Meme Training Manual: Rebirth: I Want to Be a Diamond Hand (Part 2) | Produced by Nanzhi

Original | Odaily Planet Daily ( @OdailyChina )

Author｜Nan Zhi ( @Assassin_Malvo )

In the previous article Meme Training Manual: Rebirth: I Want to Be a Diamond Hand (Part 1) , the author proposed 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.

Compared with a week ago, there have been major changes in the funding and score evaluation methods of the copy trading system . 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 13 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.

Basic information

• Account separation: In the previous article, only one account was used to follow the smart address that met the Diamond Hand System. In order to test some addresses that did not meet the system requirements but had strong profitability and profit and loss , a follower account No. 2 was set up . Due to some problems with the follower bot (explained below), multiple accounts such as No. 3 and No. 4 will be set up in the future.

• Original capital: Follower No. 1 invested a total of 7+ 8 = 15 SOL, which is also the original capital of the entire system ; Follower No. 2 invested a total of 5+ 5+ 5 = 15 SOL; Follower No. 3 invested a total of 10 SOL, but it has not yet been activated due to the lack of a secure Tg account.

• Positions and Profits: As of 07:00 on October 21 (UTC+8), the cumulative withdrawal was 60 SOL, the total assets of Follower No. 1 were about 57 SOL, the total assets of Follower No. 2 were about 14.9 SOL, and the total assets of Follower No. 3 were 10 SOL. The total profit was 141.9 SOL, and the profit margin was 846% .

Screenshot of assets before withdrawal on October 20

Diamond Hand Evaluation System

Score calculation method

The first version mentioned in the previous article designed 13 evaluation dimensions, and its core problems include: ① the evaluation system is too greatly affected by the amount of funds. For the same profit and loss ratio and winning rate, addresses with high amounts of funds have a significant advantage; ② it cannot block malicious addresses; and ③ there is a lack of statistics on the holding period.

Last week, the author made some changes to question ①. The latest evaluation system has 16 evaluation dimensions , as follows:

• 30-day win rate, out of 1 point: the score is equal to the 30-day win rate;

• 7-day profit, full score 1 point: 7-day profit greater than $50,000 full score, profit from $0 to $50,000 linear score, negative profit 0 point;

• 30-day profit, full score 1 point: 30-day profit greater than $100,000 full score, profit from $0 to $100,000 linear score, negative profit 0 point;

• (New) Profit stability , full score 0.5 points: if the profit in 30 days is twice that in 7 days, 0.5 points; 1.5 times - 2 times is 0.25 points; if the profit in 30 days is less than that in 7 days, -1 point, and 0 points in other cases.

• (Adjusted) Average profit per purchase , full score 0.3 points: 30-day profit ÷ 30-day purchase times, less than 300 gets 0 points, greater than 2000 gets 0.3 points, and the linear score in the middle;

• (New) Average profit rate of each purchase , full score 0.7 points: total purchase funds in 30 days ÷ number of purchases in 30 days, score = e^(-2.5 × average profit rate of each purchase) × 0.7.

• (Adjusted) Average profit per sale , full score 0.3 points: 30-day profit ÷ 30-day sales times, less than 300 gets 0 points, greater than 2000 gets 0.3 points, with a linear score in the middle;

• (New) Average profit rate of each sale , full score 0.7 points: total purchase funds in 30 days ÷ number of sales in 30 days, score = e^(-2.5 × average profit rate of each sale) × 0.7.

(Note: Through the above adjusted indicators, the high profit-loss ratio goal is strengthened and the impact of the amount of funds is weakened)

• Trading style, full score 1 point: average profit of each sale ÷ average profit of each purchase, calculate the score if it is greater than 1.5, otherwise 0 points. The score is equal to average profit of each sale × 50%. (Note: I hope to find an address that buys in small quantities and sells in batches at high points)

• (Adjusted) Total profit, full score 1 point: less than $10,000 gets 0 points, more than $50,000 gets 1 point, and the middle is linearly scored; (Note: Total profit = the sum of the profit of the 30 tokens with the largest profit + the loss of the 30 tokens with the largest loss. The upper limit is reduced from $100,000 to $50,000.)

• (Adjusted) Profit TOP 1 ratio, full score 0.166 points: calculated as the most profitable token ÷ total profit, score when the ratio is greater than 0% and less than 50% , otherwise 0 points. The score is equal to total profit score ÷ 6;

• (Adjusted) Profit TOP 5 ratio, full score 0.166 points: calculated as the top 5 profitable tokens ÷ total profit, score when the ratio is greater than 0% and less than 75% , otherwise 0 points. The score is equal to total profit score ÷ 6;

• (Adjusted) Profit TOP 10 ratio, full score 0.166 points: calculated as the top 10 profitable tokens ÷ total profit, score when the ratio is greater than 0% and less than 100% , otherwise 0 points. The score is equal to total profit score ÷ 6;

• In addition, there are three indicators: loss TOP 1, TOP 5, and TOP 10. The methods are similar to the above three indicators and will not be repeated here.

(Note: The upper limit of total profit has been lowered to reduce the impact of capital volume. The score ranges of TOP 1, 5, and 10 have been lowered from 75%, 100%, and 150% to 50%, 75%, and 100%, aiming to find addresses with stronger profit stability. Of course, this will reduce the ability to search for diamond hands.)

Veto system

Due to the need to review too many addresses, a new veto rating has been added. If any of the following conditions are met, the rating will be vetoed and deleted directly:

• The average profit rate of each purchase and the average profit rate of each sale are less than 0.1

• The number of transactions in 30 days is greater than 3,000

• Total profit is less than zero

In addition, the following conditions will be given a risk warning label:

• The average profit rate of each purchase and the average profit rate of each sale are less than 0.2

• The number of transactions in 30 days is greater than 2,000

Did Diamond Hand win?

Judging from the results of the 13-day test, Diamond Hand won again .

Copy trader No. 1 purchased a total of 165 tokens in 13 days, with a winning rate of about 40% and a profit rate of about 900% .

Based on a profit of $20,000, the top 1 earner made a total profit of $12,100, accounting for 60% of the total profit; the top 5 earned a total profit of $21,109, accounting for 105% of the total profit.

The top 1 loser lost a total of $1,432, accounting for 7.2% of the total profit; the top 5 lost a total of $2,291, accounting for 11.5% of the total profit.

Copy No. 2 bought a total of 219 tokens in 5 days, with a win rate of about 20% and a profit rate of about -10% . This account is designed to test addresses that have a much higher PvP frequency than Copy No. 1, but also have certain diamond hand characteristics. It is planned to transfer addresses that have positive EV contributions to the system to Copy No. 1 (but no one has qualified yet, and several aggressive addresses have brought more losses).

In fact, the critical hit ability of Copy Trader No. 2 is not weak. The TOP 1 profit is 900 US dollars, and the TOP 5 profit is 2600 US dollars. However, the overly high frequency of PvP has brought a huge amount of wear and tear, and the total profit has been hovering around 0% for a long time .

(Note: Critical Hit refers to a situation where a single token experiences a significant profit.)

Second Edition of System Reflection

In the previous article, the authors reflections on the first version of the system included be sure to set the minimum copy order amount, believe in smart money chasing highs, and dont take the initiative to stop profit. These strategic guidance are fine in general, but there are still gaps in actual operation, as follows:

• Should we chase high prices? Smart address A, which is tracked by the follower number, chased high prices twice last week, with market values of 4 million and 11 million US dollars respectively, and purchased 20,000 US dollars and 40,000 US dollars respectively. The former rose by 3 times at most after purchase, and the latter was purchased at the highest point. Smart address A often holds coins for more than two or three weeks before selling them, resulting in the latter token causing the largest loss of 1,432 US dollars in the history of the follower number , which is 7 times the second largest loss of 200 US dollars . The number of samples of whether to chase high prices should be followed is too small to give a conclusion, but the author believes that it is very important to limit the position of chasing high prices , but Bot does not have the function of limiting the amount of holdings, and the author plans to achieve this through semicolons .

• Dont take the initiative to stop profit? The total investment of the most profitable token is $80, and the profit is $12,000. However, if the copy target does not move the token, the maximum amount that can be sold is $80,000. The problem is that the author intervenes in this order and takes the initiative to stop profit. If the copy target takes the profit, the profit will be $20,000. The copy targets profit fell from $600,000 to $100,000 before it began to stop profit gradually, but the strategy is still better than the authors active intervention .

• Minimum copy order amount: Previously, we mentioned that there was a third-party poisoning situation. After feedback from the author and Bot official, it has been fixed, but it is still recommended to enable restrictions to avoid other new poisoning methods.

• Necessity of semicolons : In addition to the need to limit positions through semicolons, the specific mechanism of Bot generates further semicolon needs: suppose two addresses A and B are copied, and they buy 10 tokens respectively, and the copying account also follows the order to buy 10 × 2 tokens. At this time , if address B sells 100% of the tokens (10), the copying account will sell 100% of the tokens (20), not 50% . Therefore, if a paper hand is mixed in a group of smart addresses, it may cause all tokens to be accidentally sold (it has actually happened), so semicolon division is very necessary.

Next steps

After completing the score evaluation, the copy trading system still needs manual screening. The next step is to include some manual discrimination factors in the statistical scope. In addition, the evaluation method of the copy trading coefficient is too simple and crude, and it is easy to misjudge some addresses that love Pump internal transactions as copy trading, which also needs further upgrades.