Aitor

[Aitor]

[quote quote=9864]If I remember correctly.. pile up must be true and then set these. That was from my v21.1 .0

“RESERVE_PILE_UP”: true,
“withdraw_address”: “”,
“withdraw_threshold”: 0.05,[/quote]

I will tell you something when I know if it’s possible or not.

[quote quote=9864]What is the minimum bot version to buy so I can upgrade with coins?[/quote]

Standard. Note that upgrading from Standard to Pro with Gunthies doesn’t include Backtesting Addon.

[quote quote=9864]Also.. do I need a separate wallet for the gunthy coins per bot/instance?[/quote]

If you are running several instances of your same license, you can use the same wallet. For example, if you have Pro, you can run any number of instances in the exchanges you can run with Pro. In the case you buy a new separate license, it requires a new Gunthy Wallet as it’s a different license.

Note: When I say “you can run any number of instances”, I mean, you can run for example 7 instances in your Binance account, but the apis must belong to the same Binance account, of course.

[Aitor]

Things are going good. As I said before, BTC had very low volatility last days and it affected the performance and number of trades of the pair USDT-BTC. Despite of that, the bot ended closing the cycle with profit, but I decided to change USDT-BTC pair for other one. Why? A pic is worth one thousand words:

Other pairs are doing very good. For example, let’s take a look to USDT-ETH:

Despite days 23, 24 and 25 have been heavily affected by Binance maintenances and server overload, it managed to make 111 trades with an average profit of 5,69$ and overall results for the first 12 days are very good:

Finally let’s see profit per day stats:

[Aitor]

Interesting to see how it performs with BTC as base and XRP which is a extremely volatile coin. In the first post you put “Trading Limit = 30”, I guess you have to state the TL in your base coin, that is, in BTC.

[Aitor]

A quick way to handle it would be to put the exchanges you want to withdraw from in a separate instance, so the whole amount will be used for that especific exchange.

On another note, I never used automatic withdraw function. What’s the exact name of the parameter? I can’t find it. Maybe it’s only of the parameters that got discarded?

[Aitor]

Hi Peter, several important things:

1- We can’t use Binance PNL as reference as it just states the value in dollars of your assets. For example, suppose you have 0.10 BTC and 5000$ and BTC price = 50k. In this example the value of our assets is 10k$. Ok, suppose we don’t make a single trade and BTC price raises to 100k. Now the value of our assets is 15k$, so Binance will show like we did 50% profit in 1 day… Obviously it has no sense. Gunbot would show profit = 0 because we didn’t a single trade.

Gunbot PNL calculation takes into account the profit / loss in our base coin of the trades we have closed. Our target is to increase the base coin. If our base coin is USD, we want to close trades to earn USD, and we shouldn’t care if btc price raised or dropped last day, week or month. Same happens when we use BTC as base, in this case our target is to earn BTC. If we convert 1 BTC into 1,10 BTC, we got +10%, didn’t we? Would in this case make any sense to say we lost just because btc price dropped by 20% that week? Obviously not, and that why I think we should use Gunbot PNL for serious calculations.

2-

[quote quote=9844]The PNL in Gunbot is not working[/quote]

Please take a look at this post:

https://gunbot.shop/forum/topic/usd-base-with-pnl/#post-9849

You can send me a msg via Telegram and ask me for a beta version that includes the partial fix. It just needs to stop the bot, overwrite old Gunthy-win file with the new one and start the bot again. After doing this, all the profit from previous trades / days will appear.

3- By the way, from 30th March to 26th April, BTC dropped 4,69%, and despite of that drop you still increased the value of your assets 3%+. In my case in that period of time I increased the value of my assets 9,89%. In both cases, we have earned much more profit if we just take into account the trades we closed, which is the correct way to measure earnings in my opinion.

[Aitor]

PNL is a known bug and the only thing we can do is to wait for the fix. Latest betas implemented a partial fix, and when you start the bot they show the info but they stops showing it for the new trades. So there are 2 ways to update the info:

1- Stop the bot and start it again. So latest trades will be updated.

2- Set BOT_CCLEAN to 1: This will clean cache every 1 hour, and will make trades PNL will be updated every hour.

Anyway next stable version is about to be released and PNL thing is a priority, so hopefully we won’t have to wait too much.

 

[Aitor]

Interesting. What bot version are you running? When you say USD is not properly calculated, are you running pairs with USDT as base coin?

[Aitor]

Gracias por compartir el reporte. La verdad es que has tenido una primera semana un poco accidentada: BTC en máximos, 2 días de parón por mantenimiento de Binance, otro día con el servidor de Binance sobrecargado… Solo te ha faltado que hubiese habido una invasión alienígena para completar la semanita :thumbup:

[Aitor]

Some other interesting things I would like to mention:

1- I have added STOP_AFTER_NEXT_SALE tag to the override of pair USDT-BTC. That means that once USDT-BTC pair ends this cycle, it won’t start a new one. The number of trades and profits I got with other pairs has been much higher, so I decided to run other pairs.

2- This week we have gone through the Scenario 4 I talked about in this post:

https://gunbot.shop/forum/topic/spot-grid-dca-mutation-project/#post-9385

I have seen some people getting super impatient and getting mad because the bot “got stuck”. Got stuck in this case means making trades at a lower pace and keep in the same cycle for more than 1 day.

In the pic above we see USDT-UNI graph. The price drop from highest point to lowest point was around 21%, and as you can see, the bot made a very good job lowering the average bought price DCAing and CTing. The yellow line shows the moment the bot existed the cycle with profit despite the big price drop.

3- BTC price dropped from around 64500 to 47000 aprox. That’s a 27% drop. After the drop the price stabilized for long time and the volatility was very low. That’s why this pair has performed worse than others this week. No volatility no party. Despite of this, the bot managed to lower the average bought price from 64500 to 57200 and at this moment it’s only 4,5% below break even price. So, what I am going to do? :think:  No drama, I will just wait and let the bot to continue doing its job and lowering average bought price and I’m quite sure it will eventually end closing the cycle with profit.

You know what? The most profitable trades are (almost) always made when closing a cycle after a big price drop:

[Aitor]

Hi, sorry for the delay in the response. Does it happen always? I mean, using BTC as base PNL is always shown? Recently I have been running v21.7.7 and higher that partially fix the bug that happens on v21.6.2 for example. With v21.7.7 and higher, we you open the bot, it shows the profit for each trade, however it stops showing the info of new trades very soon. So what you can do is to stop the bot and start it again. That way the profit of all the past trades is shown. Anyway the next stable version is around the corner and hopefully it includes a complete fix for this.

[Aitor]

Hi Tek, thanks for sharing your results :clap:

This week has been a bit weird because Binance made 2 maintenances that stopped the trading operations for many hours, and in the middle of it, during 24 hours aprox, their server was so overloaded that a lot of multiple buys and sells happened. I see multiple buys / sells affected you, specially at the beginning of the second pic, it’s like your bot turned into a machine gun hahaha

Regarding PNL: My idea was to post both Binance and Gunbot PNL, but I have reached to the conclusion that Binance PNL is useless. I mean, suppose you are trading USDT-BTC and you have 2000$ and 0.10 BTC. Suppose your trades are 0, you do nothing. Ok, if BTC price raises… you have more USDT as Binance just convert the value of your assets to USD.

What happens if BTC price drops? You have less USDT, so you lose even if you didn’t make a single trade? It has no sense. Also is important to remark that in this case despite the value of your assets is a lower USDT amount, BTC value would increase in this case, so it would be a win for some people.

What I think about this? After analyzing it, I think Gunbot PNL calculation is much better as it just calculates the profit / loses when a trade is closed. If we are trading USDT-XXX our goal is to increase USDT, so when a trade is completed, it calculates the USDT profit / loss. In my opinion, Gunbot PNL calculation is the way to go.

[Aitor]

Quick update: We are starting to test v3 variants with big improvements compared to v2. Things are going good and the strategy successfully handled 40-50% drops in some cases. I will start posting results for SG Mutation C6, the one I’m testing now :thumbup:

[Aitor]

Seguramente hoy explicaremos en el canal del grupo sobre la v3 y las nuevas variantes que tiene.

[Aitor]

Bien, vamos a seguir testeando todo bien, y veremos qué estrategias / pares son los que mejores resultados dan en un periodo más largo de tiempo y también partiendo de las condiciones más parecidas posible para que el resultado sea neutro.  Una vez que tengamos claro lo que nos ha dado mejores resultados, podemos implementar con el autoconfig una automatización para que el propio bot se encargue de añadir y quitar pares de la forma que queramos. Supongo que esto ya quedará para la versión v4.0 porque la v3.0 ya tiene bastantes mejoras y todavía casi ni hemos empezado a testearla.

[Aitor]

There is an easy way to update using import feature, but considering this time is a downgrade to a previous version, I would recommend you to make a fresh install.

[Author]

Posts