ETH weekly chart using silver ratio retracements & 0.4142
Most of this is explaining why I like the silver ratio: 1+sqrt(2), or 1.4142. It may be a sort of analysis of ETH, but mostly I want to explain why I found the 0.4142 fib chart level so compelling as soon as I started using it (I tried several values and that one INSTANTLY had results, where all the others failed).
I was working on this when I saw the front page idea* with my fib chart retrace values! (note: i had to abandon my other account since this had my number attached so I could buy being able to look at two charts at once... x.x. so i am starting fresh, but I have been using sqrt(2) as a fib level since I was trying to figure out my own tweaks on the formulas. In fact, 1.4142 and 0.4142 are WEIRDLY good as fib levels. They are more accurate than 38.6. Because it's sqrt(2), it also fits nicely into the overall template of a fib chart that is mostly focused around ensuring the levels are being respected in general, so that once applied further, the same levels can be predicted in the same manner. I'm not being annoyed about that I would love to see 1.4142 used more often in fib charts; it's just a fantastic level. It's also related to a more complicated topic, enumerative combinatronics, which is quantifying possible patterns... which is perfect for pattern-heavy price action!
I initially was not very confident in TA. There are so many different techniques and it is very hard to figure out which are hopeful guesswork and which have an objective relation to the price. Learning incorrectly early on also harmed me, but it was an important lesson to learn. I am not a maths major so my explanations here are pretty minor.
youtube.com/watch?v=7lRgeTmxnlg
Here's a numberphile video about the silver ratio.
sqrt(2) is a Pell number. More formally, {delta} = {delta}S = 1 + sqrt(2). I wish TV wouldn't get mad at me for using math symbols and think I'm not writing in English but okay tradingview.
If you were to replace the fib numbers and make a Pell chart instead, it would look like this:
nevermind I made it and it was terrible so I scrapped that idea. I will need to mess with the pell series more because that many irrational numbers (19601+13860sqrt(2) = 39201.99997... ok)
puu.sh/HFeTV.png Silver rectangle (as opposed to the so-called ~perfect~ Fibonacci ratio rectangles, I find this one to be more respectable!!)
I'm still pretty new and only have a few months with very scattered education regarding these topics so most of it ended up being pragmatically learned and backtested. It seems to work so well that I second guess myself so often I end up making bad trades while it just follows my prediction almost precisely. I just use the retracement because the actual uses are sorta irrelevant to me as long as the ratios and levels I want are on the chart in the way I want them to consistently be used for my own purposes.
What's that? I can toss this set of retracement levels and it just manages to fit better than the defaults with zero fitting required! usually
Indeed, 0.2 and 0.4142 were probably my favorite discoveries in January when I was experimenting with different mathematics. A runner up is 0.8, only because it usually nests the dip after the 1 (100% value) is reached. As a trading style, the point is to buy in either for a short term trade between 0 and 0.2 and then sell at the silver level (0.4142), or, anything below the silver level is a buy zone, and 0.618 would be the sell point if there is a "projected" chart, which I like doing. It is, and always will be, weirdly accurate in predicting reversals, but never when, only the price. When is the hard part to me.
puu.sh/HFdOT.png
example, this is my bitcoin chart from my prediction in march and then now. I was expecting this sort of move but getting it to the exact day and recognizing that last flash crash before it happened (because it was repeating a previous impulse, so is probably some fundamentals I don't know about, whether it's monthly miner sales or just... taxes or whatever).
So using presumptions of the asset's usual impulse distances (since, well, the same people are buying and selling, and unless the price changes significantly, the movement in prices tends to always be correlated to previous movements, with the exception of mean reversion inevitability if it's on a wrong side of an MA or vwap or something. Basically, trading probability and the averages. So far it was worked well as long as there is no flash power outage in China that messes everything up. But even then it seems to be priced into the chart levels.
My next study is to see if I can integrate Elliot wave theory and these levels into a combined impulse prediction tool that I can use for myself objectively so I can stay in a trade with more confidence. Mostly, I gotta stay off the 15m chart.. lol.
Likely fallacy pitfall explanations for the retracement fit:
1. seeing patterns where there are lots of possibilities for patterns to be formed; of course, everyone ends up ignoring the values that are not met or are blown past, since they fail to create any support or resistance. I like to use these levels that are previously unused to predict where a channel may form, because an unused channel is more appealing to price action than a used one, given the volatility a used one entails with so many orders and predictions being made for the same ones.
.4142 and .2 and .8 are all slightly off of the fib levels. They are just off enough that instead of being overlapped, they skirt the level. This looks and is a lot cleaner on the chart, but may well just be seeing what one wants to see in how important the levels are. This is why I chose to focus on the 0.4142 level so heavily. It falls just short of the halfway mark (a very important value in probability given it is the median of the range I am trying to predict), which more or less confirms a halfway point. If the trend is very bullish, you can use the 0.4142 level as the guideline to try and predict where the 1 value will end up in the future.
I spent some time adding some more predictions and using those levels as examples of their utility in future price projection, as well as how well it can backtest.
Here is a step-by-step in how I do this for a quick guesstimate of bullish impulse prices that I find a bit more adaptable than simply the golden ratio fetish, and unsurprisingly, other such irrational numbers work well when used for different purposes.
BTC
s3.tradingview.com/snapshots/o/oLLGR176.png 1
s3.tradingview.com/snapshots/e/EqhyDrYf.png 2
s3.tradingview.com/snapshots/b/BTXD1EDl.png 3
AAPL:
s3.tradingview.com/snapshots/n/nMNGPxEZ.png ~base
~adjusted
~result
The final thing I have to say and want to express is that it's fantastic to learn how all these things work in the book, but the book doesn't matter if you don't have your own understanding and conceptual view of the technique you or I am using. Most of all, it should make sense to you, even if it doesn't make sense to anyone else. I've come to appreciate that with charting and the great variety of ideas and methods used by everyone, closer or further from the standard. It's worth noting the standards didn't have a high success rate anyway when tested in a vacuum; but this methodology is less for efficient autotrading, and more for having a plan and reasons for entering or exiting trades at certain points.
The most important element of charting, that I've come to identify, is that it cements the plan into a reference work that when changed becomes useless, so the plan must be stuck to unless price itself defies expectations beyond parameters.
Thank you! I'm happy to be able to shove this all in an article finally and sort of start to work out my ideas, as I have very little concrete documentation on it yet, as frankly it's still in the growing pains stages as I am but a bab in TA.
Sofie
*
My examples focus on an uptrend.
I was working on this when I saw the front page idea* with my fib chart retrace values! (note: i had to abandon my other account since this had my number attached so I could buy being able to look at two charts at once... x.x. so i am starting fresh, but I have been using sqrt(2) as a fib level since I was trying to figure out my own tweaks on the formulas. In fact, 1.4142 and 0.4142 are WEIRDLY good as fib levels. They are more accurate than 38.6. Because it's sqrt(2), it also fits nicely into the overall template of a fib chart that is mostly focused around ensuring the levels are being respected in general, so that once applied further, the same levels can be predicted in the same manner. I'm not being annoyed about that I would love to see 1.4142 used more often in fib charts; it's just a fantastic level. It's also related to a more complicated topic, enumerative combinatronics, which is quantifying possible patterns... which is perfect for pattern-heavy price action!
I initially was not very confident in TA. There are so many different techniques and it is very hard to figure out which are hopeful guesswork and which have an objective relation to the price. Learning incorrectly early on also harmed me, but it was an important lesson to learn. I am not a maths major so my explanations here are pretty minor.
youtube.com/watch?v=7lRgeTmxnlg
Here's a numberphile video about the silver ratio.
sqrt(2) is a Pell number. More formally, {delta} = {delta}S = 1 + sqrt(2). I wish TV wouldn't get mad at me for using math symbols and think I'm not writing in English but okay tradingview.
If you were to replace the fib numbers and make a Pell chart instead, it would look like this:
nevermind I made it and it was terrible so I scrapped that idea. I will need to mess with the pell series more because that many irrational numbers (19601+13860sqrt(2) = 39201.99997... ok)
puu.sh/HFeTV.png Silver rectangle (as opposed to the so-called ~perfect~ Fibonacci ratio rectangles, I find this one to be more respectable!!)
I'm still pretty new and only have a few months with very scattered education regarding these topics so most of it ended up being pragmatically learned and backtested. It seems to work so well that I second guess myself so often I end up making bad trades while it just follows my prediction almost precisely. I just use the retracement because the actual uses are sorta irrelevant to me as long as the ratios and levels I want are on the chart in the way I want them to consistently be used for my own purposes.
What's that? I can toss this set of retracement levels and it just manages to fit better than the defaults with zero fitting required! usually
Indeed, 0.2 and 0.4142 were probably my favorite discoveries in January when I was experimenting with different mathematics. A runner up is 0.8, only because it usually nests the dip after the 1 (100% value) is reached. As a trading style, the point is to buy in either for a short term trade between 0 and 0.2 and then sell at the silver level (0.4142), or, anything below the silver level is a buy zone, and 0.618 would be the sell point if there is a "projected" chart, which I like doing. It is, and always will be, weirdly accurate in predicting reversals, but never when, only the price. When is the hard part to me.
puu.sh/HFdOT.png
example, this is my bitcoin chart from my prediction in march and then now. I was expecting this sort of move but getting it to the exact day and recognizing that last flash crash before it happened (because it was repeating a previous impulse, so is probably some fundamentals I don't know about, whether it's monthly miner sales or just... taxes or whatever).
So using presumptions of the asset's usual impulse distances (since, well, the same people are buying and selling, and unless the price changes significantly, the movement in prices tends to always be correlated to previous movements, with the exception of mean reversion inevitability if it's on a wrong side of an MA or vwap or something. Basically, trading probability and the averages. So far it was worked well as long as there is no flash power outage in China that messes everything up. But even then it seems to be priced into the chart levels.
My next study is to see if I can integrate Elliot wave theory and these levels into a combined impulse prediction tool that I can use for myself objectively so I can stay in a trade with more confidence. Mostly, I gotta stay off the 15m chart.. lol.
Likely fallacy pitfall explanations for the retracement fit:
1. seeing patterns where there are lots of possibilities for patterns to be formed; of course, everyone ends up ignoring the values that are not met or are blown past, since they fail to create any support or resistance. I like to use these levels that are previously unused to predict where a channel may form, because an unused channel is more appealing to price action than a used one, given the volatility a used one entails with so many orders and predictions being made for the same ones.
.4142 and .2 and .8 are all slightly off of the fib levels. They are just off enough that instead of being overlapped, they skirt the level. This looks and is a lot cleaner on the chart, but may well just be seeing what one wants to see in how important the levels are. This is why I chose to focus on the 0.4142 level so heavily. It falls just short of the halfway mark (a very important value in probability given it is the median of the range I am trying to predict), which more or less confirms a halfway point. If the trend is very bullish, you can use the 0.4142 level as the guideline to try and predict where the 1 value will end up in the future.
I spent some time adding some more predictions and using those levels as examples of their utility in future price projection, as well as how well it can backtest.
Here is a step-by-step in how I do this for a quick guesstimate of bullish impulse prices that I find a bit more adaptable than simply the golden ratio fetish, and unsurprisingly, other such irrational numbers work well when used for different purposes.
BTC
s3.tradingview.com/snapshots/o/oLLGR176.png 1
s3.tradingview.com/snapshots/e/EqhyDrYf.png 2
s3.tradingview.com/snapshots/b/BTXD1EDl.png 3
AAPL:
s3.tradingview.com/snapshots/n/nMNGPxEZ.png ~base
The final thing I have to say and want to express is that it's fantastic to learn how all these things work in the book, but the book doesn't matter if you don't have your own understanding and conceptual view of the technique you or I am using. Most of all, it should make sense to you, even if it doesn't make sense to anyone else. I've come to appreciate that with charting and the great variety of ideas and methods used by everyone, closer or further from the standard. It's worth noting the standards didn't have a high success rate anyway when tested in a vacuum; but this methodology is less for efficient autotrading, and more for having a plan and reasons for entering or exiting trades at certain points.
The most important element of charting, that I've come to identify, is that it cements the plan into a reference work that when changed becomes useless, so the plan must be stuck to unless price itself defies expectations beyond parameters.
Thank you! I'm happy to be able to shove this all in an article finally and sort of start to work out my ideas, as I have very little concrete documentation on it yet, as frankly it's still in the growing pains stages as I am but a bab in TA.
Sofie
*
My examples focus on an uptrend.
Publications connexes
Clause de non-responsabilité
Les informations et les publications ne sont pas destinées à être, et ne constituent pas, des conseils ou des recommandations en matière de finance, d'investissement, de trading ou d'autres types de conseils fournis ou approuvés par TradingView. Pour en savoir plus, consultez les Conditions d'utilisation.