 Show Posts
|
|
Pages: [1] 2 3 4 5 6 »
|
 Im struggling to understand the logic for the buy/sell offsets. Look at this screenshot, wouldnt it made sense that the buy line is the yellow line the blue line is the sell line? It would made a buy at around 0.00045 on the 4th or 5th candle from the right and now would be on its way to make a sell. Instead its on its way to make a buy??? I dont get the logic! That's because the EMA crossings signal a trend, i.e. a crossing from below has a high probability of signaling an uptrend market. In that case it makes more sense to wait until the trend reverses with enough confidence to then sell. The opposite is true when starting with a sell. This bot is more tricky because of its multiple settings that can heavily alter its response (and on the other hand is its strength), but the reasoning is the same. And depending on the market it may not be ideal, you really have to try different settings until you find something ideal for your strategy. Unfortunately there's no bot on the market that can precisely catch the trends without fine-tuning by the operator, specially given the highly speculative nature of crypto markets these days. Regarding your previous post, it's indeed strange, I suggest contacting Margin to see if they can clarify that behavior. |
|
|
|
Thats exactly what I am doing, trying it in the demo mode.  So what am I looking at in this picture? Because the previous candle was green it means that the buy line was above the ema at least 0.2% so it triggered a buy at the top of the candle? Did I get this right? Since I've been playing with the EMA bot *a lot* these days, and have been in contact with Margin to also help me understand more of the inner workings of this bot, I'll drop my 2 cents here. The basic functionality is this: the bot will trigger a buy when the white line (short EMA) crosses the long EMA (dashed red line or blue line if using offsets) from below, and a sell when the white line crosses the long EMA (yellow line if using offsets) from above. But since I'm not sure if the bot has this ability to know when the short EMA is sloping upwards or downwards, I'd take that into consideration when starting the bot: start in BUY mode if the short EMA is below the long EMA (as you'll be waiting for a crossing upwards), and vice-versa for SELL mode. (Margin team, can you clarify if the bot takes into consideration the direction of the short EMA, i.e. crossing from above or below?) Another thing you have to take into consideration is that the orders will be SPOT (market) ORDERS, therefore if there's a huge volatility, as we've been seeing recently, your buy or sell could happen anywhere in the current candle. In your example, if you started the bot in buy mode when the market was low, it would probably have made a profit, selling more towards the end of your image (but remember the spot orders nature...). I don't understand why it bought in the penultimate candle, though, as the short EMA didn't cross the buy offset (it never happened to me). My advice is to visually check your settings backwards to see if they are a good fit for the particular market you are operating. And sometimes a market simply won't be good for bots (ANY bot, Margin or another), because of extreme volatility, price manipulations etc. I feel these recent weeks have been a bit like that. In those cases, I trade manually, as the app is also an excellent trading terminal. Not that I always make money with that, but well, I try...  Hope it helps. Dorian |
|
|
|
|
@hackbyte : first of all, I'd like to say sorry for assuming the issue you're facing was because you were running a tweaked / different Linux distro than the certified one (and just saw your recent post stating you are in fact using an unmodified Ubuntu LTS). You were right and I was wrong: I just confirmed the bug myself! Found the URL for the latest leonArdo, and tried to run it on my Ubuntu 16.04.4 VPS.
And I also assumed Margin wouldn't let something as blatant as the app not starting pass, but apparently they did; need to improve the QA, guys...
So I even deleted my last reply and decided to post this one. As an apology and peace offering, I think I found a workaround for you: just replace the ./lib and ./plugins dirs of leonArdo with the ones form the latest Margin release (you have to replace both, otherwise you'll get core dumps and other problems). It solved the issue, at least the app ran and did some basic stuff.
Hope it helps and at least gets you running the app!
|
|
|
|
Yeah now i have 2 dedicated ubuntu 16 lts installs on 2 machines...
And can't use the latest binary because... it does not run at all..
And the latest working binary just does not care to give any sound for notifications..
Hilarious.... really...
"Latest binary"?? You are downloading the older leonArdo, not Margin. And even leonArdo, I always ran on Linux without these weird issues you are facing. Yeah, crashes, memory leaks etc. it happened in the past (and now they are vastly improved, although still happening in some edge cases I reported them), but simply not running because of lib issues? NEVER. In fact, I could even run it on different distros than the certified Ubuntu LTS. What I noticed is that you seem to have a highly modified Ubuntu and with your "hacks" you made it incompatible. Messed with LD_LIBRARY_PATH, perhaps? Why not install a VANILLA Ubuntu 16.04.4 LTS and run Margin (or even leonArdo!) from it? I would even bet some Satoshis that it will run straight out of the box on a vanilla install. "Ah, but I want to tweak my Linux to the extreme and still run software xyz." Well, that's up to you, NO SOFTWARE DEVELOPER can guarantee their software will run on every freakin' tweaked Linux distro out there. But I'll show you what I just did on a Kubuntu 18.04 LTS right now to check if there was something weird with the paths, perhaps they wrote a shitty shell script that expected Margin on a particular base dir? I noticed you are running from ~/opt, so I did the same: vm@vm-kubuntu01:~/opt$ wget https://margin.de/download/file/273 -O margin-linux.zip
--2018-05-15 01:56:17-- https://margin.de/download/file/273
Resolving margin.de (margin.de)... 104.20.3.168, 104.20.95.167, 2400:cb00:2048:1::6814:3a8, ...
Connecting to margin.de (margin.de)|104.20.3.168|:443... connected.
HTTP request sent, awaiting response...
200 OK
Length: 28657584 (27M) [application/zip]
Saving to: ‘margin-linux.zip’
margin-linux.zip 100%[====================================================================================================================>] 27,33M 4,28MB/s in 8,0s
2018-05-15 01:56:26 (3,44 MB/s) - ‘margin-linux.zip’ saved [28657584/28657584]
vm@vm-kubuntu01:~/opt$ unzip margin-linux.zip
Archive: margin-linux.zip
creating: margin-linux/
creating: margin-linux/licences/
inflating: margin-linux/licences/readme.txt
inflating: margin-linux/run-margin.sh
inflating: margin-linux/margin.bin
creating: margin-linux/plugins/
creating: margin-linux/plugins/platforms/
inflating: margin-linux/plugins/platforms/libqxcb.so
creating: margin-linux/icons/
inflating: margin-linux/icons/margin.svg
creating: margin-linux/lib/
inflating: margin-linux/lib/libQt5Core.so.5
inflating: margin-linux/lib/libQt5Widgets.so.5
inflating: margin-linux/lib/libQt5WebSockets.so.5
inflating: margin-linux/lib/libQt5XcbQpa.so.5
inflating: margin-linux/lib/libQt5Svg.so.5
inflating: margin-linux/lib/libicuuc.so.56
inflating: margin-linux/lib/libQt5Gui.so.5
inflating: margin-linux/lib/libicui18n.so.56
inflating: margin-linux/lib/libQt5DBus.so.5
inflating: margin-linux/lib/libQt5Network.so.5
inflating: margin-linux/lib/libQt5NetworkAuth.so.5
inflating: margin-linux/lib/libicudata.so.56
inflating: margin-linux/margin.desktop
vm@vm-kubuntu01:~/opt$ cd margin-linux/
vm@vm-kubuntu01:~/opt/margin-linux$ ./run-margin.sh
installing margin icon to user account
Not running on Ubuntu 16.04
-: FAILED
sha256sum: WARNING: 1 computed checksum did NOT match
Downloading and extracting https://margin.de/static/libssl.tar.gz
VOILÀ, Margin runs!
|
|
|
|
The app is stable for me on Linux, no crashes and constant memory usage, although I'm not heavily using bots at this time. Good job guys, really like the new features.
Can i ask which specific linux distro and version you are using? Sure: Ubuntu 16.04.4 x64. More specifically, the Lubuntu variant, because it runs on a VPS and I wanted something lightweight. |
|
|
|
|
This new release is simply awesome, two of the most asked features have been implemented. The drawing tools are just... beautiful! I've been playing with Fibonacci retracements, in parallel with some TA studies, and it's a valuable tool. By playing with it I've discovered some nice features such as "extend lines", the color / alpha transparency customizations, labels and, most importantly in this case, the ability to select which Fib levels we want. The simple tools like lines are also important, I'm personally using horizontal lines to mark price levels relevant for me (where I bought, sold etc.).
Regarding the EMA bot, from a "visual backtesting" I've been doing, it shows a huge potential, but I still haven't left it running "for real". By the way, I think that's a good tip: you can visually check how your bot & settings would perform visually, by scrolling the charts. While not a full, proper backtest with simulated P/L, it allows us to quickly have an idea. And I've noticed that the "best" bot will heavily depend on the market behavior, you'll have to do this visual backtest with each of the bots (including even the simpler ping-pong) to determine the best one. And sometimes it's best to manual trade, using the indicators.
The app is stable for me on Linux, no crashes and constant memory usage, although I'm not heavily using bots at this time. Good job guys, really like the new features.
|
|
|
|
Interesting article; as always, bringing valuable contributions to the forum, thanks! My personal experience with stop losses is somewhat mixed: 1) For high volume markets, it usually works, although I was commonly using a lower % than the article suggests (and indeed, sometimes it was triggered for the market to recover shortly after). 2) For low volume markets, which are susceptible to huge variations and market manipulations, I think it's usually a trap; but I barely trade in these markets anyway, because of these reasons (easy targets for P&D groups, for instance). 3) Trailing stop is wonderful if you place it in an upward trend. Margin has support for these orders on Bitfinex, which I tend to favor given this and some other reasons. 4) BUT despite being useful, given the wild west fuckfest the crypto markets have been recently (have you noticed what was precisely called the "Bart Formations", which is basically Bart Simpson's head, happening all the time with the BTC market some time ago?), I've been focusing on manual trading, and usually go to sleep with USD/USDT. Too many shocks after waking up when bought in crypto... |
|
|
|
You won't have any issues with API quotas if you use different ISP connections!
Hi Jonathan, I have one question regarding that: let's say I have Margin on a remote VPS, and a local installation. Since they are on different ISP connections, can I leave both running at the same time, on the same exchanges, for the following scenario: - Margin@VPS: bots. - Margin@local: manual trading. Are both instances going to be synchronized in real-time, meaning if the "bot" instance makes a trade, I'll see it immediately on my local installation and vice-versa? Any possible side-effects with that scenario? |
|
|
|
A great article to get people up to speed to what this is about. https://www.crypto-news.net/fibonacci-sequence/From the article: "The truth is that nobody really knows why the Golden Ratio appears in trading charts, in the size of our hands, in the way we build our cities and make our art. It’s literally everywhere and all evidence suggests that people are using it subconsciously for the most part." -Very thought inspiring, there is so much more to be gleaned here if one wishes to go down the rabbit hole, it is very enlightening. Having a great day, hope you are too!  Hey that's a very interesting article, thanks for sharing! By the way, you always share some interesting things here on the forum; thanks for that too. Have a good day! |
|
|
|
|
Merging functionality with astonishing visuals: that's Margin. When the market isn't favorable to bots, I do manual trading using these awesome features (the multiple indicators, candle sizes quick switch, market depth etc.). The drawing tools with Fibonacci will add a great value to the app. Looking forward for the release!
|
|
|
|
Hi, i try to run margin on Debian Jessie 8.2 server but there are many errors. i fixed most of them excpt this error (version glibcxx_3.4.21' not found debian) btw libstdc++.so.6 is lastes version on the server. plz how i can run the margin?
You should use Ubuntu 16.04, which is the certified release by Margin. The Ubuntu Server 16.04 release is very nice to use on servers, I prefer it instead of "plain" Debian: it's stable (being a LTS release), but with newer software than Debian -- meaning you almost 100% of the time can install software certified for Debian, plus the ones certified for Ubuntu, like Margin. A glibc version mismatch, especially when the software you're installing require a newer version than the base OS, is a pain in the ass to solve (if not impossible without changing the base OS to have matching glibc versions). *Possibly* if you upgrade your Debian installation to Debian testing or unstable, you'll have an up-to-date enough glibc, but if you're going this route, I'd suggest a clean Ubuntu 16.04 install, which will get you up and running in no time. |
|
|
|
|
Guys, the official Binance mobile apps now support stop orders. That led me to investigate if their API support these types of orders: it seems it does. Do you think it would be possible to add stop and take profit orders for Binance, the same way you implemented for Bitfinex? Thanks!
|
|
|
|
Wow, Aggressive! I guess the truth really does hurt! Now I can see that you do not treat your customers with the professional courtesy that any reputable company should.
Who would want to do business with you and buy into your bait and switch subscriptions if you're going to disrespectfully treat your customers this way? Your attitude here has dictated what will happen to your company next. It doesn't look promising.
"Your customers" => no, just you, because this is becoming trolling from your part. If you read this thread you'll see that they are nice, honest and fair with most of the customers. I don't have anything to gain personally by defending Margin, but I believe in fairness: if a business behaves badly, I always complain (but don't troll like yourself); however, if it has a record of good practices, I'll also stand to defend them. This subject of the business model change has been discussed to the exhaustion now. What's tricky about trolling is that people feel compelled to reply, but it just serves its purpose. Maybe I also contributed with this reply, but thought it was important given the strong words you're using, and don't plan to continue replying anyway. |
|
|
|
It is about greediness to capitalize on past users who were fooled to pay a "Lifetime License".
Where did you read a promise for lifetime UPDATES? If they stated that, you may have a case, otherwise you DO have acquired a lifetime license. You can use your product forever. Yes, yes, APIs can change, the old product won't be updated anymore etc., but that would fit in a "lifetime updates" model. There's simply almost no business that can survive without either charging for updates, support, or subscription, unless they have a constantly growing user base, which is not the reality. Almost all software that come to my mind fit in one of the models I mentioned. And calling them "greedy", "misleading", "cheaters" or whatever else you called them is absolutely unfair. First because they didn't break a promise. Second because they listened to the users and offered a very good deal for existing ones. And third, because everyone here who have done business with them, used support etc. know they are extremely honest and decent people. |
|
|
|
Edit: Before we get to a strategy editor we will add a trend watcher as a general feature to spot order bots like the Bollinger Band bot to help ensure that the bot does not buy or sell when the market is trending
Hey, that sounds interesting. Can you share some details on how it will operate? |
|
|
|
Update to the announcement
After discussing what we could do for our Hero early adopter customers, in recognition of the fact that you guys got us to this point, we have decided to give a special status of Hero to every leonArdo customer going forward. It won't just be that you have a Hero badge in your version of margin, but it entitles you to a 50% discount on all future margin terminal subscription payments. The offer will kick in if you sign up for a subscription within a month of your free period ending.
We wanted to also make the point that switching to a subscription based model is absolutely necessary for us. We are a real company, with real employees who have families and mortgages and have to be paid at the end of each month. If the market only saw month on month growth then it might be possible to survive by sticking to the lifetime license model, but this is not the case. Catastrophic events like what happened in mid January, where a lot of cryptos lost up to 70% of their value, have a huge effect on our revenue. This puts everything we have built up in danger and if not dealt with could mean us having to close for business - which would also mean no updates for leonArdo. This was not a decision we took lightly, rather one we couldn't avoid.
Thank you Jonathan, really appreciate this offer! I think the change to a subscription model is reasonable; from a business perspective, people need to understand a few key points: 1) If the model is a one-time charge, the business can only survive on the following situations: 1.1) Make it really expensive, so you have enough savings to keep it alive for some time. However, it may not be successful, as not many people will be willing to pay the price, and the savings will not last forever. 1.2) Charge for updates, as some big software companies used to do. May not please everyone, just like a subscription-based model. 1.3) Charge for support, as some big software companies. Not really applicable for this business, it's usually sustainable only for corporations with big customers, that sign long-term support contracts. 1.4) Ensure you have a constantly growing user base. I don't think this is a reality for almost any business; the user base will flat out at some point. 2) Most businesses are, therefore, changing their business models to a subscription-based one. Look at Microsoft and Google, for instance... It ensures the business survives, as long as the price is fair, they commit to quality and constant updates, and provide good support. If they offer all that, it may be beneficial to the customer, as he can expect future improvements and have a reason to keep demanding them. So far Margin Software has been providing all that, so I expect it will continue (and even improve) with the change. The price and conditions can be discussed, of course, but the model makes total sense for me. And I think they have been fair with the transition conditions offered for existing, early-adopter customers. |
|
|
|
|
A feature request for the Bollinger Band Bot: when, given the nature of spot orders, it can't buy or sell the max amount configured, only a partial amount, keep trying until it maxes out. It could keep trying with repeated spot orders, or wait for the next trigger (e.g., when waiting for a buy, on the first signal, it bought 5/10; if a subsequent buy signal happens, try to fill the rest, or keep trying while the signal is there).
Sometimes we minimize our gains with the current behavior, because it only bought or sold a small part of what was configured, flipping to waiting the opposite signal.
|
|
|
|
|
@hackbyte, what I understood is that you have the funds (i.e. they are not "gone"), but they don't show up in leonArdo, right? If that's the case, I haven't noticed a similar problem (also using Bitfinex). However, I usually only keep a few pairs open and don't run bots in all of them; the most I used to keep open is 6 (now it's usually 3) per exchange. Perhaps with too many pairs open, the amount of network traffic triggers the bug? My suggestion for them to test.
The minimize and slowness problems I also haven't noticed, but again my setup is different. Have you tested with fewer pairs open? Could help nail down the issue. Noticed any memory leaks or CPU spikes after 24h that could indicate something?
And regarding Bittrex, I agree; I simply gave up using the exchange. Don't know how much Margin's team can improve that, because it seems Bittrex's API itself (their side) became too slow recently, after the huge influx of new users. It was OK before that. It would be great if the devs could somehow improve it, though.
|
|
|
|
|
@Margin team, I have another suggestion for a future release: keep the market depth open when we switch to a different time period. Everytime I switch periods, I have to drag the chart window to reveal the market depth again. It should be saved together with the other indicators and layout choices.
Nothing urgent, of course; just would really appreciate if it entered your to-do list. Thanks!
|
|
|
|
|
