sorry for disconnecting.

@sarang just cleaned up our econ data set

Color is block size

I think most of the low-payout blocks are due to altruistic miners that include lots/large transactions even when mining an empty block would have paid more

Note that altruistic mining is an entirely separate matter from underclaiming (not collecting all of fees + emission)

*separate phenomenon

Isthmus: is it possible to alter create_block_template to maximize short term mining profit?

I didn't know that It can grab large tx while decreasing block reward.

only 21 days old

sorry, wrong channel*

Sure @cohcho, if you don’t mind tinkering with the source code, you could program it any way that you want (within limits of consensus rules)

Side note:

Let’s say the mempool contains apathetic or many transactions, such that including them results in a decrease in the *total* payout. We define two types of miners: altruistic miners mine the transaction(s) and take a hit to their bottom line (i.e. make a donation) so that transactions get processed. A rational miner will only mine subsets of the mempool that don’t decrease the total reward

I don’t know which mining software is rational versus altruistic

Core or pools, but I’m curious now.

Tentative analysis suggests that both are exhibited in the wild, but that piece is preliminary and may be wrong

iirc the core template generator makes some shortcuts. tries to be rational but may get it slightly wrong at times

most pools ask core for a template, though some may not

if you are looking at old blocks, core used to be much more altristic or arguably just 'dumb'

What tests can mine blocks with randomx?

Some core tests do.

There's a PR that adds tests specific to it, but it's broken on current master and I've not debugged yet,

Does any among some core tests mine long chain of randomx blocks in a row?

Yes, about that long.

The PR I speak of mines more than a rx period and re-mines it (that's the part that broke though).

Ah cool, thanks @smooth. I didn't know if they implemented from scratch or not.

In that case, runs like this suggest that the core software is probably altruistic

Since there were no rational blocks for an hour

color = block size

(blue for small blocks, yellow for large blocks)

It stops adding txes when the blockreward starts decreasing by more than some small threshold.

Hmm. The threshold is 0. So it really tries to get the last atomic unit.

Though it considers txes linearly, it doesn't try to fit different (smaller) ones at the limit.

Did the threshold get set to zero in v8?

Cuz a lot of the jitter disappears around 1685555

However there is still *some* jitter, as seen in these very recent blocks:

Looks like it as 1% till january 2017.

was*

Also, it may pay off to include a bit more tx data and get less block reward if it makes the block size go up so you can fit more for a bigger reward later. That's what the threshold was meant to do (but the value was pulled out of my... hat).

Isthmus is that reward+fees?

@koe yes/essentially.

Calculated as total payout minus total fees

its a cowcow

Photographer Nina Leen has an extensive series of photos of cows wearing hats

@moneromooo makes sense RE leaving room for block size growth

Though eyeballing from the second to last plot, the altruistic donation looks closer to 10% than 1%

If the leeway is now set to 0 atomic units, we can conclude that the outlier blocks in the last plot (very recent data) are not constructed by the core software, right?

wait isn't total payout minus potential block reward the better metric?

That'd be ideal, yes.

I was just working with column operations in SQL

so subtracting SUM(fees) from the payout is muuuuuch easier than implementing the whole reward calculation

I'm really time constrained these days, but happy too share data/access with anybody who wants to whip that up

ah fair enough, it can get a bit involved especially nowadays with block weights

Argh I was just closing a bunch of tabs, and a coworker walked past right as my browser flipped to a page full of cows wearing hats 🙈

hard at work or hardly working? :p

does the core implementation optimize for total payout? what's the function/algorithm used to select transactions for inclusion?

I'm curious about this too

i found this code once

once upon a dream

gingeropolous im gonna take a nap, will you dream-mail me the code reference?

is there a guide on how to set up a monero testnet?

im utilizing it to test my new consensus rules and see if it actually works

Insight: do you mean configure a node to use the testnet?

Insight: github.com/moneroexamples/private-testnet

Yes, on github. I think the person has a nick with "emil" in it.

ah, thank you.

Is there anyone around that knows Borromean ring signatures and would mind helping me through some areas? Specifically, if the key we know is the last one in its group, doesn't that conflict with e0?

Also, the diagram given in the paper is for (P0 | P1 | P2) & (P0' | P3 | P4) . What does it mean by P0' ?

the full algorithm is here web.getmonero.org/library/Zero-to-Monero-1-0-0.pdf

moneromooo: not sure if you are reading on Github but ph4r05 will fix core tests on Monday