Admittedly I’ve not read the attack specifics, but thats not how I imagined it would work, needing to plot 512.
I imagined you’d wait to receive that info of the required proof to pass that filter, then create a plot with that proof within it.
3 Likes
I have to say that after you wrote it, I am really lost how it works. Reading that attack chapter again, I am not sure whether I understand how all that works anymore.
The most confusing part for me is that based on that chapter, even if you are able to produce such plot, that doesn’t mean that you are winning that challenge.
1 Like
Thats correct, you can create the plot that passed the filter, but that doesnt indicate that you won the plot. When that happen, the idea is to move to K33 or k34.
1 Like
Indeed your correct, it doesn’t mean you will win every time, but it makes you at least 512 x ish more likely to win every challenge, as you’ve cheated the initial proof.
https://www.chiablockchain.com/2021/04/28/winning-process-explained/
2 Likes
Thank you for that doc! Good start but looks like I need to read more of the original docs.
When you can plot fast enough (you just need to complete phase 1 in under 30 seconds)
Then you can make a plot on the fly that passes the plot filter each time.
Passing the plot filter just means that this plot now actually is competing to win a block.
(Normally, for every time a challenge is issued 1 out of 512 on average of your plots has a chance to win. This is done to reduce the I/O load on the nodes/disks)
So you can look at it like this: A plotter that can produce sub 30 second plots, essentially replaces 512 plots.
This is where the economic argument comes in; the cost of having 512 plots is really quite low. Fat chance there will be a machine that can plot that fast for the same investment.
If you could make it so that it wins a block every time, then it would be quite different. But this is not the case. You are just bypassing the plot filter.
Before moving to K33 or K34, the first knob Chia has to dial on therefor is the filter. For example, if they reset it to 1/256 plots, suddenly your “Asic” is only worth 256 plots, so it would need to be half price also to make it economically viable.
Reducing the filter setting will increase the I/O load on the nodes though.
That is the balance act that has made the K size 32 and the plot filter 512.
5 Likes
Thank you for the e planatiin and this is completely diffrent info that I was lead to belive from another source. I appreciate that and perhaps we can have a YouTube video explanation for other people who are confused as I was. Let me know if you are interested.
Bye