Does the number of connections affect farming?

Chia Farming & Harvesting
1

Apart from the number of plots, what does winning depend on?
Does the number of connections affect farming?

2

As far as I know, winning has something to do with having quality connections. You really won’t know a connection you have is of good quality or not.

Also, Chia winning is basically, 1/512 plots pass filter every period (10 secs?). If your plot passes filter, then it is 1 of the 512 competing to complete the transaction the fastest. Whoever turns in the complete the fastest gets the 2 XCH. Winning would depend upon a decent internet connection and computer fast enough to serve a completed transaction faster than the other completes; but people are winning on a Pi which is relatively slow.

People keep asking what the magic “do this and you’ll win” is. As far as I know, there isn’t any setting you can change to start magically winning. There is no way to keep winning if you are on a streak. You just have to be lucky I guess and winning is the luck of the draw.

2 Likes
3

in response to the fastest wins… not true

if I submit a Proof with quality A, and you submit a Proof with quality B for the same signage point… and both are within accepted quality needed by the network difficulty… then both of us wins a block and XCH… doesn’t matter who was first as long as within the 30 seconds window from when the challenge was issued.

You see multiple blocks per signage point and also no blocks… this is normal… the difficulty will make it so over 24 hours the target number of blocks is created… and then difficulty is recalculated.

3 Likes
4

in response to if your plot passes filter then you win a block…

passing plot filter is only 1 step.

  • the challenge needs to be in the plot

  • the plot quality score needs to be good enough
    then a Proof can be submitted

  • a Proof needs to pass verification (header values OK etc)

  • a Proof needs to fall within the network difficulty

  • a Proof needs to be received and checked within 30 seconds of challenge issued.
    then a Proof participates in creating a block

so if you create a Proof then there is a chance you win a block, but not always.

1 Like
5

You need enough connections to be at the correct height all of the time and to be able to communicate your proofs when you find them to the rest of the network in time - technically you only need to have one ‘good’ connection for that, but practically you want some redundancy (a lot of redundancy).

I know people do run huge farms on rPis with external USB, but I don’t - I haven’t had great experiences with external USB disks. Other than making sure you have enough connections, watch challenge response times like a hawk, I average 0.2-0.3sec, and IMO anything over 1 sec needs an investigation.

1 Like
6

The consensus I thought was that any plots consistently pulling over 0.5 seconds needed looking at to determine if that plot is “bad” or if there is a system issue. If there are a lot of plots going consistently over 0.5 seconds, then there is a system issue, at least that is the way I look at it.

I don’t have to worry about any plot going over 0.5 seconds as all the ones I checked are well under.

1 Like
7

Challenge response is not something you can check once, it is not plot specific (although it can be disk specific). It is much more to do with how loaded the system/controller/disk/port/bus is at the time it’s looking for a given challenge.

You need something like chiadog set up IMO, which will warn if challenges are more than 5 sec - but I also have a log grep pattern looking for anything 1 sec or more, usually that correlates to a time when I’m moving files between disks, or once when I had a failing disk that was making a whole SAS controller run slow.

You need INFO logs enabled, and to look at these lines:

2021-06-25T21:11:29.957 harvester chia.harvester.harvester: INFO X plots were eligible for farming H... Found F proofs. Time: T s. Total P plots

All. the. time.

2 Likes