LOL! Someone definitely made a profit during the outage. In a 30 minute period, 350TH was bought on the market, bringing the price back up to 8400.

The good news – for now – is that 4 days from now the difficulty is expected to drop by about 7%, which will likely decrease the maintenance cost for some time. (source: https://bitcoinwisdom.com/bitcoin/difficulty). Also, when the S9 hits the market, I think people want to get rid of their S7 in favour of the S9, so I believe the S7 is going to be cheap as hell while still proftiable to use. I bought my lot around 8400, if I can sell it for 9000 before then, I'll do it, otherwise I'll just hang on to it until it goes out of profit.

The bad news obviously is that due to the diff decrease, S7 will be even slightly more successful, which may postpone the S9 getting introduced in Hashnest even further.

No, it's not. The price was 14000 before the halving, when it happened, it almost immediately dropped to around 7000, and after that stabilized at 8500. This is classic market behaviour. Draw up a fibo, and you'll see it's well within range. I'm a bit surprised, I would expect it to drop a bit more, as the maintenance percentage doubled from 37.5 to 75%, one would expect the price to at least stabilize at 7000.

Also, hashnest is still selling s7 units for considerably more than the market GHS value, on the same maintenance plan. Now, if they were out to scam you, why wouldn't they simply get rid of the market and sell you shares of those machines for the advertised price?

I agree that the maintenance fee on hashnest is relatively high, but I see it as paying a premium for a trustworthy (and according to the 101 thread the only really legit) hash farm in the universe and surrounding areas.

Are they steep? Absolutely. Are they stupid (for losing their domain?) I'd say so. But calling them scammers just because they're steep, a bit stupid (and apparently because of your unsubstantiated claims that they manipulate the market) just makes it look pretty obvious to me that you don't know what you're talking about.


Oh, COOL!!! Do the miners still work? Then it's probably because of the outage, and I might take on a risky investment.

OK, I made a few changes:

• I now use the numbers from the Zpool API. Until I understand what each number means, I take the lowest amongst all 4 numbers calculate with that, whichever it is.
• I disabled all coins except for Sha, Scrypt and X11, coins with decent amounts of orders on NH. L2v2 would be next as things look now, but I leave it disabled for now.
• I changed my config:

I still need some help with this one:

My setting is purely based on guesswork, I have no idea what kind of impact the diff will make, so I haven't got the slightest clue as to which number I should put in, and how I can calculate that. What do you think is reasonable?

I am actually doing that: 135% or more = buy, 125% or more is go ahead, 115% or less = throttle down and decrease as fast as I can, <115% for more than 15 minutes = sell. I decrease the price as often as I can without losing substantial hashrate, and if my hashrate is too low I increase to a level where it substantially grows while still making at least 125% profit. Also, right now I set it up so that I don't buy hashrate that can't complete an order in 24h. (That is, it looks for the NH price of enough hashrate to mine 0.01 BTC worth at that same price point, in 24h or less.)

As for the buy/sell hashrate fees: I know. It seems to be like this:

If you buy hashrate, you pay almost 4% fee. (That is, buy 0.01, get 0.0096). BUT: Once you remove an order, you get 3% of the remainder refunded. So in short, if you buy 0.01, you pay 0.0001 + 3% of what you use.


I'd really appreciate that! One advantage is that you can get those numbers in less than one second, whereas the page I used consistently takes 20 seconds to load.

I think if I use 140% (like ax66 suggests) I could use those numbers, even if they're a bit more volatile.

I still don't think 140% would get me to break even though, or I should fine-tune my sell vs decrease policy.

I read almost the whole thread (yes!). I scripted a Zpool-based nicehash-order-manager, but I've lost quite a bit of cash, and I'm having a hard time figuring out what I'm missing.

Here's what I'm trying to do:

1. I pull the average rewards from ZPool algos from THIS PAGE. (Why that page? Those numbers seem more conservative and less volatile than what Zpool reports on its site, and I've read somewhere that ZPool has a tendency to have its averages a bit off.) I look at all numbers (i.e. current, last 2h, last 24h) and take the lowest as the 'right' one, just to be a bit conservative.

2. I pull the orders from Nicehash to see what hashrate goes for.

3. If the reported reward is more than 135% of the selling price of some substantial hashrate on nicehash, I buy the hashpower and point it towards zpool.

4. I poll the rewards every 60 seconds, and if they drop such that profit gets below 115%, I limit the hashrate on NH to the minimum, and if it doesn't come up after 15 minutes, I remove the order. When it rises again to above 125%, I set the limit back up to my calculated maximum (see below).

So, basically, I would think that 3% nicehash mining fee, 1% buying fee, 2% pool fee and 5% delta (spilled hashrate) adds up to 11% potentially loss to cover for, so I think the 15% is a reasonable selling-threshold.

I ran this script for about a day, I think it picked up about 10 orders, and while I was hoping to make some profit, I only found out that I lost about half what I started with. Bollocks!

But seriously, I'm a bit dumbfounded right now. So, what am I missing?

Here are the decisions I made that I'm uncertain about:

1. d=<max>. And by max, I mean the max in the NH allowed range. I found out I had to specify it when Zpool gave an out-of-NH-range number, and when I set it to minimum, my hashrate delta was like 40%. I set it to the highest possible to prevent pool flooding, and now the hashrate delta stays nicely <5%. Is this totally idiotic? And if it is, how can I calculate what diff value I should provide? (As obviously I can't trust zpool to set them in NH's range)

2. I limit the orders to consume up to 35% of ZPool's total hashrate, and I have seen orders get there. I initially didn't think this could be a problem, but when I initially did 50%, I saw the rewards dropping significantly a few minutes later in some cases, so I thought those might be related to one another. I found out that I could take about a third without brining about major impact on the rewards. Maybe I'm wrong here though. Should I have only taken 10%? Or just a solid fixed amount based on the algo I'm trying to do?

3. I connect from both the EU and the US servers at Nicehash, whichever gives me the best rate. But EU-CA obviously has way more latency. Is that something to keep in mind, and should I just stick to Westhash?

4. I'm having trouble understanding the numbers at https://www.zpool.ca/api/status. For example: Estimate Last 24h: 0.00052364, Actual Last 24h: 0.5686. The difference in order of magnitude, is that just GHS vs THS? And is the estimate made 24h ago, or is this really an estimate of the next 24h? Also, what does rental_current mean? Is that the reward based on the last mined coin? And last, but not least, what attributes to the difference in the ZPool API and the page I mentioned above? (And which one do you guys think I should be using?)

5. Could it be possible that this loss can be attributed to a single algo? If so, is there any way I can still figure that out without having to test every damn algo again? The vast majority I've mined is X11, I've had multiple of those orders, all very profitable. I've also seen orders on Lyra2v2, a bit of Neoscrypt and an odd SHA256, it was highly profitable for about an hour today. All on the above policy. (Although I might stop doing Neoscrypt, because at times the minimum limit is already something like 20% of Zpool's hashrate)

6. Anything else / more numbers I should keep in consideration?

Sorry for all the things I don't quite understand yet! I'm passionate about learning all this stuff.

Thanks for the compliment! I would love to brush up on whatever relevant stuff they taught me those 35 years ago. Unfortunately, as a CFO of a successful e-commerce business, I can't find the time...

I don't remember exactly which phase in my education taught me that based on such a small sample (both blocks and hash rate in my case) spending more than 60 seconds on a calculator could be considered a waste of time, but I remember that as a great time of my life, and every time I encounter a great mind that is still looking forward to that stage of development, I feel warm and fuzzy again.

On topic: I'll make a more in-depth attempt when I know a few more reliable numbers. And because you're such a fine fellow, I'll throw in some tasty percentages.

Cheers!

I did a small, simple rudimentary calculation.

I'm a small-time cloudminer. (Really small-time.) I have 78 GH in Hashnest S7 power. Now here's what I see:

With antpool's current hashrate, in the last 3 blocks (the first 3 after the halving) I got paid BTC 0.00000294 per block. This breaks even on maintenance costs for my hashpower when a block takes about 60 minutes. Given the fact that on the current hashrate Antpool reveals a little over 30 blocks per 24 hours (consistently over the last 4 days), according to blockchain.info/pols, my preliminary conclusion is that mining using s7 on Hashnest is still profitable, albeit very marginal compared to what it used to be before the halving.

P.S. I believe this is not a function of the total hashrate, because if the hashrate drops, obviously the amount of blocks per day will drop as well, but the earnings per GH should increase proportionally, as there are less miners in the pool to share with. Obviously this changes when the difficulty increases, but given the fact that S7 miners are supposed to be amongst the most efficient ASICs, I really don't see how the difficulty would increase dramatically after the halving, as I believe a lot of miners will stop hashing.

I also remember reading rumours about the S9 getting into Hashnest early next week, but I can't find the reference. I believe it was somewhere on reddit. (I understand, as this is my first post on the forum, I don't have a lot of credibility to make a statement like this, but let's just say I am not associated with Bitmain/Hashnest in any way.)