Topic: HashFast announces specs for new ASIC: 400GH/s - page 388. (Read 880461 times)

No, it's the opposite. Once we hit equilibrium, pretty much NO ONE will make money and those who do are the ones who have free electricity and cheap time.

I feel like the moral of the story is that no one should buy mining hardware until we hit mining hardware equilibrium where it is readily available for anyone who may want it.

I'm batch two with no MPP...

Not even batch 2 Avalons! My batch 2 Avalon has only made 70 coins and it isn't likely to make more than 1 or 2 more through the rest of its life.

Yes, but this is the world of bitcoin hardware where nobody has ever delivered on time ever (though KnC did come impressively close I will admit) where if you think about all the possible lost BTC you could have theoretically mined you surely go insane.  I'd be worth millions by now if BFL had delivered in 2012 like they had claimed but they didn't, and I'm not a millionaire but that's just how it goes.

Don't buy mining hardware assuming you'll be able to get whatever you paid in BTC back, I don't think any hardware has managed this except batch 1 and 2 avalons. 

what a pleasure to see someone receiving what a company promised to deliver in time they promised to deliver...   happy christmas !  thank you hashfast(slow...) ... thank you knc !

This.

They screwed things up on multiple levels with their marketing. I still believe that the engineering side of things is trying to do the best worldly job possible to fix the issues - but you can beat a horse only so much...

With their mining equipment they sold an expected amount of future bitcoins for a set amount of USD, which should be around the exchange rate they quoted when providing the contracts. The statement of John wrt BFL in an interview with Neal Blackburn (thank you by the way) is appalling when viewed in hindsight: https://www.youtube.com/watch?v=5yj-mmnRWYc&t=185

The good news is that with the increased purchasing power of the bitcoin, the cost for a lawsuit becomes negligible. As such I second the assessment made here:

Think of it from a business perspective. It is more beneficial for HF right now to sacrifice Batch 1 customers at the altar as they will reimburse batch 1 in US $; gained 9x the value of their BTC just because BTC value exploded. And on top of that they won't have to send the MPP chips (using our machines and extra chips for the other batches). We, the batch 1 customers, were just the cow giving away our milk for free. This is why HF "is taking it slow" and making sure they won't ship on time. Too many complications and pissed off customers on Batch 1 that will be looking for any issue with their rushed assembled machines when receiving them. Might be too many problems with shipment too during x-mass time and potential loss machines or broken/defectives ones.   

Red flags we should have considered right from the gecko: HF accepting to be paid just in BTC or dollars but not via Paypal or CC so you have no recourse complaining to Paypal and Visa protection (because of the non delivered product) and requesting arbritration on their TOS right away.

HF is an incredible fraudulent marketing company. They knew from the beginning that they will never ship on the 20-30th october. It was just a strategic date against their competitors. Then they wrapped this lie with an MPP that they will never have to ship and that doesn't concretely exist as all they have to do is just reimburse $$ after 31 dec. What a profitable and smart move!

They are playing game theory and calculating their odds and ruling for their best option, that is sacrificing us!

Thank you HF the grinch for a totally ruined X-mas!

No truer words have ever been spoken.

For a limited time only I will share the secrets of my pickaxe revenue boosting system.  We all know that mining gold is getting harder and a pickaxe in January is probably only worth $3 per gigarock.  Honestly it is probably worth less than that but gold miners have gold fever and you can find enough people willing to pay $3 per gigarock in January.  Still is it your fault that you can't delivery until January and thus are limited to only $3 per GH/s (probably 1000% gross markup)?  Now some honest suckers might say "sorry you are going to have to live with "only" $10M in gross revenue for a couple months work".  You know what I say .... that is stupid honest sucker talk.  Don't listen to it for a second.  I am going to show you how you can double the $10M into $20M without delivering more or earlier.  Just free bonus money.

See the first thing you do is promise to deliver the pickaxes in October and mark them up from $3 per gigarock to a much more profitable $14 per gigarock.  Now you might be confused because you already know your pickaxes won't be ready until January so how can you promise October delivery?  It is what we in the creative pickaxing business call "telling your customers what they want to hear to get the moneyz", I have heard that honest suckers might call it lying, or fraud but I wouldn't know. Now this plan isn't for everyone, if you are looking for an honest profit where both you and your customers benefit then you should probably stop reading because this plan isn't for you.  Also if you are in that category want to buy some cheap pickaxes?  

Ok are all the honest suckers gone now?  Good now lets get down to business.  So instead of selling pickaxes in January for $3 per gigarock the revenue boosting secret is to pretend you will have them in Oct and offer them for $14 per gigarock.  We will "limit" this to half a million pickaxes.  We want to make it seem limited.  Nobody want to pay a premium and be at the end of the line.  Now if you get lucky some large group like IcePicking ("always bet on Ice") will do a lot of the work for you and buy about half of those early pickaxes.  Now just sell the other half retail.  Now some of those customers might be cautious so if sales are slow then give away some of the pickaxes for free as "protection".  Hang on trust me it will be worth it, these "free" pickaxes aren't going to make us a fortune at the expense of the customers who trust us.   See if the customers don't trust us they won't buy regardless but if they trust us but are worried about difficulty then the free pickaxes will seal the deal. So reserve 800K pickaxes for the PPP (Pickaxe Protection Plan).

Now you should know that after all this plays out as you know it will, your rep is ruined, so lets double down.   Your customers might not know you can't deliver on time but you do so don't hold anything back. You might as well offer "Dec" pickaxes you can't deliver either!  I mean if you are going to defraud one batch of customers why not defraud two for twice the profit and twice the fun?  So sell off the rest of your January pickaxes as "December" pickaxes and offer a deal of only $8 per gigarock (for a $3 per gigarock pickaxe).  See it is a "deal" it isn't as horribly bad as the first batch?  

Now stop a second and just think about the beauty of it.  We only have one box of January pickaxes.  That is it, yet with some creative pickaxing we have now sold two batches at different price points.   Some customers probably agonized over which was the better deal and they are all coming out of the same box!  ROFL! I mean all we have are January pickaxes, some just have an Oct sticker and some have a Dec sticker.    See who says you can have fun making millions?

So lets punch this into the ole pickaxe calculator. The honest promise and delivery in January at a reasonable price nets us $10 million in revenue, but with some creative pickaxing we can still ship the same number of pickaxes at the same time but DOUBLE OUR REVENUE! Don't believe me?  Here is the creative math.  Honest (sucker) January delivery: 3.3 million pickaxes @ $3 per gigarock = ~$10M.  Creative Pickaxe Super Sales System:  0.5 million (promised for Oct) * $14 + 0.8 million * $0 (protection plan promised when we feel like it) + 2 million * $8 ("cheaper" batch 2 promised for Dec) = ~$20 million.

Both methods sell out our 3.3 million pickaxes.
Both methods don't ship until January (or maybe a few days prior).
The creative one makes twice as much money.

Now do you see the magic of creative pickaxing?   Well the bad news is if you want to get technical there is no such thing as something for nothing; this is just a wealth transfer from your customers to you under false pretenses (lawyers might call that fraud).  They never had a chance at getting units before January so $10M extra for "early" delivery is just a $10M reduction in net profit.  With mining margins as tight as they are that will be enough to push them into the red.  See they are paying $20M for pickaxes which really are only worth $10M because ... and this is the funny part ... they actually believed you when you said Oct and Dec delivery.   What a bunch of suckers.  

Of course the only downside is you have to deal with "whiners" who get mad when their Oct & Dec pickaxes doesn't arrive until January. You can't really blame them I mean the high priced pickaxes only make economical sense for the easier Oct rocks.  $14 axe in Oct = profit.  $3 axe in Janaury = profit.  $14 axe in January = profit (well profit for us and massive losses for them).  That is why the Pickaxes were only worth $3 per gigarock in January, but if we had sold them as January pickaxes we would have only made $3 per gigarock and that just isn't fair.  It isn't fair other companies were faster and thus were able to charge more.  I mean $10M with 1000% gross markup how many yachts can you buy with that.  The funniest part is many of your victims did the math and made all the right calculations.  They never would have paid $14 per gigarock for a January pickaxes because they KNEW it would be a loss.   Still with the ole switcheroo you can trick them to.  They have now absolutely no hope of even breaking even.  If you start to feel bad about all the loses your customers are taking and the fact that it was completely avoidable by being honest (and accepting a fair price for a fair product), just buy another sports car, or snort some more premium blow off an expensive hooker. Nobody said creative pickaxing was going to be easy, but a "bonus" $10M should help your guilty conscience.

So , If I'm getting it right , the most profitable way to mine is to sell your pickaxe!
Of , great !

... yup .. that's true, ,
they need batch one customers to keep their goal , and now they screwed us..  
NEVER EVER HASHFAST !!

I agree with Cevadid analysis upthread.  Essentially the MPP will be +400% hashing power but hashing power alone is irrelevant.  What matters is hashing power / current difficulty.  If difficulty growth averages ~27% per adjustment then difficulty will roughly quadruple between 1 Jan and 1 March.  So 4x the hashrate for "free" but at 4x the difficulty means the additional hashrate will produce ~ the same amount of BTC (before electrical cost which is now 400% as much) as the original unit did.

TL/DR the MPP will probably roughly double your lifetime revenue (whatever that might be).



In Cevadid numbers it comes out to ~9 BTC for a BabyJet and ~8 BTC more for the MPP in lifetime revenue.
https://bitcointalksearch.org/topic/m.3932737

Now like any projection the output is based on the input so a couple caveats:
a) It assumes you have your Batch 1 rig delivered on 01/01/2014. A delay means less revenue and an "early" delivery means more revenue (~0.2 BTC lost/gained per day before/after 01/01).

b) It assumes you have your MPP delivered on 03/01/2014.  A delay means less revenue and an "early" delivery means more revenue (~0.2 BTC lost/gained per day before/after 03/01).

c) It assumes that difficulty growth will average 27% per adjustment at least in the short term.  In reality difficulty can't keep doubling every month forever but so much of the revenue is front loaded that difficulty slowing 6 months out doesn't have much of an effect.  By Oct you are down to 0.1 BTC anyways so lets say difficult slowed to only 10% per adjustment and thus doubling every three months.  That would only add another 0.6 BTC lifetime revenue.  Still if you want to be a little more hopefully you can probably pad Cevadid's numbers by 1 maybe 2 BTC so say 20 BTC total for original unit & MPP.

d) The MPP is completed boards only.  So you will still need to supply a case/rack/frame/cardboard box, power supply, some controller (rPi, BBB, dedicated atom based system, or your home PC/laptopt would work fine), and cooling system (Corsair HydroCool is retail version of loops that HF is using).  So there is some additional expense, but the parts are "off the shelf" and shouldn't be too expensive.  I estimate maybe $600 (for 4 boards, 1,600 TH/s using 1400W or so) if you are a DIYer.  Not sure if HF will supply kits but I would just buy parts yourself and have them ship "naked boards".  I wish I could do that for the ordered product as well.  Hey HF how about you ship me just the boards on my Batch 1 order and refund $500?  I mean in BTC terms you still come out ahead (I paid ~50 BTC so 0.5 BTC refnd would be 1% ).

e) If you wanted to wash your hands of the whole thing the value of the MPP really just depends on what you think you can sell the boards for in March.  What will (ready to ship not pre-order) mining hardware be worth in March?  2 mBTC per GH/s?  Well that would be 1,600 * 2 /1000 = 3.2 BTC.  3 mBTC per GH/s? That would be 4.8 BTC.

f) Of course all this assumes HF can actually deliver, doesn't go bankrupt, or the chips have some as of yet unannounced flaw.  At this point nothing really would surprise me.

g) (only if you want to hope against hope and have your heart crushed by the HF sledgehammer again) technically the MPP should be delivered by in Feb not March.  HF has such a bad track record to date I think March ("only" a 30 day delay) is more realistic but difficulty is going up so fast if they actually got their act together and did something to ease the pain of their customers (which built the company BTW) a delivery on time would roughly double the lifetime revenue from the MPP (16 BTC vs 8 BTC) http://mining.thegenesisblock.com/a/0eea988786.  A delivery "early" when the ship Batch 1 order on or around 1 January would double that again to ~32 BTC.  If you consider 8 BTC for the original unit that puts lifetime revenue around 40 BTC and the closest anyone could hope for break even.  Of course HF won't be doing that but dreams are nice.


* since revenue is front loaded the first 3 months are the most important.  If difficulty grows ~27% for 3 months then slows down it won't matter much.  Difficulty can't keep growing exponentially forever but lifetime revenue is so front loaded the difficulty falling off or even flatlining for some time doesn't really matter if it is on the tail end of the curve.

I would rather refund my batch 1 order and maybe buy 400% worth and it would still be less than what you paid for any batch 1 user. Batch 1 users are screwed and hashfast knows it...

Well , I really love how you explained the whole picture.
Right now I  think that most of us if not all already accepted that It was a bad idea to buy from them.
Any thoughts of how the MPP might lick our wounds?

Well I hate to seem like an ass towards Batch 2 customers but that would be great news for Batch 1 customers after massively overpaying and being delayed months, having more than 2 PH/s from the same company crash onto the network right after they get delivery would just compound the loss.  If Cointerra (another 2 PH/s) managed to deliver at roughly the same time then difficulty for Batch 1 customers would skyrocket 50% to 60% the very next adjustment after they starting mining.  Think 0.2 BTC per day is bad?  Watch it crash to <0.1 BTC per day within days of when you start.

Still at this point I doubt batch 1 was ordered separately.  Here is why.  TSMC is a pretty good company, and they charge top dollar because of their reputation so when they give you a date they move heaven and earth to make sure that date sticked.  I have worked (indirectly) with them at previous employer (no I don't design chips I wrote custom software for high end networking equipment which use custom chips).  In Hashfast's case the time from tapeout to wafer delivery was 78+ days.   Rocket runs are usually 30-45 days especially for as mainstream as 28nm is.  75+ days is a slow for even a regular run (60 days is more common).  Foundry schedules aren't guaranteed but their rep (and billions of dollars in business) depend on timely estimates.  TSMC missing by a day or week is one thing.  TSMC missing by over a month on an rocket (engineering) run?  It just doesn't fit but then again 90% of what this company has said since launch no longer "fits".  

Still lets assume you are right.  Marginally better news for Batch 1 customers but utterly brutal for Batch 2.  If your right and the Batch 2 and MPP are part of a second run and that run wasn't ordered until now (after the chip bring up) well we are looking at two months until delivery of Batch 2, upgrade modules, and MPP.  Even if HashFast paid extra for a rocket run (note the claimed time from tapeout to delivery on the first run was 78+ days) you can figure 37 days (median between 30 and 45) for fabrication, another week for packaging, and lets be optimistic and say a week for assemble and ship more than 4 PH/s of hardware.  Throw in two days for delivery and you are looking at 46 to 53 days from order to delivery.  Add in a week of padding for problems, delays, component issues and two months looks like a pretty good estimate for a rocket run dropped today.  The bad news is rocket runs are 50% to 100% (depends on foundry capacity) more expensive.  Add another month to that for a normal run.

Semi OT but this is likely why KNC "Q1/Q2" for 20nm means "end of Q2 but we will throw in Q1 because we won't technically be lying because we say Q1/Q2 and it will boost sales but those counting on getting it before late Q2 are going to get burned".  For the record I have no hate for KNC they are the only hardware which after some DOA problems has performed above expectation, I just call em as a I see em.

I think its highly optimistic to think that batch 2 HF will get theirs same time as batch 1. Batch 1 was rocket run "engineering run" in their own words that just got here in nov. given the planning and project management seen so far by this company, i wouldn't be surprised if the order just went in for 2nd batch wafers.

indeed, in the end MPP is all that matters

Well the network is ~7 PH/s and KNC two batches in less than two months was >4 TH/s of that.  So the others are shipping but nothing like the volumes of KNC (and volumes planned by Cointerra & HF).  As the network gets bigger it takes bigger and bigger daily volume to keep growing at the ~30% per difficulty.  A slow rollout of 50 GH/s and 100 GH/s rigs isn't going to move the needle as much anymore.

Still HF customers had better pray that Cointerra is delayed at least 30 days after HF ships as that is another 2 PH/s that originally was going to be months and months after HF delivered and now could be roughly the same time.   Still even if Cointerra is delayed, batch 1 HF customers are still screwed as originally Batch 1 was Oct and Batch 2 was Dec.  That provided a nice gap between the 0.5 PH in batch 1 and 2.0 PH/s in batch 2.  Now essentially there is just one massive >2.5 PH/s batch.  Sure batch 1 customers might get their units a few days early (at 2x to 3x the cost) but when HF ships difficulty is going to explode upwards.  So either HF doesn't ship in which case batch 1 has no hashpower or they do ship in which case their difficulty is going up another 35%+.

To put it into perspective to produce the same amount of Bitcoins per day requires:
400 GH/s on Nov 1  (Original Batch 1 delivery date even padded 10 days)
908 GH/s today (if HF's BabyJets output 908 GH/s customers would be no better than getting 400 GH/s on time)
1,040 GH/s on Dec 31 (this assumes HF batch 2 is delayed 30+ days after batch 1 so it doesn't blast the difficulty)
1,352 GH/s on Dec 31 (w/ back to back delivery of batch 1)

This also assumes no more shipments from KNC this year and that Cointerra will be delayed 30+ days and the rest of the players continue their slow rollouts with no massive surprises (as in stepping on a landmine kind of surprise not a you just won the lottery surprise).

So pretty much at this point if HashFast converted all BabyJet's into Sierras and they were overclocked 15% it wouldn't be so massive bonus it would only be enough hashpower to give what was promised for Oct delivery.  Now is HF going to do this?  No but it illustrates just how screwed Batch 1 customers are.   Even IF HashFast upgraded all Batch 1 orders to Sierras, delivered all of them before 31 DEC, AND the chip operated 15% over spec, that would only be as good for Batch 1 customers as getting what they actually ordered on time (Nov 1).

I agree, difficulty explosion does not stop because we can't even account for half of the mining power out there..

Chinese companies still making chips. Bitmain, avalon. Plus KNC is most likely making their first gen chps for themselves to mine. Don't forget bitfury as well.  Diff will keep rising.