- June 10, 2020 at 4:38 pm #9079
I hope everyone is safe and doing well.
Today I was shocked when a BTC buy signal was issued at $9,970.
Within the last 4 weeks, BTC crossed the $9,900 mark six(!) times and was repelled each time at around $10k with a significant drop. Instead of buying at a range of $9,350-$9,400 on June 7 (three days ago), the Crypto-ML algorithm for some reason decided to buy at an almost three month high only to see the price fall to $9,800 the next minute. The buy signal was literally issued when the price was at the highest point for a minute.
The last 9 months have shown how difficult it is to break through the $10,000 barrier and close above it. It does not take too much intelligence to know that buying at $9,970 will have a very low RoI if any. I would have expected to buy low at around $9,500-9,600 levels and sell around $10k levels in expectation of bouncing back as it has happened so many times.
Decisions like this make me wonder if the auto-trade is worth the money.
Thanks for your time.
- June 11, 2020 at 12:37 pm #9082
There seemed to be uncertainty on more than just the prediction engine. Independent analyses suggested it could go either way from the top resistance, which seemed reasonable given how often that resistance was tested over the past few weeks with whales seemingly cashed-out with the last drop. Change of say 20% gains (some claimed the next resistance past the 10-10.5K resistance was at 12K) with risk of dropping a bit or staying the same seemed reasonable with the model closing the trade with a 1-2% drop. While it has dropped further since that point, this is historically a tenuous point that is more unpredictable given the recent halving with frequent testing of the resistance.
- June 11, 2020 at 12:38 pm #9083
Whales being already cashed-out could also open the situation up to them cashing in and potentially overcoming that resistance.
- June 11, 2020 at 1:09 pm #9084
Thanks for your quick reply.
Yes, I understand that there could be a chance of breaking the resistance and moving up 20%. But issuing a buy signal at $10k or even above like on June 2 means that the ML algorithm is repeating the same mistakes again. I would understand a buy at a lower price like today ($9,300) knowing that the $10k level will be tested again and then sell at around $10k.
I understand that the price for BTC is unpredictable but I was hoping that ML-crypto would do a better job than my intuition. Whales tend to sell at $10k meaning that this is more the level to sell.
On the positive side, the algorithm sold at $9,800 so the loss was only 2% but in my view still unnecessary. Hopefully the AI will make better decisions in the future. I still have hope.
Once again, thanks for your reply.
- June 13, 2020 at 11:43 am #9087JustinModerator
This is a good question–ML is about probabilities. And what it’s trying to predict is price 24 hours from now. What likely happened is it played out a scenario where $10K is crossed. If that happened, price would like jump greatly.
Instead, price stalled out and the ML (not stop) issued a sell to back out of the trade.
I don’t think this was necessarily a bad trade and it was managed well.
Had $10K been breached, we may not have been able to get in the trade effectively.
As @Julian mentioned, it’s great to bounce this off of your own analysis as well. And given the signals were pretty mixed, this might have been one to sit out of.
- June 13, 2020 at 12:18 pm #9088
Are there any additional updates coming up for the models? Aside from giving the models more scenario training.
- June 13, 2020 at 2:00 pm #9089JustinModerator
@Julian, yes–there will always be updates. We have a few items coming soon that I’ll announce shortly.
And scenario training will also continue.
In my response above, I meant to say it’s a risk/reward probability that’s being calculated. So if there’s a chance for a large break, the ML may open a trade. I think that is what we saw in this scenario.
- July 27, 2020 at 8:54 am #9432
Hello Crypto-ML team,
Aftern long thoughts I have decided to end my membership with Crypto-ML. There were too many questionable decisions that turned out to be wrong in timing. I had the feeling that volume or past resistances were not included in the decision making. For example the algorithm issued a buy signal when the price spiked but was not supported by volume. As a consequence the price declined sharply after a few minutes.
And the last signal was a sell for no obvious reason to miss a $1000 increase. I did not understand too many decisions.
I really wanted it to work better so that I don’t have to worry but I ended up worrying even more.
So I believe the program is not yet where it should be.
I liked the quick replies by the team.
Thank you and good luck for the future.
- July 27, 2020 at 2:28 pm #9434sravanupadrastaParticipant
yes, i want to end my membership too. It was almost 1500$ increase in price but didn’t get a buy signal for BTCUSD.
- July 27, 2020 at 4:13 pm #9435
I personally will be keeping my membership. Using any sort of predictive tool is accepting that it won’t be perfect. This current peak is a demonstration. It is not a reason to see events like this and have fear of missing out and feeling like you were cheated. by signing up with this type of service you’re essentially accepting that you won’t be getting all the gains. The alternative is to invest into the markets on your own that is practically gambling because of how divergent the market can go with outlier events like this happening not infrequently, ie questionably making gains at all. I was reminded of this time and time again when I attempted to balance use of the service and my own trading skills. As many can attest on the service, this program does produce significant gains, more than you would get if you HODL. I believe a recent post stated a user who had 72% gains over the period of less than a year. The benefit of an algorithm running everything is because even if you did capture the gains with this current peak on your own, there is a distinct possibility emotion overtakes reason and you keep on holding on thinking that it could go further up. No matter how objective you are, this behavior can happen and cause problems as it has with my own trades on occasion. Inevitably, things will go back down and with how crypto markets work, whales typically sell around this Mark and things go down fast. For a number of traders, all their gains will likely be negated by whales selling because of greed overtaking investing reason. We are all free to do whatever we want with our money, including leaving this service. There is no need to publicly shame the company in the process considering users accept risk while signing up. It is not a professional thing to do.
- July 27, 2020 at 4:21 pm #9436
Thanks for the replies.
I did not want to discourage anyone who wishes to use this service and believe that anyone truly interested will join anyway. Still, I did not want to offend Crypto-ML.
I wished that the algorithm had worked better better but after four months I have to say that hodling and dca would have created higher returns (my personal opinion). Sometimes it was frustrating to see that the buy and sell signals came almost randomly.
Good luck everyone and I apologize if I have offended anyone.
- July 27, 2020 at 4:38 pm #9437
We all have a right to do what we want with our money and I too have the pinges of wishing I had ETH as it went from I believe 250 to 325. We are all professionals here and surely wishing that things were a bit different. It is understandable to want to find the best way to optimize assets and moving between systems to accomplish this. Trial and error absolutely includes using services such as this. What you stick with is personal and that is perfectly okay.
- You must be logged in to reply to this topic.