The viability of Trading Bots.

[mikeywith]

Traders have different opinions in regards to the viability of trading bots, in this topic, I attempt to study that with real-life trading.

It's important to know that this does not test the random bots available online where users have no clue about how and when trades are made, this study is intended to confirm or deny the theory of fundamental importance in trading as well as human emotions and their effect on the trading results.

What I aim to do is take a simple strategy that I use in my normal trading and automate it.

Bot Rules :

• Number of pairs traded = 10
  
• Trading Platform = Binance
  
• Position direction = Long only
  
• Position direction = No duplicate positions
  
• Compound?= Yes ( 10% of the the avaiable balance will be used at all times)
  

Date of the first trade = 2020-08-21


** I initially deposited $102.86 and then found out it wasn't enough for 10 pairs (some orders could not be activated)
** Later added another $99.68
** Finally I had to deposit another $300


Initial Total Investment = $502

Update : 25-11-2020

Balance = $480.74
Unrealized PNL=  $98.38
Current Status Including U-PNL = $579.11


I will be updating this thread regularly, it would be great if I manage to do this experiment for 12 months with a combination of some bear, bull, and side-ways markets.

[maxreish]

I am still studying these kind of bots and planning to buy some but still collecting and researching which the best bots are suited for me since trading bots have different styles.
What's the name of this bot you are using so I can further study how does this works?

Also, the minimum trading pair is 10? Or we can set it up as we like to trade just 2 pairs or so? Most likely if ever I just have small bank roll or fund as $100.

[mikeywith]

I think I made this clear enough in the OP, I am not using a commercial bot, I made my own, nothing fancy it's just a simple strategy that I use in my manual trading, it's now run by the bot, it's merely a pine script on tradingview that sends notifications to a script which interacts with Binance API to execute the trades, so no bots are for sale here and there are no names.

Also, the minimum trading pair is 10? Or we can set it up as we like to trade just 2 pairs or so? Most likely if ever I just have small bank roll or fund as $100.

The reason why I am limited to 10 trades is my tradingview plan, it's a Pro subscription limited to 10 pairs notifications, aside from that the only limit would be the exchange you use if you are very good with python or any other programming language - you can bypass tradingview and write everything from scratch, run the code on VPS and let it do the work.

But if you don't have a working strategy that you need to automate then all of this would be a waste of time and money.

[tbct_mt2]

Can you give me advice on wash volume on exchanges and how they affect trading bots, please.

I know it is a stupid question and if I know one exchange is a wash volume exchange, I would not use bots for trading on that. On some small exchanges, bots for wash volume are easily to be detected by eyes. I only don't know how you can control it on real environments for your tradings.

Another suspicious question is how fast or low the signal from your bot to exchange can affect the trading results?

Thank you.

[Tytanowy Janusz]

Hmm... this study will not prove anything without control sample since 100% of your bot performance depends on strategy you feed him. It is simple recreating the strategy but much faster, better, 24/7, 0 emotions. It must be better that manual trading with same strategy. If strategy is bad, performance will be bad, conclusion will be - bots doesn't work. Maybe you should think about going something like - 1 month manual and automatic trading comparison.

I am still studying these kind of bots and planning to buy some but still collecting and researching which the best bots are suited for me since trading bots have different styles.

As mikeywith said - its the best to do your bot by yourself and feed it with the best strategy you have.

if you are very good with python or any other programming language - you can bypass tradingview and write everything from scratch, run the code on VPS and let it do the work.

Try it. First program in python that I made was trading bot and it started to earn starting from day 1 giving me extra motivation to carry on with it. Now trading bots are my main income. Before python I had basics in c++ and java but it was elementary school knowledge level.

Can you give me advice on wash volume on exchanges and how they affect trading bots, please.

I know it is a stupid question and if I know one exchange is a wash volume exchange, I would not use bots for trading on that. On some small exchanges, bots for wash volume are easily to be detected by eyes. I only don't know how you can control it on real environments for your tradings.

If your bot does not use volume as indicator than wash volume does not affect your bot. If it use volume than you can get this data from other exchange.

Another suspicious question is how fast or low the signal from your bot to exchange can affect the trading results?

Thank you.

Depends on strategy. If bot is scalping on binance that it would be good to run it on vps located close to binance to avoid up to 500 ms ping. If its trading on 4h candles than this 500 ms will not make much difference.

[gotmilk_]

Hey guys.

I just wanna to share with you one great project we are working on - Signal MONSTER
CONVENIENT WAY TO BITCOIN TRADING WITH CONSISTENT RETURNS
These days everyone’s finances are affected, confidence is challenged, generating consistent returns becomes more difficult, and fewer opportunities exist. We are very well aware of that and we are here to help you!

Signal MONSTER provides convenient and reliable mid-term trading signals based on the quantitative analysis of the most popular and liquid cryptocurrency pair in the world.

Please join us on our discord channel where u can find a lot of info about project, and now you can try our signals for free! Then tell us, if this is what you always wanted

https://discord.gg/A3q852MwZ4

If you need any other question, feel free to ask.

[pdsghb]

Can I buy robots from you?

[mikeywith]

Maybe you should think about going something like - 1 month manual and automatic trading comparison.

You make sense, but even doing the 1-month switch won't really prove anything, with manual trading there will be days when I don't trade or signals that I don't catch, so eventually I will miss many trades, and so if those were supposed to be winning trades then my manual results will be bad, if they were more of losing trades then my manual trading results would be better because a losing trade missed is a winning one.

Here are the current positions opened by the bot



I missed 7 out of 9 in my manual trading and there was a glitch with a BTC/USDT order which should have been activated at 18.3k level



I took that trade manually but the bot failed to take it, which will also affect the result unless I modify the bot's numbers, so I guess there is just no perfect way of doing this, but I will continue to update the thread either way.

Can I buy robots from you?

Please read the OP, I don't have any bots for sale, most trading bots online are fake, if you have a trading strategy that works, hire someone to code the bot/script for you, shouldn't cost you above $100-200, and if you don't have a trading strategy you should learn TA and come up with one.

[zanezane]

Considering that you are using your own bot, it is pretty impressive. If you ever plan to release that bot, make an announcement. For me, Bot viability relies on convenience, if you are not so much as involved in the trading then it is a good trade because you are using your time wisely, you are doing some other things that make you money while your bots are trading. In this day and age, time is the most priceless thing that you can have and having someone do the taxing work for you is a good thing. Good job OP.

[mikeywith]

Considering that you are using your own bot, it is pretty impressive. If you ever plan to release that bot, make an announcement.

I doubt that I will be releasing it since I don't want to be responsible for anybody's loss, keep in mind that non of these trades is closed yet, while the majority of them won't be losing (given the strategy that I use), there is still no guartnee that they will make even half the profit you see in the picture, so I woudn't say it's impressive yet, I will keep this topic updated for further disccsuion.

[zanezane]

~

I doubt that I will be releasing it since I don't want to be responsible for anybody's loss, keep in mind that non of these trades is closed yet, while the majority of them won't be losing (given the strategy that I use), there is still no guartnee that they will make even half the profit you see in the picture, so I woudn't say it's impressive yet, I will keep this topic updated for further disccsuion.

I am just giving you some possibilities, you also should be proud of your creation my guy. Even if it is not yet yielding you profits, you shouldn't be ashamed. Regarding the possibility of releasing your bot for commercial use, I have heard a lot of people that say they wouldn't release it, don't be a pessimist. 

[ranman09]

~

I doubt that I will be releasing it since I don't want to be responsible for anybody's loss, keep in mind that non of these trades is closed yet, while the majority of them won't be losing (given the strategy that I use), there is still no guartnee that they will make even half the profit you see in the picture, so I woudn't say it's impressive yet, I will keep this topic updated for further disccsuion.

I am just giving you some possibilities, you also should be proud of your creation my guy. Even if it is not yet yielding you profits, you shouldn't be ashamed. Regarding the possibility of releasing your bot for commercial use, I have heard a lot of people that say they wouldn't release it, don't be a pessimist. 

Maybe if he releases this for free with the "use it at your own risk" tagline then there would not be a problem in his conscience. I get his point that he doesn't want to be responsible for any loss as he is just sharing a study.

Based on the first logs on the study, the bot seems good. It's positive still so it is still good.

[zanezane]

~

Maybe if he releases this for free with the "use it at your own risk" tagline then there would not be a problem in his conscience. I get his point that he doesn't want to be responsible for any loss as he is just sharing a study.

Based on the first logs on the study, the bot seems good. It's positive still so it is still good.

Yeah, it would be the best course, I was on the profiting side of the story so never considered that haha. I really think the bot will be a good help for some traders who want to spend some time away from the monitor to do other things besides watching the valleys and mountains in the markets form.

[Lakai01]

I really think the bot will be a good help for some traders who want to spend some time away from the monitor to do other things besides watching the valleys and mountains in the markets form.

I think the problem is simply that many people believe a bot is a "get rich fast" tool that you set up once, run for months, and then just cash out 1xx BTC in profit.
In my experience, a bot is simply a very good support to execute your own trading strategy much faster, more targeted and most importantly, with pinpoint accuracy. Nothing more. If your strategy is bad, so is the bot.

[Yamifoud]

I really think the bot will be a good help for some traders who want to spend some time away from the monitor to do other things besides watching the valleys and mountains in the markets form.

I think the problem is simply that many people believe a bot is a "get rich fast" tool that you set up once, run for months, and then just cash out 1xx BTC in profit.
In my experience, a bot is simply a very good support to execute your own trading strategy much faster, more targeted and most importantly, with pinpoint accuracy. Nothing more. If your strategy is bad, so is the bot.

This means that you'll need to figure out first if your strategy is really working before applying to the bot. The same is that be doing trading manually and the result will also be depending on us, not the bot itself because we are making the final decision and the bot only just to follow what we input to them.I'm not stopping people to use trading bots but we should have to consider first our self if we are knowledgeable already into trading otherwise, we just lose and blame the bots for not functioning.

[darewaller]

I have to say that trading bots are doing it 7/24 and that is the difference because of that we are not really doing the same thing as humans. Technically you could do better than a computer, that is no doubt, but it is also has some advantages of their own as well, there is really no other way around.

Bots could make a profit by checking things 7/24 and doing it non-stop and every time whatever you told it to do, it does that without missing a beat, that is some marvelous thing. What makes bots good or bad is not the bot itself, bot just does what you tell it to do, what makes bots great is the fact that you may not be able to find a right indicator or anything to put in there, so it would look like your system you put in the bot could be wrong. Hence this is why I do not think that we should be really doing comparisons between real people and bots.

[ranman09]

I really think the bot will be a good help for some traders who want to spend some time away from the monitor to do other things besides watching the valleys and mountains in the markets form.

I think the problem is simply that many people believe a bot is a "get rich fast" tool that you set up once, run for months, and then just cash out 1xx BTC in profit.
In my experience, a bot is simply a very good support to execute your own trading strategy much faster, more targeted and most importantly, with pinpoint accuracy. Nothing more. If your strategy is bad, so is the bot.

Is there a bot you know that users can configure for their own strategy available in the market or is there nothing like it? Do you really need to be good at scripting to build the bot?

[so98nn]

Dont forget to consider that we had unusual hike in all the crypto currencies this year due to outburst of bitcoin's market cap. I think it's very obvious that bitcoin will be affecting your rest of the portfolio in very unique way.

On a normal day this would have been little harder to attend the current hikes that you show up in the above window. I wanted to study that one because in crypto we wont be seeing such hikes on regular basis, it has limitations!

Also, does it mean that bots are not working up that perfectly as compared to manual trading?

[zanezane]

~

I think the problem is simply that many people believe a bot is a "get rich fast" tool that you set up once, run for months, and then just cash out 1xx BTC in profit.
In my experience, a bot is simply a very good support to execute your own trading strategy much faster, more targeted and most importantly, with pinpoint accuracy. Nothing more. If your strategy is bad, so is the bot.

That will not be on the bot's problem then, it will be on the strategy that the user is employing on his trading. AFAIK, there are some bots that are automated although at lower volumes.

[Lakai01]

that you'll need to figure out first if your strategy is really working before applying to the bot. The same is that be doing trading manually and the result will also be depending on us, not the bot itself because we are making the final decision and the bot only just to follow what we input to them.I'm not stopping people to use trading bots but we should have to consider first our self if we are knowledgeable already into trading otherwise, we just lose and blame the bots for not functioning.

Right, yes. Just because you bought a bot for a lot of money doesn't mean it's going to make you money. There are enough people who buy a bot and then ask with which strategy they should configure it. That will not work.
I also don't want to stop anyone from working with trading bots. The bots are very good and do exactly what they are supposed to do - and much faster and more accurate than any human trader. Just don't expect miracles.

Is there a bot you know that users can configure for their own strategy available in the market or is there nothing like it? Do you really need to be good at scripting to build the bot?

In my "trading days" (most notably 2017 and 2018), I used Gecko. This is a freely available trading bot based on open source that you can configure very well. But now there are many platforms that host trading bots for you and also come up with a variety of predefined strategies. I personally have never used one.

[mikeywith]

Dont forget to consider that we had unusual hike in all the crypto currencies this year due to outburst of bitcoin's market cap. I think it's very obvious that bitcoin will be affecting your rest of the portfolio in very unique way.

Of course, my strategy barely makes a profit in a bear market since it doesn't short, so it needs a bull market or a coin that has some positive news whereby the coin goes up regardless of the market condition,so if your strategy doesn't lose you money during a bear market then it's good enough IMO.

With that being said, the bot has been online since Aug which the majority of it is bull, if you randomly invested in 10 coins you would be in good profit by now, so it really doesn't indicate anything yet, it needs to run for at least a whole cycle where we see a new ATH followed by a new local bottom, this means 1-2 years at least.

Also, does it mean that bots are not working up that perfectly as compared to manual trading?

It doesn't mean anything yet, I don't have a perfect sample to compare it against, I just laid out that the pros of using a bot are that you don't miss any trade and you stick to your strategy 100%, which I do fail to comply with, for an example, the current ADA position profit is 129.37%, if I was manually trading it, I would close it right now, but the bot is following the code which says don't close it yet, so this 129.37% could end up being 150% or 50% only, whereby in manual trading I would have probably closed it long before, despite that fact that I am not too prone to emotions, I don't think I can go to bed without closing a 129% profit position.

Anyway, when a year has passed and I have long term results from both my manual trading and the bot, I should be able to complete this study, for now, let's just watch the bot do its thing.

BTW, the bot lost a few trades earlier own prior to opening all of these positions, and then I had to reset my orders, ended up at about 500$ back from the 579.11$, so I call that breakeven and I loaded another $500 so the initial cost now is considered $1000.

Current balance including Unrealized PNL = $1,357.0454

[Lakai01]

Thanks for this information! I tried researching and it seems not maintained anymore. By the way, I saw that it has many forks though so I will continue my research from there.

Yes, that's right. Gecko has not been developed for some time.

However, the version uploaded on the page works without any problems. Also, as you said, there are a lot of forks on Github, some of which are linked directly on the Gecko page.

The big "problem" with Gecko for newcomers, however, is that this bot must be hosted itself. I had it running on a Raspberry Pi at the time. Solutions you have to pay for are certainly more comfortable to use, especially for newcomers in the field!

[mikeywith]

The big "problem" with Gecko for newcomers, however, is that this bot must be hosted itself. I had it running on a Raspberry Pi at the time. Solutions you have to pay for are certainly more comfortable to use, especially for newcomers in the field!

You can host it on a free AWS, it's a javascript code that runs on node.js, should be easlity and cheaply hosted anywhere, however, it doesn't seem like it has Binance Futuers ready implemented, I assume that is where most of the trading happens nowadays, with that being said, I think the Binance Futuer API isn't very different from the Spot/default API, so it should be easy implemented, and perhaps one of the forks you talked about it has already done that.

The second option would be to hire someone who knows how to code in python, show them the Binance API
and ask them to write the code for you, I am pretty sure that on Fiverr or the like, you will find people who have already done this and all they need to edit is the strategy itself which does't make much of the code, the majority of the code would be the one that interacts with the exchange to place/close orders, as far as the when to buy/sell goes, for most strategies it's only a matter of a couple of lines.

[Tytanowy Janusz]

You make sense, but even doing the 1-month switch won't really prove anything, with manual trading there will be days when I don't trade or signals that I don't catch, so eventually I will miss many trades,

If you apply same strategy for manual and bot trading the only variable (advantages of algo trading) is:

1- your presence on market
2- your successiveness in sticking to the strategy
3- reaction time

So if you ment to compare bot vs human performance than its good to include p1 factor too (missed trades by not being active 24/7)

and so if those were supposed to be winning trades then my manual results will be bad, if they were more of losing trades then my manual trading results would be better because a losing trade missed is a winning one.

They could be loosing one too.

I didn't know that your bot is going with such long trades. Algo trading show true potential during scalping/MM - basically short term trading. F.e. my trading bot is MM on 7 trading pairs on small exchange, scanning order books each few sec, placing orders all day, filling 200+ trades daily for last 6 months. There is no way for me to be as fast with decisions, as fast with creating orders 24/7. Human would catch like 5% of this trades and
die of boredom after a week. Here true power of algo trading shows up. During trades that takes months bot is not that useful and its hard to evaluate its performance especially during bull run that we have now. You will need months if not years to draw valuable conclusions.

[jaberwock]

and so if those were supposed to be winning trades then my manual results will be bad, if they were more of losing trades then my manual trading results would be better because a losing trade missed is a winning one.

They could be loosing one too.

I didn't know that your bot is going with such long trades. Algo trading show true potential during scalping/MM - basically short term trading. F.e. my trading bot is MM on 7 trading pairs on small exchange, scanning order books each few sec, placing orders all day, filling 200+ trades daily for last 6 months. There is no way for me to be as fast with decisions, as fast with creating orders 24/7. Human would catch like 5% of this trades and
die of boredom after a week. Here true power of algo trading shows up. During trades that takes months bot is not that useful and its hard to evaluate its performance especially during bull run that we have now. You will need months if not years to draw valuable conclusions.

I am not entirely sure if we are talking about the same thing here, but I want to say basically something similar about this as well. Not only I would say manual vs bot would be the same thing except bot would be basically you but 7/24 non-stop with no emotions, basically a robot version of you, but I would like to also say no bot is magical, you make the bot what it is.

What trading bots are selling to you is the right to make yourself into a trading robot, that's it, no trading bot seller can guarantee you income, no trading bot seller could tell you "just put this in, put some money in, click start and collect cash" because nothing like that ever existed and will never exist neither, every single bot project owner can come to me and try to proof otherwise and I will guarantee you they will fail. What they are selling to you is just you as a robot, so if you are a bad trader, even if bot works 7/24 that will not make a profit no matter what.

[mikeywith]

I didn't know that your bot is going with such long trades.

It trades on the 1H chart for the most part, but the take profit points are very extended, and it uses a trailing stop so some trades do extend to well over 100%, but since my main focus is not to lose money so the initial SL kind of tight, so the bot can make 10 losing trades of small loses and then makes 1-2 large winning trades.

Algo trading show true potential during scalping/MM - basically short term trading.

I could never develop something that works on 1M or 5M time frames, even when the backtesting shows great results, the real testing shows other wise, given the nature of the crypto market I just don't think I can do with these small timeframes, but this doesn't mean I don't need the bot when trying to trade 10 pairs on the 1H chart, it still requires hours of monitoring and the worst thing is the entry and exit, it's very hard to be available once you need to exit.



If you look at the last couple of close positions, in my manual trading I only entered one EOS/USDT trade which resulted in a loss, the bot managed to get all of these trades on these two pairs, and despite losing 8 out of 12 trades, it still made 60.48% profit against 8.02% loss.

But I wouldn't get excited yet, this is a bull market era, most random strategies will work and outperform manual trading as long as they only open longs, the real challenge is the side-ways and bear markets.