Today I sold all bitcoins I had accumulated during the past 2 months (about 0.110 BTC in total). Slightly less than half of that (0.05 BTC) was earned from mining via NiceHash. The remaining 0.06 BTC I bought at a whim in the beginning of August at €2,297.30 per BTC. I sold the entire amount via Coinbase today at €3,514.49 per BTC.

I did this for a couple of reasons, which I am about to share with the world. It was a probably stupid thing to do, yet… hear me out.

Reason #1: I recently bought a couple of GTX 1070 cards to mine faster

The prices of all AMD and most Nvidia cards went crazy by May 2017, and stocks rapidly diminished. However this situation began to improve by the end of July, at least as far as Amazon was concerned. So I quickly snatched one Gigabyte Nvidia GTX 1070 G1 8GB GDDR5 PCI-E to see how it would perform. Without spending too much time tweaking the overclock settings, I was able to achieve about 29 MH/s in Dagger/Hashimoto while using less than 120W (+100 MHz GPU / +533 MHz memory / 70% power limit) while staying below 65° during the hottest two weeks on recent record.

I liked the Gigabyte GTX 1070 G1 so much that I promptly ordered another one. However this meant I had to spend more than €850.- (ex-VAT price) in a really short amount of time. What if Bitcoin crashed quickly after the post-SegWit2x exuberance faded away? What if the newer and faster AMD Vega GPUs poured on the market and made mining with last-gen cards more difficult and less profitable? What if…?

Reason #2: Don’t lose sight of the main objective

When I started my own Bitcoin experiment, I assumed that whatever I was doing, it would most likely incur losses. Since the beginning I told myself I shouldn’t be interested very much in making profits, but that I should do my best to keep these losses at a reasonable minimum.

By buying two GTX 1070 boards this month (after the GTX 1080 I got back in July) I went further into loss territory than I was willing to accept. I had to do something to fix this.

When I sold my coins, I was able to off-set nearly the full cost of one of these new video cards. This meant that if I sold them both on the second-hand market at 50% off, I would cover all losses for these two. I was also willing to write off the cost of the GTX 1080, because I play a lot of games and I would consider it paid off over a period of 2-3 years regardless of any Bitcoin income.

Reason #3: GPU mining may not be dead just yet

I’ve been doing this for a really short while (a bit more than 2 months) and from Day 1 all I heard was ‘GPU mining is dead’. However this still does not seem 100% true.

First of all, the temporary shortage of high-performance video cards seems to have held back the increase of mining difficulty and GPU mining is still profitable. Secondly, the AMD Vega 64 (which was initially reported to achieve more than 70 MH/s) is actually much less efficient at about 30 MH/s (while using 350W…) so it does not seem to be that much of a game-changer. AMD said recently their plans do not rely on GPU mining as a long-term growth factor, and Vega proves this.

Which means, these humble GTX 1070 boards churning at low power will likely remain profitable for at least another couple of months.

Reason #4: Higher mining difficulty means more demand for hashing power… right?

The simple supply/demand model suggests that whoever has been buying hashing power so far will need to buy more if they don’t want to fall behind. As these buyers don’t have hardware costs to recuperate, it seems reasonable for me that they would be willing to accept smaller profit margin while buying more MH/s.

Reason #5: The BTC price surge must eventually stop

Nobody has been able to really explain the insane growth Bitcoin has achieved since its mid-July crash. From a fundamental point of view, present exchange rates really don’t make any sense. So by selling my coins, I leave this price uncertainty fully behind me.

I will now continue to mine with greater rigor. I started doing this with a single GTX 970 and now I have ~6.5× more compute power 1In terms of GFLOPS. It is not possible to compare hash rate directly because different GPUs favor different algos. at my disposal. Also, I have relative confidence that GPU mining still makes sense.

Why am I doing this when I disregard the price, you may ask. And my answer is that I am not interested in the dollar/euro value of Bitcoin. I am interested in Bitcoin itself, and I strongly believe it has intrinsic value. If Bitcoin price keeps rising, then all is good. If it crashes soon, then I won’t mind so much, because I no longer have much skin in the game.

Bottom line

In conclusion, I did this whole thing in order to rebase my operation. The situation with cryptocurrencies is very fluid, and it makes little sense to try and predict the long-term future. I think I did good during the first two months; I will now try to do the same for two more. What will happen afterwards? In November we will be going through the hard fork that was locked in on August 1st. There are still fears that the transition will not go smoothly, and this may cause another crash in the price of Bitcoin. I can now patiently wait and see.

 

Does any of this makes any sense to you, my dear reader? What would you do if you were in my shoes? Regardless of whether you thing I am right or dead wrong, let me know in the comments below. I will appreciate fresh perspective.