Introduction

In this bounty, we'll be investigating block finality, Transactions per second and overall network performance metrics for Avalanche, starting on '2022-06-20' until now.

A brief rundown of the general topics:

• Average Transactions per second by the hour.
• Average, max and min number of transactions per block.
• Average time between blocks over time.
• Network revenue instead of gas price over time.

All of the above will also be compared to other chains.

Average TPS by hour

Firstly, let's look at the average Transactions per second, hourly.

Overall

Failed + Successful

What we can see in the graph above is that Avalanche can churn about 2.5-5 transactions per second. This is quite low compared to Polygoon, Ethereum and Algorand, but it's ahead of such chains such as BSC, Optimism, Arbitrum and Osmosis on average.

Successful

Polygon is still in the lead here, with about ~20 TPS on average, whereas Avalanche clocks out around 2-4 TPS. A lot of the above networks have very similar values to the ones seen above with failed + successful combined, which is a good sign. Meaning that they have a negligible amount of failed transactions. Before we conclude that though, let's investigate which is the network with the highest TPS in failed transactions.

Failed

Polygon is still in the lead here, but we can see that the order of magnitude is significantly decreased. What we can also speculate is that due to Polygons very low fees, a lot of arbitrage bots/contracts are being run with a fixed gas price all the time, which could also be the reason for this higher-than-average (chain) failed TPS. Looking at Avalanche we can see that failed TPS start from ~0.1 with the highest value on June 24th at 20:00 at around 1.58 failed TPS.

Do note that we do not include Algorand in this bottom (failed) view compared to the others as there's no such concept of a failed transaction as far as I could find in the transactions table.

Average, max and min number of transactions per block

Secondly, let's look at how full of transactions a block is, on average, maximally and minimally. In this section, we'll be comparing with Ethereum and Polygon only.

Average

Avalanche does not score well in this metric against Ethereum or Polygon. In fact, it's only about 3% as full as that of Ethereum and only about 7% as full as that of Polygon. But, this could be due to lower transaction counts currently on Avalanche as we are in a macro bear market after all. Let's look at the MAX and MIN values and see what they tell us.

MAX

When we look at the maximum values we can see that Avalanche is up there with, and even exceeds, Polygon in terms of "fullness" of Transactions per block. What I am surprised about is the Ethereum max TX/block value which is ~2.25x that of both Polygon and Avalanche. This could be due to many reasons. One of which is that a lof of DeFi volume is still on Ethereum, and especially now, in a bear market, where TVLs are low across other L1s and L2s.

MIN

When looking at the minimal "fullness" in terms of TXs per unique block we can see that Avalanche is only about ~7% as "full" as Polygon and only about ~2% as full as Ethereum. Again the max and min values can be very sensitive to hourly fluctuations and thus may not be a reliable comparisson.

Average time between blocks

Next, let's look at the average time between blocks when comparing Avalanche with Ethereum and Polygon again.

NOTE: Polygon is not as well back-filled as the other two and the presented results should be taken with a pinch of salt. I would recommend excluding the MATIC_T_BTWN_BLOCKS by clicking on it, to get a cleaner comparison between Ethereum and Avalanche.

What we can see here is that it takes about ~ 16 seconds for each block to reach finality on Ethereum vs only ~2 seconds for Avalanche. This could also be a reason as to why earlier we saw that each block does not have as many transactions per block as on Ethereum or Polygon. It is clear to see here why that is. When we include Polygon, what we see are values that I assume could be outliers, but in the case where there aren't, it seems that a few blocks took about ~300 and ~850 seconds to reach finality, which is obviously not great. When looking at the averages below we can see that Avalanche has the fastest block finality out of both Ethereum and Polygon.

Network revenue instead of gas price

Last, but not least, let's look at network revenue and compare Avalanche to Polygon as Ethereum exceeds every single chain by orders of magnitude in terms of the amount in $ value we'll be looking at.

• Sources:

Unfortunately, I could not find AVAX price on any of the Ethereum hourly token prices tables, which makes sense, as it does not have an equivalent contract address on the network. For this reason, I sourced the fee data from Snowtrace, PolygonScan and the price data using CoinGecko API.

What we can see here is that Avalanche is netting about ~55k $, compared to ~80k $, in-network revenue from fees in $. Polygon has recently had a spike of up to ~140k $ on June the 18th. Also note that I went a little bit further back on this one, starting on the 15th of June as I got the data from external sources and the data are daily.

The calculation:Gas Price (wei) * Gas Used (wei) * 1e-18 * Price of AVAX and equivalently for Polygon.

Summary

Avalanche is a chain with very high transaction finality times and that could also be why it is suffering in a few of these metrics compared to other chains, with a less "full" Transaction Count per block. What is interesting to see is that Polygon has a lot more TPS than all of the other chains. On top of that what is also interesting to me is that Arbitrum, Optimism and Osmosis all also have very low TPS. At least for Osmosis, I was surprised as I thought it would have a lot higher TPS. In terms of network revenue Avalanche has had a steady past few days, with the exception of the spike ~ 18th of June, as well as it has kept in competition with networks such as Polygon.

Evaluation

I felt that doing unique active users or cumulative users here would not have yielded a lot of value in terms of the discussion and would have also been outside of the topic, so I decided not to plot these metrics. Also, instead of plotting average gas price hourly, I decided to plot network revenue as it's a much better metric when you want to compare multiple different networks with one another. I could have used gwei vs gwei gas price. But the gas itself is worth a lot more on Avalanche than on Polygon for example and multiples more on Ethereum than the latter two.