Write NiceHash bot, make lots of profit

JoeTheMiner wrote:

Steve Sokolowski wrote:Good morning!

We're proud to finally document our WAMP API, which allows users to obtain live statistics about the status of the pool. In the coming days, we also plan to provide methods that provide insight into users' workers and balances.

The first methods are now available at https://prohashing.com/help.html#api-wamp.

For the first time, these methods make it possible for a user to write a "NiceHash bot." Since NiceHash provides an API with all the necessary properties and methods, a bot can be written to obtain pricing at NiceHash, compare it to profits provided by the WAMP API here, and if profit here is greater, then buy lots of hashrate from NiceHash and direct it here.

If programmed conservatively, this bot should never be at risk of losing money. During times of high profitability, profits here often soar to 33% above NiceHash. That should never happen, because a bot like this can drive a significant portion of NiceHash's hashrate here during those times up to the point that NiceHash's rate equalizes.

Is anyone interested in writing this bot? We aren't asking for any cut of the profits that you earn; we just want someone to write it and run it to make themselves money so that we can take the pool fees. GenTarkin seems to have the expertise to make a lot of profit by doing this.

Hi Steve,

Just wanted to give you an update that I finally finished my bot tonight and have started it up. So you should start seeing more hashrate from me now!

Thanks,
Joseph

Remember to try "g=off" if you have problems with NiceHash penalties. That argument shouldn't be used by most people because it loses money, but for NiceHash, the losses of merge mining often are offset by lower penalties.

Steve Sokolowski wrote:

JoeTheMiner wrote:

Steve Sokolowski wrote:Good morning!

We're proud to finally document our WAMP API, which allows users to obtain live statistics about the status of the pool. In the coming days, we also plan to provide methods that provide insight into users' workers and balances.

The first methods are now available at https://prohashing.com/help.html#api-wamp.

For the first time, these methods make it possible for a user to write a "NiceHash bot." Since NiceHash provides an API with all the necessary properties and methods, a bot can be written to obtain pricing at NiceHash, compare it to profits provided by the WAMP API here, and if profit here is greater, then buy lots of hashrate from NiceHash and direct it here.

If programmed conservatively, this bot should never be at risk of losing money. During times of high profitability, profits here often soar to 33% above NiceHash. That should never happen, because a bot like this can drive a significant portion of NiceHash's hashrate here during those times up to the point that NiceHash's rate equalizes.

Is anyone interested in writing this bot? We aren't asking for any cut of the profits that you earn; we just want someone to write it and run it to make themselves money so that we can take the pool fees. GenTarkin seems to have the expertise to make a lot of profit by doing this.

Hi Steve,

Just wanted to give you an update that I finally finished my bot tonight and have started it up. So you should start seeing more hashrate from me now!

Thanks,
Joseph

Remember to try "g=off" if you have problems with NiceHash penalties. That argument shouldn't be used by most people because it loses money, but for NiceHash, the losses of merge mining often are offset by lower penalties.

Steve,

The bot I wrote didn't work out too well and I think the problem lies in the way I calculate profitability. Should I be using the "percentile_btc" or "btc" value returned from wamp? From what I noticed it seems it is profitable sometimes but then it becomes unprofitable rather quickly. Any ideas?

Thanks.

JoeTheMiner wrote:

Steve Sokolowski wrote:

JoeTheMiner wrote:
Hi Steve,

Just wanted to give you an update that I finally finished my bot tonight and have started it up. So you should start seeing more hashrate from me now!

Thanks,
Joseph

Remember to try "g=off" if you have problems with NiceHash penalties. That argument shouldn't be used by most people because it loses money, but for NiceHash, the losses of merge mining often are offset by lower penalties.

Steve,

The bot I wrote didn't work out too well and I think the problem lies in the way I calculate profitability. Should I be using the "percentile_btc" or "btc" value returned from wamp? From what I noticed it seems it is profitable sometimes but then it becomes unprofitable rather quickly. Any ideas?

Thanks.

Use "btc." Percentile_btc is the 95th percentile earnings over the last 200 responses. It's not an indicator of current profitability, but is used to display the charts at the top of the page so the progress bars can be drawn.

"btc" and "usd" are the average earnings for the last 100 shares. This is much more accurate than the previous value returned by REST because it is sent to you instantly. One of the weaknesses of this, though, is that even though 100 shares is a very short period of time, it's possible that earnings were very high during the first 50 shares and very low during the last 50 shares. This happens when a coin sharply decreases in value or increases in difficulty.

In general, if there is a huge spike in value, like to 25 cents, it means that a single block of some coin has become valuable for 10 seconds. You'll lose money because you can't place and cancel your order quickly enough. Look for opportunities when profitability increases just enough to make money instead, because slight increases are more likely to last for many blocks.

But I think the most likely issue is that the act of renting hashrate changes this value. If you bring in a lot of hashrate, it can drop profitability because you yourself cause changes in the markets.

Try doing this with very small orders and see if that makes a difference, to rule out that large changes in hashrate are causing the problem.

Thanks for the help. The bot has been running several days now after I modified it seems to be working much better now.

Steve Sokolowski wrote:

JoeTheMiner wrote:

Steve Sokolowski wrote:
Remember to try "g=off" if you have problems with NiceHash penalties. That argument shouldn't be used by most people because it loses money, but for NiceHash, the losses of merge mining often are offset by lower penalties.

Steve,

The bot I wrote didn't work out too well and I think the problem lies in the way I calculate profitability. Should I be using the "percentile_btc" or "btc" value returned from wamp? From what I noticed it seems it is profitable sometimes but then it becomes unprofitable rather quickly. Any ideas?

Thanks.

Use "btc." Percentile_btc is the 95th percentile earnings over the last 200 responses. It's not an indicator of current profitability, but is used to display the charts at the top of the page so the progress bars can be drawn.

"btc" and "usd" are the average earnings for the last 100 shares. This is much more accurate than the previous value returned by REST because it is sent to you instantly. One of the weaknesses of this, though, is that even though 100 shares is a very short period of time, it's possible that earnings were very high during the first 50 shares and very low during the last 50 shares. This happens when a coin sharply decreases in value or increases in difficulty.

In general, if there is a huge spike in value, like to 25 cents, it means that a single block of some coin has become valuable for 10 seconds. You'll lose money because you can't place and cancel your order quickly enough. Look for opportunities when profitability increases just enough to make money instead, because slight increases are more likely to last for many blocks.

But I think the most likely issue is that the act of renting hashrate changes this value. If you bring in a lot of hashrate, it can drop profitability because you yourself cause changes in the markets.

Try doing this with very small orders and see if that makes a difference, to rule out that large changes in hashrate are causing the problem.

Steve,

Is there a way to get the profitability of the individual coins the pool is mining instead of the global profitability?

Thanks,
Joseph

JoeTheMiner wrote:

Steve Sokolowski wrote:

JoeTheMiner wrote:
Steve,

The bot I wrote didn't work out too well and I think the problem lies in the way I calculate profitability. Should I be using the "percentile_btc" or "btc" value returned from wamp? From what I noticed it seems it is profitable sometimes but then it becomes unprofitable rather quickly. Any ideas?

Thanks.

Use "btc." Percentile_btc is the 95th percentile earnings over the last 200 responses. It's not an indicator of current profitability, but is used to display the charts at the top of the page so the progress bars can be drawn.

"btc" and "usd" are the average earnings for the last 100 shares. This is much more accurate than the previous value returned by REST because it is sent to you instantly. One of the weaknesses of this, though, is that even though 100 shares is a very short period of time, it's possible that earnings were very high during the first 50 shares and very low during the last 50 shares. This happens when a coin sharply decreases in value or increases in difficulty.

In general, if there is a huge spike in value, like to 25 cents, it means that a single block of some coin has become valuable for 10 seconds. You'll lose money because you can't place and cancel your order quickly enough. Look for opportunities when profitability increases just enough to make money instead, because slight increases are more likely to last for many blocks.

But I think the most likely issue is that the act of renting hashrate changes this value. If you bring in a lot of hashrate, it can drop profitability because you yourself cause changes in the markets.

Try doing this with very small orders and see if that makes a difference, to rule out that large changes in hashrate are causing the problem.

Steve,

Is there a way to get the profitability of the individual coins the pool is mining instead of the global profitability?

Thanks,
Joseph

Unfortunately not. The reason is that there isn't a simple definition of coin profitability.

The profitability of a coin is dependent upon factors specific to each miner. They include the miner's work restart time, the pool's orphan rate, how many blocks are immature ahead of the current block, what proportion of the network we are, the actual block time vs. target block time of the network, the spread on the exchanges, the number of exchanges, whether there is a market in some other coin than bitcoins to avoid the high BTC fees, and so on.

It isn't possible to provide a simple profitability figure that a person could use to make decisions without knowing a huge number of factors. That's one of the reasons we have performance problems and cannot open multiple servers across the world, because of how much information needs to be known and computed. The figures you see on the website are estimates of how much the miners assigned to those coins, who are the best miners on the pool for the task, will make. The average miner's take would be much less, and since miners are so different, there is no miner that is the "average."

Ok makes sense. Thanks for your help and input.

I'm a Computer Science student who recently got heavy addicted to the whole mining world.

Any way I can take a sneak peek of the bot you wrote JoeTheMiner?
Sounds fascinating, but my level of skill is not quite there yet to code it myself. Would love to learn from it!

TwinToe wrote:I'm a Computer Science student who recently got heavy addicted to the whole mining world.

Any way I can take a sneak peek of the bot you wrote JoeTheMiner?
Sounds fascinating, but my level of skill is not quite there yet to code it myself. Would love to learn from it!

Unfortunately not at this time. I would suggest learning how to call an API from the programming language you are most comfortable with. Then learn how NiceHash operates with their bidding system and learn/test how much hash rate to rent, what time, and what orders to bid over. Brainstorm the best step by step approach to run every so many seconds to make your bot profitable. Translate those steps to the programming language and test it. Keep modifying it as needed until you have found what works. It takes a lot of time and effort but in my opinion it is worth it. Wish you the best of luck!

Hi Steve. Anyone get this thing going? I'd like to use it if it works.