But what does it matter if no one cares and continues?

This might be a good way for people who just got scammed to blow of some steam, but that's about as far as it will get. And if that was the plan discussed in this post I apologise for ranting about how a blacklist in the sense of the word is a bad idea.

Put centralization in the protocol? With a link to a single entity that can render all coins on an address useless? As you've said that is definitely never going to happen: First the devs / foundation will never get convinced this is a good idea, after that no miner / node I know of would ever update to such a version, if it does a fork / altcoin is probably going to take over.


It appears to me that at least half of the discussion is about a list not enforced by the network but by the community (on top of Bitcoin itself).

• If a private key gets stolen in almost all cases the owner will find out after the coins have left the wallet, if sent directly offchain (like to Mt. Gox for example) are you going to blacklist the addresses of Mt. Gox?
• If this list is central in any way every goverment / legal body can attack the list, or poison it with extreme ease. If it isn't it is going to be attacked by basically everyone else because it is the power to control the whole monetary system.
• And if you have a wallet that has stolen BTC from a blacklisted address 10 transactions back, are you allowed to trade? Hell I can even steal BTC, send them to random people to have them blacklisted ?
• if I don't know about the list / don't got the most recent version / etc. and I meet up with someone and buy some Bitcoin I can't use them at all?
• Good luck finding a way on determining who scammed who in basically the whole scammer accusation forum.
• What if you later find out about coins that were stolen x time ago but are now distributed across the whole network?

This would only work when you get all exchange parties on board (including parties like silkroad), somehow deal with mixers / exchanges or other parties that don't want to be obey to a blacklist. After that you have to make sure the database won't get hacked / lobbyd / controlled / spoofed by governments (or IPSs).

And if you find a way to deal with all these problems you just introduced centralization to Bitcoin?

So it's an extremely bad idea and it can never be implemented fairly any way.

EDIT: I would suggest instead of a blacklist that if you are going to trade with anonymous people on the internet and prevent scams from happening in the first place, you use a trusting system like bitcoin-otc. And if your bitcoins got stolen you should have been more careful (it may be hard, but that is the way it is right now).

Well I think a trading bot following a strategy is way better than its human counter part. But finding a good strategy is pretty hard. I would suggest using the backtesting to verify that your current settings would have made profit in the past.

You can also use Gekko to do paper/simulation trading (with a bunch of settings simultaneously for example), this way you know whether a strategy is currently effective. Do note that you are always looking a the success after a while, in which case the success lies in the history and won't guarantee anything anymore for the future.

EMA is a indicator that you can use to ride the trend (buy low - sell high), however because it uses moving averages to verify the trend it is always lagging behind. This in combination that trends can reverse at any time makes it so that you don't always make profit while trying to ride the trend.

To answer your question: I am currently on a 10% profit measured in BTC in a little over one month.

I heard a couple of things like this, I'll look in to this.

--

I don't have as much free time as I had last month. I am still working on Gekko, however not on the daily basis I was before.

You are thinking about EMA in the wrong way, in a perfect world you would actually want all data.

The way EMA works is that the previous interval is more important than the the interval that came before it, that is because you are only interested in trends that are currently happening / might happen and not trends that happened in the past. If the bot would not look at any historical data it would not be able to calculate if there is a trend for after a while (gathering data from the start which becomes old data after a while).

If 2 people (A and B) would run a bot that doesn't look at any past data and A is running the bot for months and B just boots it up, only A is able to determine if we are currently in a trend and B wouldn't because you can't just look at the current price to say wether the price is going up or down (you need to compare with past data).

The beauty about EMA is that the relevance of each interval is exponential decreased. This means that all data points are reflected in the current EMA calculation, however they are getting exponentially getting less relevant. The number 144 is was in the free bot this bot is based on, and is a nice combination between having enough past data (having ALL past history would only help you a tiny bit in better calculating the EMAs, since old stuff of exponentially less important) and not spamming the Mt. Gox API to hard / having to wait to long for the bot to start.

There was a discussion in the old thread (for the free chrome browser bot this bot is based on) on what this number should be and it was changed from 20 to 144. In most other bots this is configurable, but you should never change this number except when you know that lowering this number will make your EMA less accurate.

That's fine, you can also print out a QR of the address.

Though I strongly advise either escrow or building trust with the other party first instead of going to a meetup with a big bag of cash.

Here is mine

http://cryptotrader.org/backtests/pojZycZDpDM62CZfQ

The goal of 100k USD is already reached, they still can use more money though.

I like the idea! I already created mine

Couple of questions:

• We submit a strategy that will paper trade for 1 week in a couple of days? (And you can optimise it by backtesting before that week)
• So: I am supposed to keep the link private until after 07/15 00:00 GMT? Am I allowed to change the parameters during / after the week? Else I just need one now and optimise it after the week?
• Since the first round is about EMA indicators I take it that it is only allowed to change the parameters and not implement another indicator?

If you are still running into trouble I would suggest updating node, v0.6.19 is really old and I think a lot of dependencies need a newer version.

You can run two gekkos (you just have to copy over the config and point one Gekko to the copied and changed config). But once you start Gekko node will load and parse the files and won't use the files on disk anymore at all (they are kept in memory). So with the exception of an error inside nodejs/raspberry or a really strange API response that Gekko cannot handle I won't see any difference in running two of them (unless you want to monitor multiple exchanges, etc.)

I just looked into this and Bitstamp has changed their API, Gekko asks for data in a format not supported anymore (old API, new API). This means that Gekko currently does not work on Bitstamp. I am looking into a fix and hope to resolve this ASAP.


Well the main reason is that the config is not solid, the settings may change and using an old settings file can result in Gekko crashing. You could place the config somewhere else (or rename like you did) and make a small script with `node gekko config=config.js.bak` to force Gekko to use another file as config. You can read more about that here, though I would advice against this.


Gekko has dependencies (like momentjs), you can install them by running `npm install` in Gekko's directory.

That's a bug in Gekko, if you could let me know some details (like are you running the latest version, what exchange and interval) I will look into it!


There are a couple of ways to run Gekko silently and log to a file instead of into the terminal, you can use tee if you are using a UNIX platform (AFAIK) like OS X or Linux. You can read some alternatives in Gekko's advanced features documentation.


I had this before but changed it to log only position changes (buys and sells) so that it's easier to read back a couple of days later. As suggested you can turn on debug mode to get a lot more info but you can also turn on verbose in the profitcalculator part of the config. This way you'll get an update every interval (hourly on default) about your current (simulated) profit.

Thanks! Just make sure to tune it in a way that it's profitable. Else it can be a pretty expensive toy.

I am still improving the data retrieving part of the bot (when I have some free time), but I think that the way Gekko asks data is sime time behind the websocket (the source Clark Moody is using). I'll look into this.


Good point, I was thinking about this the other day and I think that I will implement a version with an increment first and log all results in a CSV. Thanks for your thoughts! I'll put it on my todo list when I have some more time.

For a long term goal I was thinking about either particle swarm optimization or an genetic algoritm to optimize the parameters automatically. But at that time I want to have a lot more indicators and a way to use many at the same time (maybe neural network). But this will take a lot of time.


Gekko is named after this pretty rich guy. Profile pic is some other guy who can fly.

The public opinion about Clark's service is pretty good and I think the best way to persuade people into buying into your service is to proof that it is rock solid by letting user test it. It also depends a lot on the feature set I think. I would suggest a trial period or a freemium model.

Thanks! I changed the OP and the topic name to reflect that it hosts a list of downloadable candles (I also included yours, let me know if you mind). I've pointed to your tool for everyone looking for a calculator. I still need some time to check it out by the way.

The 5 minutes was on a VPS with an SSD, so that boost in IO probably did speed it up a lot.

When Gekko is backtesting it uses the close of candles both as the price to determine trends using EMA and the price at which to trade if a new trend is detected and it simulates a trade. The fee applied to the simulated trade defaults to 0.6% (the default at Mt. Gox with low amounts). You can configure this fee amount in the config.

Your node version is pretty old: you are running v0.6.19 while the current version is v0.10.12. I think that this is your problem. You can quickly update node by downloading the npm module n globally: `npm install -g n` and than run `n latest` and it will install the latest node version. Or you can compile it yourself.

Yea, I've never experienced something like this. You can try to create a new file called `test.js` with the following content:


Run it with `node test` and see if it logs test or not (if it doesn't your node is not working / outputting correctly).

Good point, I'll also check a couple backtesters with different EMA parameters to see if they would come to the same profit conclusion.


Hmm that's strange, do you have the newest version of Gekko? And what platform do you have? All logging goes to stdout (which means the terminal session most of the time).


Well you would have to run two Gekkos: One where you disable buying by making the buyTreshold "Infinity" (without the quotes). And another one where you'd set the sellTreshold at "-Infinity", if you put those in two different config files you can specify to Gekko what config to use on running. Read more about that here. This way you can also configure the MtGox one to watch MtGox and the Bitstamp one to watch Bitstamp (if you want them to).

Gekko has plans, but those are deffered until the new BTC-e API stabilizes.


Good idea, I'll look into what will be nice beginner settings. But I do need to make it clear that Gekko is more an trading algo platform then something that works out of the box.


I like it a lot actually, this way the bot doesn't change position as agressive as the default one which saves a lot in fees. But this also means that the bot may not pick up light trends.

I am running a couple of paper traders but my real trading Gekko has thresholds of -0.9 and 0.4 (with other EMA short & long) as it appeared very stable over the last couple of weeks.

---

By the way, I wanted to use your backtester to verify my EMA and backtesting implementation. But I can't find the raw candle date. How did you calculate your candles? Similar to my candle fetcher (which uses the trade data MtGox put in Google BigQuery)?