Great video explaining how probabilities of winning a block actually work, and what to look for in case you think farming isn’t working

@storage_jm published an absolutely awesome video that everyone should watch!

14 Likes

I would really like to see the graph in that spreadsheet with the following:

  • 100 TiB farmed.
  • 1 day time farming.
  • Netspace at 250, 500, 1000, 2500, 5000 PiB.

I think the basic explanation is useful, but I’m not sure if it’s a good representation of your real chances of winning unless you’re plotting at the exact pace of the netspace growth. Seeing 5 graphs with the netspace growing and a fixed farm size would help to imagine what the graph would look as time goes on.

The example at 5:30 showing 5 days farming isn’t realistic because in 2 days the netspace will be much larger and the graph will shift making your odds of winning much smaller. The formula to figure out your real odds of winning without matching netspace growth is going to be much more complex. Does anyone know how to figure that out?

1 Like

Here is a great tool from Chia.Garden very similar to the one from @storage_jm :

https://chia.garden/win-probability-calculator

(thanks @Alex)

6 Likes

sorry for the crappy audio quality, was testing some new mic settings (and forgot to record audio the entire first run of the video). So for longer periods of time I have a netspace model that I use in similar fashion to chiacalculator.com, I just sent them my model to see if we can come up with something…who knows that is going to happen with price (these things break all models)

5 Likes

We should first model how these things break models :laughing:

2 Likes

Could you share the link to the spreadsheet that was in the video?

@storage_jm do you have that link handy?

1 Like

good old sheets

4 Likes

I’m certain There’s a massive variable not accounted for…: with any model out there.
My expected time to win with 50tb is right now…
6 days…
it can’t be a bug.
Iv only followed the GitHub in its entirety to the letter.
I am Linux savvy…
I haven’t looked at the chia code itself. yet.(prolly way over my head)
But Something.
With my setup. BY THE BOOK.
Is defy the “numbers/probability”
I’m running the latest.
Does anyone know what EXACTLY. Determines the number in the GUI. And cli. That gives u the
Expected win time…
What are the chances of it reported “incorrectly”
My Logs look phenomenal… and frankly defying laws of Whats known as -speed within a private network.
I do offten get 10 plots passing filter…constantly… with proofs being found. With my 50tb setup….

Statistics is incredible. I’m a big fan and everything is okay with Chia, I guarantee xD.

But just to blow out Your minds: let’s consider, that we have expected time to win a block: 2 months. And let’s keep it simple, so the network doesn’t grow - it’s stable. Now let’s say, that You haven’t won a block for 2 months. Do Your expected time is shorter? You haven’t won for 2 months, so it should be much shorter! But it’s not true, and expected time after no win for 2 months is exactly another 2 months. So after 2 months of no block, your total expected time for a block is 4 months. It is, because every draw is completely random and independ. Just sayin :smiley:

You’ve been prattling on about this for 6 days by now, have you won any XCH?

Number of plots passing the filter means nothing. Not when you’re touching the same plots multiple times.

You won’t tell anyone you’re set up, but I think it’s similar to mine. I have cloud plots that I am farming with two sets of harvesters. So I know that the metrics don’t understand that, and they double count the space. But the Chia network does know, if and when I find a block twice at the same time, I will only get paid once. You will too, and you’re real expected time to win be calculated from your 50TB