1.Statistical Samples

The statistical samples are 85,195 LPs from the inception of Uniswap V3 to 27 Oct 2022. Among them 1,066 LPs are removed from the samples because they did not continue to hold the position and transferred to another address after the first creation of the position . So there are 84,129 samples left.

Based on the related events of all positions in the metadata on the chain and income index of LPs and its position summarized every hour ,we derived the hour-level LP characteristic index , which was used for LP characteristic distribution statistics , LP clustering and user portrait establishment .

2.LP characteristic distribution statistics

2.1 LP’s largest token pools held in history and the distribution of the number of positions

We respectively counted the largest number of token pools , the largest number of positions and the largest number of positions in single token pools held by LP at the same time period . The quantity distribution as followed chart 2-1-1 ,We found that more than 66,000 LPs only make market in one pool at the same time . More than 66,000 LPs hold one position in single token pool at the same time . More than 57,000 LPs only hold one position in uni-V3 at the same time . That means most LPs are holding only one position in one single pool . Only a small number of LPs choose investment diversification.

We divide LPs into single_position holders and multi_positions holders . The position distribution of multi_positions holders in a single token pool as followed chart 2-1-2:

LP address：0x741aa7cfb2c7bf2a1e7d4da2e3df6a56ca4131f3；

POOL address：0x8ad599c3a0ff1de082011efddc58f1908eb6e6d8；

2.2 Statistical analysis of LP’s total mining time and the frequency of adjusting the position

Total mining days is the time from the first time LP add liquidity and the last time remove liquidity . The distribution of total mining days is mainly concentrated in 0-100 days and it’s like  long-tailed distributions . Also we found about 20,000 LPs with a total mining days of more than 200 days. There are around 16,000 LPs with a total mining days less than or equals to 3 days .

Operate_cnt means the number of mint and burn actions LP performs on the position. For statistical purpose, all adjusting movements in a single position are counted on the last owner of the position . It can be seen from chart 2-2-1 that the times of position adjustment are mainly concentrated in 2 or 3 times. There are about 54,000 LPs whose positions are changed less than or equal to 3 times .

As a single position which has been removed liquidity at least has two adjustment actions which one of them is add liquidity and the other one is remove liquidity . Therefore LPs with operate_cnt less than or equal to 2 times are considered single_operation. LPs with operate_cnt  between 2 to 10 are considered as multi_operations . LPs with operate_cnt greater than 10 times are considered as many_operations . The distribution of adjustment as followed chart 2-2-2

According to the number of adjustment times , we draw the chart of net APR . We found these three types of LPs yield distribution are obviously different . And those LPs who have many_operations their distribution of yield is most concentrated. There’s no LPs with extremely small yield . And LPs with single adjustment their distribution of yield is most decentralized.

2.3LP’s last removing liquidity time and distribution of APR

According to the scatter plot of net APR and LP’s last time removing liquidity , and the net APR distribution of removing liquidity month , There’s no obvious connection between LP’s total APR and its last removing liquidity time .