned in the second article, some addresses like to buy tokens with a market value of millions to tens of millions of dollars multiple times without setting stop losses. These tokens are often only born for a few hours to a few days, and their stability and certainty are not strong, so the profit and loss ratio of such transactions is extremely poor. An actual scenario is that when a token has a market value of 11 million, the copy object buys 20 consecutive transactions, and copy No. 1 sets a single purchase amount of 0.8 SOL, buying a total of 16 SOL. A few hours later, the token plummeted to a market value of 3 million US dollars, causing the largest loss in the account.
Buy at low positions, high frequencies, and low winning rates : Some accounts are keen on buying in the Pump market or at a market value of hundreds of thousands. After losing more than ten transactions, they will make a critical hit profit once. If these addresses insist on the strategy of not stopping losses and holding for a long time, they will also meet the Diamond Hand requirements. However, due to the low price fluctuation problem, the profitability of the copying address will be significantly lower than that of the copied address, and the EV will be negative in the long run.
Let’s get back to the basics. What should be the profile of the person you’re following?
Therefore, the author re-sorted the portraits of the intended follow-up objects to guide the upgrade of the evaluation system. The original text of the thinking logic at that time is as follows:
What kind of address should I look for? An address with low transaction frequency, long holding time, entry at the bottom rather than FOMO chasing high, good profit and loss, no cutting, and low retracement.
Which factor is most important to me ?
Low retracement ranks first, so we need to find out the investment ratio when the loss is the largest, and how many consecutive losses are required to get a golden dog;
Secondly, in order to make money, in addition to the total profit and loss, the details of the input cost must be calculated. It is not desirable to use a large bet on a small address, and the proportion of high-level FOMO in the loss must be calculated;
Dont bet big on small, but also need to reflect the ability to enter the bottom, and calculate the proportion of low-level entry in the profit;
The low transaction frequency is also reflected in the low number of purchases;
The length of time you hold it can be counted in the 4th point statistics, and the same applies if you don’t cut anyone off.
In-depth evaluation system
According to the above portrait requirements, the author has completed the addition of two in-depth evaluation systems, the details are as follows:
Trading point evaluation: small investment, big investment, anti-FOMO
The idea of this system is that for each address:
Get a list of its most profitable and most losing tokens;
Filter tokens with profit or loss greater than a certain threshold. For example, the author selects tokens with profit greater than $5,000 and loss greater than $2,000.
Get the transaction details of the first purchase, second purchase, and so on of the token by the address until the first sale;
Count the market value of each purchase and calculate the average market value of purchases;
Count the number of purchases before the first sale;
Calculate total profit and total cost;
Count the holding time from the first purchase to the first sale.
The statistical results are shown in the figure below. From this table, we can get:
The ability of each address to “buy at the bottom” , for example, the first 7 addresses were able to buy below the million dollar market value, and the profit and loss ratio was extremely high, with only one or two transactions;
The risk of losing money due to “high FOMO” . For example, the last address bought in only when the market value was an average of 20 million USD, which is obviously not suitable for copy trading.
The ability to make a small investment for a big gain is obtained by dividing the profit portion by the average profit and loss ratio;
Stop loss style , obtained by dividing the loss part by the average profit and loss ratio. The closer it is to -1, the less this address likes stop loss.
The degree of “diamond hand” is obtained by the average holding time.
Recent trading levels review
The table obtained in the previous section shows the top profit and loss of this address in history, which cannot reflect the recent transaction level. On the other hand, the data provided by the data platform is not completely accurate. The authors statistical path for recent transaction level is as follows:
Get the 100 tokens recently traded by an address;
Exclude tokens not created by Pump;
Exclude tokens created more than 30 days ago;
Get the total profit and loss of the token;
Statistics of the winning rate of tokens after screening.
Document system data reporting and document system settings
Although it has experienced a round of pullback, the stability of the system has been significantly improved after the upgrade. Currently, the No. 1 and No. 2 copy orders have returned to ATH, while the No. 3 and No. 4 strategies failed to be verified and have been liquidated .
Here, we disclose the copy target of Copy Order No. 3, which is HdxkiXqeN 6 qpK 2 YbG 51 W 23 QSWj 3 Yygc 1 eEk 2 zwmKJExp, known as Alien Cat, which is an address known to the whole universe. Copy Order No. 3 is intended to test this type of address with strong profitability but known to everyone, and the final test result is that the account loses all its money.
The data of each wallet as of 09:00 on October 29 (UTC+8) is as follows, in SOL:
Copy order setting details
The bot used by the author is Dogee , and the details of most copycat address settings are as follows. The key points are:
Set a minimum order amount to prevent the poisoning mentioned in the first and second articles ;
Buy all, in fact, should be set separately according to the internal market winning rate and profit and loss ratio;
Sell in proportion. The ultimate goal and core judgment factors of this system include buy once at a low price and sell in batches at a high price . Therefore, you need to choose to sell in proportion with the copied order object;
Market value restrictions make it extremely difficult to make profits by chasing high prices. Multiple addresses have set a market value cap of 10 million US dollars. But from an objective data perspective, how profitable is the chasing high strategy? A special article will be published in the near future .
