BTC transaction size, fee per kB (satoshi/ USD) -- median, p75, max, % in 2020

[tranthidung]

ABSTRACT

• The network conditions are reflected by block size that in turn causes fee is high or low. The median value of daily block size in October is above 1.23 MB.
• In 2020 so far, the median of daily block size is 1.27 MB
• October is the month with transaction size are stable around 1.27 MB (see box plot with vertical line, box, whisker and raw results). In median, the latest 3 months (August - October) have same value at 1.27 MB
• In differences (%) between median of fee (in specific day to all days in 2020), people are paying more expensive fees in October: around 508% (in satoshis) or 20224% (in USD).
• The medians (all days in 2020) of fee per kB are 32467.2 satoshi or 0.1 USD
• In the last 3 months, more people accept higher fees that reflects by narrower gaps between differences of max - median and p75 - median (see plot)
• In median of daily max fee: people tend to accept 10⁺ USD for each kB in October. However, October is not the highest month, it is a third one instead. The 2 highest months are May and August. Reminder is outliers in October (latest days) are much higher than outliers in May or August.
• In median of daily daily median fee: people tend to accept 5 ⁺ USD for each kB in October and it is the third highest month, after May and August. The same data interpretation for outliers.

Time series plots

Block size in MB (it is after I calculated medians for daily block size).

Code:

Summary for variables: sizeMB
     by categories of: month 

    month |         N      mean        sd       p50       p25       p75       min       max
----------+--------------------------------------------------------------------------------
   2020m1 |     31.00      1.13      0.17      1.21      1.05      1.22      0.51      1.26
   2020m2 |     29.00      1.23      0.04      1.24      1.21      1.26      1.13      1.28
   2020m3 |     31.00      1.21      0.10      1.25      1.22      1.27      0.85      1.33
   2020m4 |     30.00      1.19      0.12      1.23      1.17      1.26      0.79      1.30
   2020m5 |     31.00      1.29      0.03      1.28      1.27      1.31      1.22      1.36
   2020m6 |     30.00      1.23      0.12      1.27      1.24      1.29      0.78      1.33
   2020m7 |     31.00      1.27      0.04      1.28      1.27      1.30      1.12      1.35
   2020m8 |     31.00      1.29      0.02      1.29      1.27      1.31      1.25      1.34
   2020m9 |     30.00      1.28      0.04      1.29      1.27      1.31      1.15      1.39
  2020m10 |     29.00      1.29      0.02      1.29      1.28      1.30      1.23      1.35
----------+--------------------------------------------------------------------------------
    Total |    303.00      1.24      0.10      1.27      1.23      1.29      0.51      1.39
-------------------------------------------------------------------------------------------

Block size (for all blocks in 2020). I don't calculate daily median for this one.

Code:

Summary for variables: size
     by categories of: m 

        m |         N      mean        sd       p50       p25       p75       min       max
----------+--------------------------------------------------------------------------------
   2020m1 |   4737.00 974062.11 440761.55  1.19e+06 637357.00  1.29e+06    244.00  2.21e+06
   2020m2 |   4154.00  1.10e+06 381206.27  1.24e+06  1.05e+06  1.32e+06    293.00  2.28e+06
   2020m3 |   4255.00  1.10e+06 384647.89  1.24e+06  1.06e+06  1.32e+06    244.00  2.14e+06
   2020m4 |   4513.00  1.08e+06 390384.36  1.22e+06  1.05e+06  1.30e+06    244.00  2.42e+06
   2020m5 |   4192.00  1.27e+06 225990.99  1.28e+06  1.20e+06  1.36e+06    280.00  2.07e+06
   2020m6 |   4549.00  1.15e+06 373225.21  1.26e+06  1.13e+06  1.34e+06    244.00  2.20e+06
   2020m7 |   4589.00  1.23e+06 285914.06  1.28e+06  1.20e+06  1.35e+06    244.00  2.29e+06
   2020m8 |   4521.00  1.28e+06 230549.31  1.29e+06  1.21e+06  1.36e+06    217.00  2.33e+06
   2020m9 |   4531.00  1.24e+06 292441.24  1.29e+06  1.20e+06  1.36e+06    200.00  2.19e+06
  2020m10 |   3977.00  1.25e+06 266874.59  1.29e+06  1.21e+06  1.36e+06    213.00  2.33e+06
----------+--------------------------------------------------------------------------------
    Total |  44018.00  1.16e+06 349851.87  1.26e+06  1.15e+06  1.34e+06    200.00  2.42e+06
-------------------------------------------------------------------------------------------

In October 2020

Data was imported on 29 October 2020 (ends at block #654708), not recently but the plot shows time points at which mempool was clear sometimes in October. Whenever the p25 value of daily block size falls back to below 1 MB, mempool has chances to get clearer.

I made a typo in plot title: The period is 01 Oct 2020 - 29 Oct 2020.

All-time and 2020 plots

[tranthidung]

Bitcoin has its public ledger and all data is accessible so any analysis is reproducible. It is only a matter that which skills you have and which code you use for your analysis.

Today I see this tweet and the chart is similar to mine in OP (look at a bottom chart). It is a proof of transparency and reproducibility. This tweet gave me a information that I didn't know when I find the visually strange flatting period in my chart.

Some #Bitcoin history visualized:

The average block size chart beautifully depicts the SegWit soft fork in August 2017, that increased the 1 MB limit for block sizes.

In my chart. I understand the increase of block size after the soft fork in August 2017 (both values for p25, p75, and median block size moved up after that) but the flat area since late of 2014 (you see both three lines - median, p25, and p75 tend to be overlapped each other) make me confused.

Does it present that the block size hit its limit in almost more than 2 years and a half?

[bitmover]

Does it present that the block size hit its limit in almost more than 2 years and a half?

I think it is not that simple.
Blocksize depends on what type of transactions are in it. Because weight is more important than transaction size now, after segwit adoption.

After segwit hardfork blocks do not have 1MB, 2MB, etc... they are measured in virtual bytes, so the total blocksize is 1vMB,  or 4 million Weight Units.

All transactions have a different weight according to their address format. Legacy transactions weight is the same as their size in megabytes. But segwit transactions are much lighter.

If you have a block filled only with native segwit transactions it would have 4MB. More than double the size of the stats you posted.

The maximum size of a block in bytes is nearly equal in number to the maximum amount of block weight units, so 4M weight units allows a block of almost 4M bytes (4MB). This is not a somehow "made-up" size; the maximum block size is really almost 4MB on-disk and over-the-wire. However, this maximum can only be reached if the block is full of very weirdly-formatted transactions, so it should not usually be seen.

https://en.bitcoin.it/wiki/Weight_units

[tranthidung]

Does it present that the block size hit its limit in almost more than 2 years and a half?

I think it is not that simple.
Blocksize depends on what type of transactions are in it. Because weight is more important than transaction size now, after segwit adoption.

Thank you. I meant about the period (~2 1/2 years) before the soft fork, not after the soft fork. I guess the block size reached limits for more than 2 years (of course with ups and downs depend on transactions on the network). I am surprised that it takes the community more than 2 years to solve it (if what the chart discloses is true).

If you have a block filled only with native segwit transactions it would have 4MB. More than double the size of the stats you posted.

My stats are for median of block size every day, not max or min values. You are mentioning about max value of block size.

[bitmover]

Does it present that the block size hit its limit in almost more than 2 years and a half?

My stats are for median of block size every day, not max or min values. You are mentioning about max value of block size.

I know. But as you asked if we hit the block size limit... to reach the blocksize limit the median for every day would need to be 4mb. Isn't it?

Or, 1vMB median blocksize, or soemthing close to that. As your data is in mb, we cannot know how much space is really left in those blocks.