Medium term forecast for US$ / BTC price.

[organofcorti]

After a number of posts about expected Avalon ASIC earnings I realised that in the medium term - within a month or two - the US$ / BTC price is likely to fall, or at best level out. This has nothing to do with market players or bubbles, and everything to do with mining.

1. When a constant increase in network hashrate occurs, difficulty retargets cannot completely negate the increased number of blocks solved.

For example, in the chart below (estimated increase in hashrate from various sources) by July this year the expected blocks per hour for the last retarget is ~ 160 blocks per day.

A constant increase in blocks solved will increase supply substantially, and since these devices are electricity efficient the increase in network hashrate is mostly limited by the cost of ASIC devices - so we can expect the increase in network hashrate to continue for significantly longer than a few months.

2. Since the cost of the device is a limiting factor and in the short term a higher device hashrate will lead to significantly larger earnings, many miners will pay much more for a hashing device than they would have for a GPU last year.

This means that many miners will obtain investors or bank loans in order to purchase, for example, a mini rig, when they might have been previously satisfied with a few gpus. These larger hashrate owners will need fiat currency to pay for electricity - which although small compared to the near term btc earnings will still be a substantial sum. If miners have taken loans and investors they will need fiat currency to pay for bank loans and reimburse any investors who want to be paid in fiat.

Smaller hashrate miners who may have purchased their device with their own cash and can afford to pay for electricity using other revenue sources may hold on to their btc, but based on other analyses I've done (unpublished) I estimate that as was the case prior to ASICs, the top 20% of hashrate owners will account for 80% of the network. Smaller miners that hold btc will affect the price less  since their total btc holdings will be smaller.

The over-supply of coins in point 1 will be exacerbated by larger hashrate miners regularly selling a large proportion of their btc earnings.


tl;dr  In the coming months, the price is more likely to drop or stabilise than continue to increase. If you are purchasing a high cost ASIC it would be best to assume lower USD BTC prices rather than higher.

[01BTC10]

There will be no increase in blocks solved.

[organofcorti]

There will be no increase in blocks solved.

Yes, there will.

[Bowjob]

I'd prefer stability than a drop, tbh. I bought an ASIC with my own money, I plan on holding the BTC for a while, then again it was just ~$2000 dollars.

[01BTC10]

There will be no increase in blocks solved.

Yes, there will.

Bitcoins are created each time a user discovers a new block. The rate of block creation is approximately constant over time: 6 per hour.

https://en.bitcoin.it/wiki/Controlled_supply

Maybe after 4 days of winter skiing/camping and some wine I'm a bit slow. I thought supply should remain the same even if hashrate is growing.

[organofcorti]

There will be no increase in blocks solved.

Yes, there will.

Bitcoins are created each time a user discovers a new block. The rate of block creation is approximately constant over time: 6 per hour.

https://en.bitcoin.it/wiki/Controlled_supply

Maybe after 4 days of winter skiing/camping and some wine I'm a bit slow. I thought supply should remain the same even if hashrate is growing.

No, but it is a little tricky.Point 1 covers it:

1. When a constant increase in network hashrate occurs, difficulty retargets cannot completely negate the increased number of blocks solved.

This is because difficulty is calculated using the average hashrate since last retarget, so while the network hashrate constantly increases, the difficulty calculated at each retarget will not be sufficiently large. It will always be behind until either the added hashrate slows or the increase in hashrate as a percentage is negligable.

[dree12]

There will be no increase in blocks solved.

Yes, there will.

Bitcoins are created each time a user discovers a new block. The rate of block creation is approximately constant over time: 6 per hour.

https://en.bitcoin.it/wiki/Controlled_supply

Maybe after 4 days of winter skiing/camping and some wine I'm a bit slow. I thought supply should remain the same even if hashrate is growing.

No, but it is a little tricky.Point 1 covers it:

1. When a constant increase in network hashrate occurs, difficulty retargets cannot completely negate the increased number of blocks solved.

This is because difficulty is calculated using the average hashrate since last retarget, so while the network hashrate constantly increases, the difficulty calculated at each retarget will not be sufficiently large. It will always be behind until either the added hashrate slows or the increase in hashrate as a percentage is negligable.

In addition, if hashrate more than quadruples, even the retarget will not be enough to slow blockrate.

[organofcorti]

There will be no increase in blocks solved.

Yes, there will.

Bitcoins are created each time a user discovers a new block. The rate of block creation is approximately constant over time: 6 per hour.

https://en.bitcoin.it/wiki/Controlled_supply

Maybe after 4 days of winter skiing/camping and some wine I'm a bit slow. I thought supply should remain the same even if hashrate is growing.

No, but it is a little tricky.Point 1 covers it:

1. When a constant increase in network hashrate occurs, difficulty retargets cannot completely negate the increased number of blocks solved.

This is because difficulty is calculated using the average hashrate since last retarget, so while the network hashrate constantly increases, the difficulty calculated at each retarget will not be sufficiently large. It will always be behind until either the added hashrate slows or the increase in hashrate as a percentage is negligable.

In addition, if hashrate more than quadruples, even the retarget will not be enough to slow blockrate.

I don't think it would be a big problem for very long. If the hashrate (eg) increases sevenfold, then the the next retarget would be two days later. If that rate continued there would be seven retargets in two weeks.

[Dargo]

I'm shocked, a drop prediction that is actually based on something that makes sense. I'll just mine and hold with my ASICs, but I'm also small-time with only $2K invested. Bigger players might sell as they mine since it would be riskier to hold. So I think this idea has some plausibility - will be interesting to see if this does have an impact.   

[organofcorti]

I'm shocked, a drop prediction that is actually based on something that makes sense.

Yes, not actually being able to provide an estimate of the loss and the exact timeframe is a bit of a giveaway, eh?

I'll just mine and hold with my ASICs, but I'm also small-time with only $2K invested. Bigger players might sell as they mine since it would be riskier to hold.

This is what I think will happen for the mojority of miners. But the majority of miners will contribute a minority of the network hashrate and earn a minority of the coins.

The small number of large hashrate miners will contribute a majority of the network hashrate and  would have more significant effect, and as you mention they're the ones who will be more risk averse.

So I think this idea has some plausibility - will be interesting to see if this does have an impact.    

Yes, hard to say. If something happens to massively increase the general usage of btc, the increased miner-side supply could be much less significant.

[Bowjob]

^ Wouldn't Mt.Gox switching to coinlab impact anything?

[organofcorti]

^ Wouldn't Mt.Gox switching to coinlab impact anything?

Maybe, but I have no idea how to account for it.

How many miners are also holders who traded off a large portion of their holdings, which now needs replenished, in an attempt to be an ASIC early adopter?

I think you're on to something there. The model I use assumes a constant 3.2 Thps added per day; if only half of that is purchased using bitcoin that's ~ $30000 demand per day for bitcoin. That could easily soak up the excess.

I'll have a think about it and post an update later. Well spotted, Holliday.

[Dargo]

^ Wouldn't Mt.Gox switching to coinlab impact anything?

Maybe, but I have no idea how to account for it.

How many miners are also holders who traded off a large portion of their holdings, which now needs replenished, in an attempt to be an ASIC early adopter?

I think you're on to something there. The model I use assumes a constant 3.2 Thps added per day; if only half of that is purchased using bitcoin that's ~ $30000 demand per day for bitcoin. That could easily soak up the excess.

I'll have a think about it and post an update later. Well spotted, Holliday.

Yes, good point by Holliday. If I had purchased a few BFL SC rigs with BTC early in the pre-order, I would be pissed as hell about unloading all those BTC, which are now worth over 5x what they were when I spent them. So I might not be looking to part with those new mined coins right away. Personally, I spent about 50 BTCs for two Jalapenos (later converted to a half-single order by paying the difference with Paypal), and it's bad enough to think that those BTCs now worth over $1700 USD. In terms of current BTC price, the half-single is effectively costing me over $2000 USD vs $1300 for the full single paid by bank wire, lol. Just glad I didn't use BTC for the whole purchase, which I was considering at the time.   

[Global BTC]

I'm shocked, a drop prediction that is actually based on something that makes sense. I'll just mine and hold with my ASICs, but I'm also small-time with only $2K invested. Bigger players might sell as they mine since it would be riskier to hold. So I think this idea has some plausibility - will be interesting to see if this does have an impact.   

That's an interesting point - not all newly mined coins go into the market. Are ASIC miners more or less likely to hold on to their mined coins than FPGA miners or GPU miners? ASIC miners have made a larger investment. Maybe they need their money back? This would suggest they should also be less likely to hold on to their coins.

[piramida]

So people who invested half a year ago in a very risky hardware that has only one purpose, at a point where 1 btc was worth 10$, did so on their last lunch money, waited half a year patiently and now they desperately need the cash back because they don't believe in this project? yes, extremely plausible...

[cedivad]

So an increase of 3600BTC/day worst case of available coins, not on sale coins, on a market that does at least 36kBTC every day, should do something?

[organofcorti]

So an increase of 3600BTC/day worst case of available coins, not on sale coins, on a market that does at least 36kBTC every day, should do something?

I think you misread the OP.

I was assuming in the OP that the excess coins would be on sale. A continual surplus and large hashrate miners selling coins in order to pay fiat debt could have a significant effect on the market.

Holliday's point was that, some miners will want to replenish their previous BTC holdings. I extrapolated from that comment that the 3.2 Thps per day added to the network would cost a considerable daily sum, and if a significant proportion of that is paid using bitcoin then this could reduce or negate the effect of the surplus.

Do you not think that the mechanism by which btc become available has an effect the price?

Also, the worst case scenario is not 144 blocks of surplus coins per day. Depending on the amount of hashrate added it could be more than that.

[cedivad]

Do you not think that the mechanism by which btc become available has an effect the price?

Also, the worst case scenario is not 144 surplus coins per day. Depending on the amount of hashrate added it could be more than that.

I think it has, but not at the level you are predicting. What are 100 or 1000 bitcoins on sale on a market that is at least one order of magnitude bigger, and sometimes 2?

[organofcorti]

Do you not think that the mechanism by which btc become available has an effect the price?

Also, the worst case scenario is not 144 surplus coins per day. Depending on the amount of hashrate added it could be more than that.

I think it has, but not at the level you are predicting. What are 100 or 1000 bitcoins on sale on a market that is at least one order of magnitude bigger, and sometimes 2?

On one day, probably not much. But consistently every day? If they are not purchased, then the surplus will build up quickly.

Besides that though, right now I don't think it will have a negative effect on the market, for the reasons outlined a few posts above.

[phelix]

(log scale, ignore green curve)

Please note that did not happen with previous hashrate jumps. Now the impact is much less because of larger coinbase and lower reward.

Also in you diagram the most additional coins are generated right now. the price did not drop today.