Can we objectively measure (point in time) relay filter effectiveness?
We can by looking at the friction a Tx/usecase faces getting into a block.
Filtering on datacarrier (preventing OP_RETURNs from exceeding 83 bytes) lives in the [pink box]
Sharing this as a discussion point.
nostr:nprofile1qyg8wumn8ghj7nn0wd68ytnhd9hx2qg4waehxw309ajkgetw9ehx7um5wghxcctwvsqzp6y2dy0f3kvc0jty2gwl7cqztas8qqmc5jrerqxuhw622qnc2pq3x3x7wj Based on popular demand.
anything outside of the pink box is uncontroversial, so you need to justify why something should be included
nice, the X axis is important here. It's where the "filters work vs dont work" discussion is breaking down.
You are likely pink, most are pink
Yes we can run epidemiological simulations and Laurent a big defender of filtering has just proven that it doesn't work
nostr:nprofile1qyt8wumn8ghj7etyv4hzumn0wd68ytnvv9hxgtcpzemhxue69uhks6tnwshxummnw3ezumrpdejz7qpq2rv5lskctqxxs2c8rf2zlzc7xx3qpvzs3w4etgemauy9thegr43sugh36r, I'm curious if you agree with the chart.
~85% is generous imo. You could push it more aggressively to 95–99%, but I figured 85% would serve as a starting point, since I didn't want the whole discussion to revolve around the arbitrary majority needed for a filter to be noticed.
Yes, I think it is a good way to visualize the concept. Thank you for creating it. Two thoguhts: I would suggest using a different label for the x-axis, perhaps something more specific like "monetary transaction fees," instead of "transaction use case."
I also think that without filters, monetary txs can displace jpgs, which doesn't look like it is the case in this graph. Thanks again. Wdyt?
> I also think that without filters, monetary txs can displace jpgs, which doesn't look like it is the case in this graph.
I would generally agree. The minimum demand/value of a monetary tx/usecase would need to be higher than the highest demand/value of a "jpeg" tx in order to have zero spam/jpegs in the next block. This isn't intuitively represented I suppose for the graph (based on the x axis as you point out), but it's how I intended this to be interpreted.
I wonder of min relay fee would be a good model to analyze. Most nodes filtered less than 1sat/vbyte, but from a bystander's view when fees dropped below that, it didn't seem like there was any trouble filling blocks with sub 1/sat/vbyte. How would you measure the friction?
