Topic: Help with research questions (Read 468 times)

Hello folks,

thanks for the replies.

blockchain.info has time-series data on hashrate and difficulty, but only on bitcoin and no altcoin. I would have to use difficulty from the blockchain.

On the issue of multipools: I am interested on how individual miners collectively converge to a specific coin over time or how a large collection of individuals react to shocks. Now, multipools strip away the individual decision making part of this study because small number of analysts make a decision for possibly a large number of miners. My question is how frequent are these multipools? are they popular among miners? or miners still make individual decisions?

merged mining would also have the same effect, because the miner does not have to decide where to invest his resources: he can use the same recourse on all coins. So how popular is merged mining?

Seconded. Find and chat up some multipool operators, and you'll learn a lot. Wrt looking at the halvings: Litecoin is near one now; you'll see that in this board's threads on the coin. From those threads, you'll get some worthwhile members for further questions.


I believe the polite translation-word for the above is "stochastic."

I favour merged mining. So there is no moving hashpower from coin to coin, there is just adding more coins into one's merge.

When DOGE showed that Litecoin was so pathetically insecure that just some stupid meme could whip up almost overnight more hashing power than Litecoin had, I first urged them to implement merged mining then, when they refused, I dumped all my litecoins. And did not buy into DOGE, because who knows, the next stupid meme might be a "lets destroy all the scrypt based coins with the superior hashpower we memesters can whip up" meme blowing away DOGE as easily as DOGE could, had it been a destroy litecoin meme instead of a compete with litecoin with yet another crappy clone meme, have destroyed Litecoin.

Now that one can merged mine them I am re-thinking, heck maybe scrypt ASICs might be worth looking into if enough coins can be merged that between them all they can maybe attract over half the planet's scrypt hashing power and not lose it overnight to some stupid meme.

-MarkM-

The blockchain of each coin has a history of the difficulty target on each block. That that should -more or less- reflect the hashing power put on each coin. Difficulty is not perfect beacuse it takes some time to adjust to the variations on actual hashrate, but it's the best data there is.

Look at multipools to see how miners switch coins automatically.
You may also want to look at reward halvings to see effects in price and hashrate.

But if you want to understand the markets, you are doomed. It's just gambling. Sometimes manipulated, sometimes random, but always gambling shit.

markm and Nxtblg, thanks for the great insights.

From markm response, it sounds like delay in reacting to market trend contributes to exacerbating the very same market trend, either rising or dropping. This is good for me, because I would still consider it in the basket of "herding effects", rather than attackers (i.e. pump and dumps). I am mainly interested in studying the herding effects, so I am very concerned about capturing trends that are due to pump and dumps. Based on your experiences, how often are these deliberate pump and dumps and do they ever result in big shocks?

I have another question: I would like to study how miners migrate from altcoin "a" to other altcoins after extreme shocks in altcoin "a" price. In other words, I want to understand what strategies miners adopt for distributing their computing resources across different altcoins. What are some  accepted strategies in the community regarding distribution of your computing resources?

Ideally, I would like to have access to a data source on total computing resources spent on each altcoin over time. I could also use a proxy for the total computing resources such as hashing difficulty. Do you know of any such data source?

Thanks in advance.

That's a damn good point.

eamanj: You should look into those strategies and run them in a "sandbox market" with random data to see which ones of them have a positive-feedback effect. Any grad student in a field that would be interested in this here jungle should have the analytical tools to do so. 

A lot of folk jump on things that are already rising in price, thus buying them higher, especially when you also consider they often lack the patience and/or the faith in exchanges to leave funds on an exchange in the form of a buy offer waiting to be taken up on the offer, so they buy directly from the sell side instead of placing orders on the buy side.

Also many of them have regular jobs, so maybe cannot respond instantly. A lot might snatch a glimpse at the internet during the workday, tell themselves damn gotta react to that real soon, maybe also already have plans that night so might not even get to it that day, so there can be quite a bit of delay in reacting even without taking into account most exchanges want some number of blocks/confirmations before crediting deposits.

Thus often when something goes up folk not only don't say to themselves damn should have bought that when it was low, oh well maybe I will buy it when it goes down again, but instead jump into the bubble; then it takes them as much as a day or few to actually get liquid on an exchange ready to buy; by which time profit-taking is starting by folks faster than them; so often after all that hassle of digging up old daemon and waiting for blockchain to update so as to send buying-coin to an exchange to buy with and waiting for the deposit of the coin you're buying with to confirm and be credited to your account, the price is already rapidly falling.

Remember that folks holding the coin are in same delay boat, they see their coin is going up so want to dump it but, like the buyers, they take time to get them to the exchange, because they mostly have not left them on the exchange on the sell side all along ready to automagically profit from the skyrocket. So often by the time they get them ready to dump they missed the peak.

Then instead of saying to themselves oh well that took too long maybe I should always have funds on the exchange to buy with (and buy stuff that hasn't started to skyrocket yet, placing sell orders on sell side to be first in line to profit when the skyrocket happens) they go like damn well that was too much hassle to undo, heck with taking my funds back off the exchange and heck with leaving them there, I'll just dump them while I am here since I do not want to go again through the hassle of digging up my wallet catching up the blockchain and sending to exchange.

I saw that a lot with DeVCoin, it would go up but dumpers would take so long to get coins to an exchange to dump them that prices would not be good by the time they dumped, but they would dump anyway after going to all the trouble of actually getting the coins to an exchange.

EDIT: Part of the problem might be trading strategies one finds in books, things like moving averages and such, which wait for the sign of the slope of the curve to change (upward to downward or vice-versa) before acting instead of predicting inadvance what ought someday to rise and buying it ahead of time to be already in position when it happens. Problem being such strategies assume you are ready at the instant to react, maybe with automatic triggers. When it might take you a day or few to react, you become a sucker reacting not a canny trader getting in as near the peak or trough as technical trading (math such as slope detection) can figure out what to get in or out of. If it will take hours or days to react, don't bother, you will tend toward too late. By the time a balloon starts to rise it is damn near too late unles you are awake and ready and liquid; the thing to do is buy stuff that is as close to rock bottom as you can find, but that you have good reason to elieve will take off again some day, and get your sell orders already in place on the sell order-books long before it actually does start to take off.

(TL;DR buy things that are dropping, sell things that are rising, not the other way around; and place your orders on the buy or sell side instead of buying directly from the sell side or selling directly to the buy side. Sure the exchange could run off with all your funds but that is only an issue at startup, once you have withdrawn back your capital so that all left on the exchange is already pure profit all they'd be stealing is some of your profligate profits, none of your initial capital.)


-MarkM-

Hello,

I am a graduate student starting to work on a research project related to fluctuation of BTC and altcoin prices. I am mainly interested in understanding the dynamics of the big rises and drops shortly after a new coin is introduced to the market. I have noticed that majority of altcoins follow this pattern at some point during their lifetime. I am aware of the existence of pump and dump schemes, but from what I have gathered pump and dumps usually happen within a period of hours whereas the big rise/drops happen within a period of days. I appreciate any insight into the following questions:

1- Are these big rise/drops really pump and dumps? or there are actual herding effects going on without the existence of attackers (pump and dump players)?

2- I need historical data on total hashing/compute power that is put into every coin per day or alternatively hashing difficulty per coin per day. Do you know of any source I could get this data?

Thanks in advance