Username: Password:

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Topics - Clannad

Pages: [1]
1
Hi all

I noticed I am changing very few parameters in each pair's config file, that's why I thought we could do the following:

Have ALL DEFAULT parameters in ALLPAIRS-params.js

If a parameter is present in the PAIR config file, it would OVERRIDE the default value set in ALLPAIRS

This way, we could even have some PAIRS config file just empty, if their config matches the default set in ALLPAIRS
Some other  pair config files would just have different values for strategy, BB values, BUY and SELL Levels

This would make the PAIRS config files much more simple, they would just contain the parameters we want different from ALLPAIRS's default

What do you think about this ?



2
Hi Gunthar, and Hi all,

I wonder if there could be something wrong with the step-gain strategy, or me doing something wrong ?
It seems that whatever I set my Sell Levels to, Gunbot always sells around 2%

With Gunbot 2.04b using Step-Gain, I was tuning my Sell Levels for each pair according to their recent moves, and I was getting this kind of percentages:

3.57%
4.62%
2.81%
4.97%
12.28%
5.16%
6.00%
5.15%
4.67%
8.34%
4.16%
7.00%
13.43%
6.25%
3.08%
5.39%
13.02%
5.01%
26.52%
8.66%
9.39%
7.66%


I am using exactly the same strategy with Gunbot 3.3, running since 36 hours on ~15 pairs, and here are the percentages I am getting now:

3.16%
2.41%
1.82%
2.27%
2.46%
2.25%
3.44%
1.68%
1.59%
2.32%
1.80%
2.14%
1.94%
2.31%
1.60%
1.70%
2.16%
2.14%
2.12%

I have been observing closely Gunbot running, I saw it a few times sell at 2~3% whan the SellLevel 1 was 5% or 7%, and the price was still climbing, and it could have sold at the instructed percentage if it waited a little more...

Did anyone noticed the same issue or could have an explanation for it ?

Thanks all for any help, would be much appreciated
Clannad


3
Hi Everyone, Hi Gunthar,

There are cases where a "BUY NOW" feature would be really helpfull, it is when a pump starts and when we think price will keep getting higher, in such case Gunbot would never buy during the pump and thus we can miss great opportunities.

There were attempts for Gunbot to take into account manual buy orders but it didn't work, I can understand it is not easy for Gunbot to deal with manual orders, but it should be easy if Gunbot places a buy order himself upon request.

I thought about an easy way of coding this: Gunbot already watches for config files modifications. We could add a new parameter called "COMMAND" (for example) at the end of config file. By default this parameter would be empty.

Now if we want Gunbot to buy now, we could just set "COMMAND=BUY", Gunbot would notice the config file change, see we tell him to buy, it would then buy for the set amount of the coin and reset the COMMAND parameter to empty.

We could also use this parameter for other purposes, for example setting "COMMAND=STOP" would tell the bot to stop in a clean way, Gunbot would then reset the COMMAND parameter to empty, and EXIT, this is much better than killing the process.

There could be many other usefull uses for this COMMAND parameter I have not thinked about yet...

What do you think about it ?



4
Hi all,

I'm working on a script that will run above Gunbot, and add some more controls. Basically the idea is to start the bots that have no Alts balance when Bitcoin is high and Alts are low, and stop the bots once they have sold and when Bitcoin is low and Alts are high.

I did this 'manually', cause I had to work on my PC for a few hours, and the result was great! Hence the idea for a script that would do the same automatically. I even consider storing pairs settings in a database, and alter them dynamically according to market conditions, and from these paramaters, create a new config file right before launching each pair.

I'm on Windows 10 64, I already know how to collect the indicators I need, and how to start a pair from a script.

The issue I have is stopping a bot. I'd like to stop them in a clean way (rather than just kill the process) to make sure I don't get corrupted save files. I have no clue about how to do this in a gunbot-friendly manner so any help or suggestion would be welcomed :)

Thanks and have a great sunday
Clannad

5
Hi evryone,

Gunthar, here's a "cosmetic" improvment suggestion

Actually Gunbot clears the consoles every ? minutes, but it does it right after displaying a new cycle. Which means the screen remains black for Bot Delay time until nex cycle is displayed.

When we have like 20+ pairs running, and check all tabs from time to time to make sure everything goes well, there are always 2 or 3 black screen, and we have to wait up to Bot Delay to know if it is a clear screen or if the bot crashed (which happens less and less, but still sometimes).

Could you please proceed to the clear screen just before displaying a new cycle ?

Thanks and keep on the great job !


6
Windows Defender finds the folowing malware in the last release (PatchFix601)

Trojan: JS/Jorv.A!cl


7
Technical Support & Development / Found a litle glitch...
« on: April 30, 2017, 06:01:10 AM »
Hi Gunthar,

There's a misstype that appears in all BTC_XXX-save.json:

Quote from: BTC_XXX-save.json
"pirceToBuy3":null

8
Hi all,

I was thinking, about the Bolinger Bands, instead of making it a separate strategy, we could add it to the Step-Gain strategy, so we would benefit from both the powerful Step-Gain strategy and the additional BB indicator in the same strategy.

Well it is how I was thinking it should be coded, I see the Bolinger Bands and the %b as an additional indicator, like an optional filter that would help buying or selling at the best time.

We would just need to add two fields (with a check case in front of them) for the %LowBB and the %HighBB

Some might say (and they would be right) that there would be too many fields on the Step-Gain setup screen... but the good news is we can get rid of 4 fields that are actually completely useless

Indeed, in the Step-Gain strategy, we use only two levels, so we don't need 3 fields for buy and 3 fields for sell, but only two fields for each. We could get rid of the Buy Level and Sell Level fields too, we would have:
- Buy Level 1
- Buy Level 2
- Sell Level 1
- Sell level 2
And if we want to use only one level, we can just set the same value in the two fields

This is how the new interface would look like, more simple yet more powerful:



Please let me all know what you think about this proposal...

Thanks for reading






9
Hi all,

In the last release changelog it says
Quote
Implementing manual trades: you can now play manually @Poloniex and the bot will know. If you buy and let the bot sell for you, it will follow the logic of supergun and sell at profit. If you sell what the bot bought, it will look to buy again.

I gave it a try but it did not work... but maybe I did something wrong.

What I did was:
1) Launch Supergun on the pair I wanted to buy, with 0.01 max balance to trade
2) manually bought 0.01BTC worth of that coin on Polo

The Supergun didn't notice my buy (despite it was showing the correct balance of altcoin), it kept saying "we have enough balance ... waiting for the price to drop more..."

So I'd like to know if this feature works, if some ppl already used it successfully and how they did.

This feature is really interesting when you catch the start of a pump and I'd like to have it work

Any help would be greatly appreciated

Thanks

10
General Discussion / Improvment suggestion to minimize bagholdings
« on: April 16, 2017, 10:47:08 PM »
Hi all :-) and Hi Gunthar :-)
First of all I want to say I really enjoy Gunbot and Gunthar's constant efforts to improve it !

Now from my experience and from what I read in the chat, the main problem remaining on Gunbot is when it buys at top price, leading to bag holdings for days (weeks... months...) before these alts recover

I have a suggestion for an option that would allow to minimize these situation
I already posted this idea twice on the Telegram group and it received very good feedback, but I'm sure many users missed it (sometimes hard to read 1000+ messages daily), and that's why I'm posting it here on the forum so everyone can see it and comment.

This options consists in an additional parameter that would be % of bolinger band where bot is allowed to buy
The bot would measure the distance between the lower and upper bolinger bands, and if the price is at a higher distance than set, the bot would not buy

We could set this parameter to 500% by default, with this setting the bot would behave exactly as it is doing now, and it would not bother  those who don't want to use this feature.

This would be an additional filter that would be applied just before the bot buys, so it would not change at all the actual gunbot strategy, but only prevent buying too close to upper Bolinger Band if the feature is activated.

For those who want to use the feature, they could set this parameter to 50% for example in which case the bot would buy only if Buy Price is below or at 50% distance from lower bolinger band to upper bolinger band. Depending on each coin behavior the parametre could be set to 10%, 0%, or even -10%

In the picture below for example, the parameter is set at 50% and the bot would be allowed to buy only in the yellow range (sorry the drawing was quickly made and not perfect).



For myself if this was implemented I would use it a lot and set it between 20%~30% I think, but that must be adjustable differently for each pair according to their recent behavior.

Quote
In Spring, 2010, John Bollinger introduced three new indicators based on Bollinger Bands. They are BBImpulse, which measures price change as a function of the bands; percent bandwidth (%b), which normalizes the width of the bands over time; and bandwidth delta, which quantifies the changing width of the bands.

%b (pronounced "percent b") is derived from the formula for Stochastics and shows where price is in relation to the bands. %b equals 1 at the upper band and 0 at the lower band

So basically if we call BBSetParam this new parameter, we would have something like:

Code: [Select]
if (AllOtherConditionsOk && (%b*100 <= BBSetParam))
BotCanBuy=TRUE;

(Note: someone on the Telegram group said that it is useless because we have EMA's, but this is wrong. Bolinger Bands are really a different indicators, and EMA's are useless in what I'd like to see, because EMA's always tend to follow the upper Bolinger Band during pumps, and the lower Bolinger Band during dumps.)

I would like to hear everyone's opinion about this suggestion
Thanks for reading :-)




Pages: [1]