I really don’t get the negativity here. Sure we should be cautious, and clearly some algos are better than others and clearly some market conditions will suit algos and some won’t. Algos are not free money and its incredibly hard to even code a winning “fake” algo over a long enough period.
But I was under the impression that the majority of members are here because they actually believe algos can make a profit – medium / long term. And if you dont believe that, then save yourself and all of us here some time and leave now.
Right, so.. hopefully the people now reading agree algos can make a profit.
So, your algo can make a profit, its ticking along nicely.. whats next? Sell it.
Are many of these algos junk? Yes. But will some make a profit? Yes. So if you want to buy, buy and if you don’t, stop moaning and take your negative energy elsewhere,
BTW did someone tried those Algodax algos? Is the seller legit and are allgos from them working? Just comparing them with cfd autotrading and prorealalgos they seem so much less expensive.
https://market.prorealcode.com/store/algodax/
For example, ALGODAX M1 looks like a perfectly optimized backtest without any losing months over the last 3 years, and the equity curve looks almost like drawn with a ruler. Now, May 2021 as the first month after the optimization period is the first month showing a loss. Surprise ?
For example, ALGODAX M1 looks like a perfectly optimized backtest without any losing months over the last 3 years, and the equity curve looks almost like drawn with a ruler. Now, May 2021 as the first month after the optimization period is the first month showing a loss. Surprise ?
No suprise there. Overoptimization. Let us highlight good algos with liveresults here
Both Pro Real Algos and CFD Auototrading provide trial weeks. My last rental was from CFD Autotrading and the first two weeks paid for the coming 6 months of all the algos.
Pro Real Algos offer trial periods regularly and CFD Autotrading still offers a two weeks trial period and backtest algos can be downloaded.
2 months sounds like a long time but maybe the competition will make it a standard later on
Did u use all the CFD-algos or just some of them?
Pro Real Algos: I chose a handfull for each trial period, haven’t rented any of them yet.
CFD Auototrading: Mainly running the “NQ Talos V2.1” it has been working well and I enjoy the telegram group with good info from the supplier
NQ Talos V2.1
only 2 weeks running NQ Talos V2.1 paid off the coming 6 months of all the algos?
For example, ALGODAX M1 looks like a perfectly optimized backtest without any losing months over the last 3 years, and the equity curve looks almost like drawn with a ruler. Now, May 2021 as the first month after the optimization period is the first month showing a loss. Surprise ?
It’s sadly normal. We speak about that on the french forum (sorry)
The most beautiful backtest you have, the most curve fit you have, the least corrélation between IS/OOS you will have
It’s like à perfect curve fit with à polynomial regression with n parameters. More n is high, more beautiful will be your backtest and the least corrélation you will have. That’s why we need to have ony 2 /3 parameters on algos
Sadly bis, in some ways it means thé least beautiful you have, thé Best corrélation you will have, but it’s against our natural instinct 😋
Bye
Did just 2 weeks of running NQ Talos V2.1 pay off the next 6 months of all algos? Kindly show us this fantastic gain !!! Put the gain you made as in my previous post …. you know I tell you this from direct experience. A friend of mine has a small family hotel, he made a hundred fake E-mails with invented names and then got five stars on “tripadvisor” …… I only believe what I see words are just words ….. how to see only graphs it shows nothing.
Hello
I did a backtest of 1million bars on PRA IBEX. How are they even allowed to sell this?
Why is not a MUST for a vendor to show all possible data for its customer (I am not only talking about previous post). I am talking about things in general that should be a must to become a vendor, maybe it is, maybe I am wrong. But 1 million bars of backtest I think should be a must in the information.
@
You clearly See thé période of backtest at right 😉
thanked this post
It did pay for the 6 months package from that supplier but those weeks are far from representative. As i wrote before, I was very lucky and it’s mainly a story of a successful coincidence.
Algos are tricky and choosing is not a walk in the park.
Backtest, walk forward and Monte Carlo are interesting but there are a lot of things affecting the market. I like to run several algos in different markets in order to minimize the risk of having many DD simultaneously.
Is this kind of information considered to accept on the marketplace ?
