 Show Posts
|
|
Pages: [1] 2 3 4 5 6 »
|
|
Hi all,
I'm working on a hobby project to basically set up a solar powered raspberry pi with a couple of usb miners on it. It'll have a couple of 10w solar panels and a charge controller with a battery pack sufficiently sized to continue operations for an hour or two should a cloud pass overhead ... which it does a lot here in Denmark :-)
Anyway, as far as I can see there are four different Raspberry models. I'm looking at the first generation B+ model because it uses a bit less power than the 2. generation. Downside is that it is single core vs. the quad core 2. gen.
With these self contained usb miners, does it really make a lot of difference how much cpu power you have on the Raspberry? Am I better off saving a few milliamps of power by buying the 1. generation B+ model, or will I be paying for that later on because it mines a bit slower? Will a single core be enough to fully take advantage of one or two or perhaps even ten avalon nano's?
And yes I know that this thing will never pay for itself. It's just a hobby project to give me someting to play with in the weekends ... nothing else :-)
-Michael
|
|
|
|
|
If "private mining" becomes unfeasible/unprofitable, it kind of defeats the original intent of having bitcoins distributed fairly evenly, and not under the control of one or more "power hubs".
Once all coins have been mined the entire mining economy will come from transaction fees. Kind of difficult though to see how private miners will see this as an attractive way to spend their kWh, unless there is a profitable upside. I wonder if the transaction fees will skyrocket in the future to provide incentive for private miners. Having a diverse and distributed mining community is vital for bitcoin I think.
-Michael
|
|
|
|
You should not include an exchange rate increase in your profit calculations. If you pay 0.04BTC (or the equivalent in another currency) for a device that will only produce 0.01BTC (after operating costs) in its lifetime, then it is not profitable. If you want to bet on bitcoins increasing in value over other currencies, then just buy bitcoins. This assumes your only motivation is profit.
I was going to write a bunch of clever calculations but you are right. It is actually more sensible to just exchange the $15 to 0.03 btc. But if you really wanted to make a profit with a mining device of some kind, how would you be able to gauge the performance? Like "price per ghash per watt < x"? Did anyone make an online calculator for that? -Michael |
|
|
|
Hello all, Many people will say that buying a block erupter is a bad idea because at the rate the network difficulty is rising, they will never produce enough bitcoints to break even. I had an opportunity to discuss this with a friend today, and the fact is that you can actually still make money with a low yield device like a block erupter, so long as you pay for it in some form of fiat currency. This is true because of the rise in value of bitcoin against these currencies - a trend which is likely to continue. For example, take a place like this (randomly picked, no affiliation). Buy a block erupter here for $15 and put it to work today. According to the profit calculator at bitcoinx, the unit will produce 0.00033 btc/day at the current difficulty level. At $326/btc that's 0.10758 per day, which gives a ROI of 139 days. But, some say, difficulty increases will make the unit unable to ever break even. Let's say the unit produces the promised 0.00033 btc/day for just one month. The total sum produced would then be 0,0099, or roughly 0.01 BTC. You paid $15 for it, so if the BTC/USD rate just swings up to $1500 you have in fact broken even. These are super conservative numbers, it is realistic to assume the unit will produce 0,03 if you let it run for a year I think. It is also likely that you will be able to buy them for $10 rather than $15. With these numbers the BTC/USD rate in one year will only have to be 334, which is pretty much where it already is (my guess is that in one year, BTC/USD will be close to $1000). So, as long as you buy units with fiat currency, it should still be possible to not just break even but make a healthy profit with lowly block erupters. What do you think? -Michael |
|
|
|
|
DisplayName: ohlsen
UserID: 158779
Thanks!!
|
|
|
|
I don't get it. Where's the number? |
|
|
|
As a JD Investor, lowering the max profit decreases variance. For instance betting your life savings with 51% is stupid. Betting your life savings in 100,000,000 bets, with a 1% edge, will put you ahead by a large amount, with little variance, in total gain.
You keep returning to this variance thing. What you call variance other people call luck. There is no such thing as variance, just the house edge and a nice bell shape called normal distribution. Betting your life savings over 100,000,000 bets and betting it all on a single bet carries the exact same risk and reward. You risk less by wagering less on each bet, but if you sum up all those little "risks" across all those millions of bets you end up with the same as just making one single, large bet. In your terms, placing many small bets means you have to be lucky and win 50,000,000 times. Placing a single bet means you only have to be lucky once. -Michael |
|
|
|
This is very important to understand. Most players go bust because there is nothing stopping them from betting their entire bankroll in one hit. Nakowa will take longer to bust because he cannot in any way risk 5,000 on 2x payout at once.
I'd like to challenge that idea. If your purpose is to go from 5000 to 10000 on a casino with a 1% house edge, can it be shown - statistically, that it is a better idea to play 1 unit 5000 times at 2x, or 5000 units 1 time at 2x? Forget about what your intuition is telling you and look at it from a mathematical point of view. You have precisely the same risk when you are making 5000 2x bets with martingale as you have if you make a single large bet. Slicing things up into smaller bits makes the risk "less" in your mind, but you forget the fact that you are making so many more bets that combined they make up as big a "risk" of losing your bankroll as that single large bet. This guy nakowa could just as well have gone to 10000 by placing that large 5000 2x bet, it would have saved him time (for the same outcome, win or lose the bank). The only reason I think he didn't is because of the house cap on bet size. -Michael |
|
|
|
What we have here is a situation where someone appears to be actually beating the house edge over many thousands of bets. And this is intriguing, because from a statistical point of view this should not be possible. You can't win a coin toss came if the coin is balanced towards the house, no matter what system you use.
-Michael
It's called variance. Are you saying that if I toss a coin 1000 times, I won't - on average, hit 500 heads and 500 tails? If you are saying that, then there can't be a 50/50 chance of heads and tails. Both statements can't be true at the same time. And no I'm not saying that if you play 3 times you couldn't win 3 times, that will surely happen. But at JD we are not talking about 3 bets, we are talking about thousands of plays where the house edge seeming doesn't provide the house with any edge at all. This "whale" seems to be playing martingale with 100 btc bets. Martingale is *not* a winning strategy long term, otherwise anyone could walk into a casino, sit down by the roulette and gain financial freedom. Eventually red will come up 12 times in a row, and you will lose your bankroll. So how is this whale winning long term by using a martingale strategy? It's certainly not because of "variance", the number of bets and the casino bankroll drawdown is just too consistent and too deep. This looks more and more like someone found some statistical exploit in the way lucky numbers are generated based on nonces, hash values and transaction id's. It looks deliberate, not just like a guy with a really big wallet or a lot of luck. -Michael |
|
|
|
We profit 76 % of the time over 1,000 bets - not 50.8 % of the time like with 1 bet.
THAT'S being an investor, not a gambler.
Also - has anyone noticed that despite the 1 percent edge JD has never made a 1 % profit? It was about half that - before the current fiasco - now we know why.
Variance. Lets lower it by refusing big bets.
I'm sorry, but this is the biggest load of BS I have ever read. Clearly you have not done the math. If you make 1000 bets with a 1% house edge, you will win 490 times and the house will win 510 times. The house will not win 76% of the time. Variance means nothing as far as house edge. Playing many tiny bets and playing one huge bet provides precisely the same house income over time. As a gambler, you will have no advantage by placing smaller bets and winning and losing small amounts than by placing a single large bet. Over time it's just a matter of scale. A casino runs the risk of losing the bankroll only if they don't cap the size of their maximum bets. Because obviously there's only a 51% chance of winning that bet, and thus a 49% chance of going bust from the casino's point of view. But JD does have these limits. Noone is placing bets of 1000 BTC or more. The max win is 1% of the invested capital, which makes good sense. There is no reason to limit that further. What we have here is a situation where someone appears to be actually beating the house edge over many thousands of bets. And this is intriguing, because from a statistical point of view this should not be possible. You can't win a coin toss came if the coin is balanced towards the house, no matter what system you use. -Michael |
|
|
|
Actually I remember reading an article a long time ago that said the way to make $ at any table is to bet the max and quit, win or lose.
I actually wrote a simulation in QBASIC for that and it worked.
This is what's happening here.
No. Playing once is not a winning strategy. If it is, you'll have to show the math that proves it. -Michael |
|
|
|
We are not being killed by huge martingale bettors, we are being killed by someone who bets 100 BTC on 49.5% 12 times and wins 11.
Quite the contrary. You *want* these people in your casino. Because in a second, he's going to bet 12 times and lose 12 times. That's what the house edge is all about - Martingale is a losing strategy and the more martingale betters you have, the better. Why do you think almost all the online dice houses have "auto" modes that specifically cater to Martingale betters? Because these stupid people think they can work around their statistical disadvantage using a top secret system they bought from some internet marketeer. Look at how the balance is decreasing. It's slow and steady, like satoshis are seeping though the creeks somehow. -Michael |
|
|
|
|
I tried primedice and found the site to be very nicely done, but slow and unreliable. Might have been during a DDoS attack, which seems to be happening to all the dice sites constantly.
Right now I'm with Just-Dice but I fear for the site because someone appears to be consistently winning and have drained the bankroll for an amount in excess of 6000 BTC over a few days. Very nice interface, more reliable than primedice.
coinroll.it is what I want to try next -- very nice interface as well, and they offer API gaming which is what I need. Only downside is the relatively high minimum bid (x1000 compared to most of the other dice sites).
letsdice I just discovered, and it looks very nice and seems to have a very healthy bankroll. House edge is higher than 1% though, although most of that is going into a jackpot so it evens out the odds long term I guess.
- Michael
|
|
|
|
|
