Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Good eye as usual Alex.....yea, we need to change that so both have $150 as a minimum.
Thanks again for your continued interest and updates.
Much appreciated!
Good day, Myst:
I hope you're doing well!!!
My name's Pedro (AIM follower from Spain). I'm just writting to ask you where I can get some information to get the X_DEV spreadsheets, if available.
Thanks in advance.
Hi Myst,
I hope you are feeling better.
I have noticed that while the minimum $150 size “rule” is being applied to SELL trades, it is currently not being applied to BUY trades. Do you want the rule only to apply to SELLs or is it something we want to change?
Best regards,
Alex
Hi Myst,
How are you feeling? I hope a lot better.
I have sent you an XDEV sheet in which I updated the SELL Forecast column to include your $150 minimum limit on the number of dollars that can be sold to initiate a sell signal. So that the SELL Forecast column will show “0” if the signal calls for less then $150 dollars to be sold.
Best regards,
Alex
Hi Myst,
The mistake happens because cell AA4 in the Suggested Shared Bot/Sold formula rightly refers to cell H4 but as a result wrongly references the High % Dev Band multiple 5.53% in cell H4 in the yellow box.
Try manually entering "0" into Cell AA4. Do you agree that the problem is then resolved?
We might have to make that a hard coded "0" in all future sheets.
Best regards,
Alex
Hi Myst,
I have shared an updated version of the XDEV “1-day lag” sheet with you in which I have made one small change to formula in the BUY Forecast column. I altered the formula to show “0” when the forecast trade is smaller than the minimum signal % shown in cell AB2.
All the best,
Alex
Alex, I have updated the following sheet:
XDEV - YINN - DAILY - 21 4 21 - PRESENT - V2 - RL
Note that while the ROI is still negative it is beating the B&H'r by over +48%
I found something interesting as well. Not sure if it's a formula error but wondering if you could check.
Input 2,000% into the SELL multiple and 30,000% into the BUY multiple and see what you come up with in the % gain cells. Then check the "Suggested shares" column. Error?
Yes...it is all corrected. Working great!
Generally speaking I've noticed in strong uptrends you want to buy more and sell less, and vice versa in downtrends....ie, buy less and sell more. Therefor the buy multiple settings on uptrends will generally be higher than the sell multiple, and vice versa for downtrends (sell multiple higher than buy multiple).
I've traded in all scenarios, so yes....I've achieved results similar to what the sheet is displaying.
I have a good feel for when an optimization is good enough....or is actually optimal. The data charts at the bottom of the sheet are very useful. Things to look for are a good clean uptrend in the "B&H vs. XDEV chart" and also a good turnover between stock and cash on the Portfolio chart. The ethereum examples below are good examples of this:
The easiest way to get back in sync with the sheet is to change the starting "SHARES HELD" until it matches the current shares you actually own. (Change the top cell until the last cell in the column matches your current holdings).
Changing this metric does not effect any of the other parameters that affect ROI.
I have done this on multiple occasions with my 4 hour sheets.
Hey Alex....I'm feeling a bit better but still not out of the woods so I'll try to answer some of your questions.
I just gave your Tezos sheets a quick glance. Very nice results! I've done the experiments you have been doing many times in the past which is why I have a very high confidence in the program. I just wish you came along earlier and I had your improvements back then....lol. Even without them though the system has proven itself time and again.
I like to check optimizations every 2 to 3 weeks or so (on daily data), every week or so on my 4 hour charts. There is no hard and fast rule on these time frames so I'm not sure if these time frames are best, but it seems to work for me. I always want to make sure my signals are optimal for the current trend and make sure they have been adjusted if a trend change has occurred. Trend changes are easy enough to spot using the charts we have on the sheets but you never know if the trend change is gonna stick or just go back to the original trend in whipsaw like fashion. Frequent optimization checks help either way.
Do you re-optimised over an (expanding) full length of data when you trade?
Do you re-start a new chart using 3 months (or 2 x buy or sell signals) at some point?
Have you observed whether it is better to use longer strips of data?
3 months of data (daily sheets) is plenty of time to fully optimize any play. I will usually start with that then continue to use that sheet for around 6 months. After 6 months of the same sheet I will begin to think about cutting it back to 3 months again. I haven't found that the additional data is necessary to keep achieving high quality signals as long as you maintain 3 months of data or better.
Hi Myst,
I hope you will be feeling better soon.
I have shared two charts of Tezoz (XTZ) starting 14 6 20.
In the first, I optimised 15 9 20, using 3 months of data and then walked forward month by month without re-optimising the settings to see how well XDEV would continue to do in a scenario in which no further amendments were made to the settings.
It is reassuring to see a continued out-performance versus buy & hold using the initial XDEV settings, from using 3 months of data. XDEV performed well even when the settings were not changed.
In the 2nd sheet, I re-optimised the settings on 31 12 21 utilising the full strip of data back to 14 6 20.
Although the results in the first un-optimised play are impressive, the re-optimised results are more incredible still.
The question is whether re-optimising every month (or making small incremental changes weekly) brings one to this end result slowly over time?
I will, perhaps, try to do the same “walk forward” experiment again, this time re-calibrating the settings each month using the data from the months before.
Do you re-optimised over an (expanding) full length of data when you trade?
Do you re-start a new chart using 3 months (or 2 x buy or sell signals) at some point?
Have you observed whether it is better to use longer strips of data?
Kind regards,
Alex
Hi Myst,
I have shared a MARA sheet in which I tried an alternative approach.
Since you have more firsthand experience in the problems you encounter trading XDEV, you will be able to share what you think is useful.
Instead of calculating the size of the BUY or SELL % as a % of the Portfolio Value, I tried turning the $ values in the BUY Forecast and SELL Forecast columns into a % of Stock Value.
The Market Order is not always exactly what turns out to be traded as the BUY signal in the Market Order column sometimes exceeds the CASH VALUE available for a BUY.
I thought sizing relative to Stock Value might be easier to follow. If the STOCK you hold is different to what is shown in the spreadsheet then you will simply BUY or SELL a % relative to that amount of STOCK.
The question of how much stock to hold in the first place is one we will do more work to try and answer, perhaps calculating a maximum % of the total trading account and perhaps with reference, to the Zone % idea you discussed previously.
I would love to hear your thoughts on how and when to bring an investment inline with the $ value shown on the sheet, or, whether, in your experience, it is necessary at all. In particular, after a re-optimisation, as it is after a re-optimisation that the numbers can sometimes on the sheet indicating one's exposure can dramatically change.
Best regards,
Alex
Hi Myst,
I hope you are feeling better soon.
Further to this earlier message I shared another YINN sheet, the same as the previous version except that I added to the SELL Forecast formula so that it never forecasts a SELL larger than the actual current holding of STOCK shown in Stock Value - Column "V". There may be pros and cons to this. We'll see.
Best regards,
Alex
Hi Myst,
Sorry to hear you feel poorly. Get well soon!
Alex
Alex....I just caught the flu and feel like pooh
I may be out of commission the next few days. I will answer your questions as soon as I am able.
Hi Myst,
I have shared a sheet for YINN. Like LINK, which has gone down 50% in the period covered in your 4-hour sheet, this play has gone down about 60% in the period covered by the sheet. The optimisation is quite strange. I suppose because the daily data is less granular than the information you are trading on the 4-hour sheet. All the money in this optimisation was made in one trade taken in 27 July 2021. I am not sure if it would be successful going forward. Have you traded assets in a downtrend successfully in the past? Perhaps, if you don’t mind, you could optimise the sheet yourself and share the results?
Kind regards,
Alex
Hi Myst,
Did you get the BUY and SELL forecast to work?
The 4-hour sheet you sent me still did not seem to show the correct numbers.
Alex
Hi Myst
Thank you for the 4 hour LINK sheet.
I think you are trying the 4-hour time frame for the first time. Are you proposing to trade the LINK chart with the 4 hour timeframe going forward?
Optimisations for assets in a downtrend seem to be different to optimisations in an uptrend. Are there any observations you want to share?
Have you traded assets in a pronounced downtrend like this with XDEV before and achieved similar results? I'd be interested to hear if you have had similar results in the past.
With regards to the 2 drawdown column... On each row, for both XDEV and B&H, the column calculates how much the equity curve is in drawdown since your last high. Whenever you see a 0% (in black) it means your equity curve is back at all time highs or above. This allows for comparisons between XDEV and B&H in terms of depth of drawdown and of duration of drawdown. It is interesting to compare the timing and depth of XDEV and B&H drawdowns.
The top 2 cells show the MAX drawdown experienced over the period modelled in the sheet. The calculation in the cell simply looks through all the data below and finds the largest negative value (in other words biggest drawdown) and this is shown at the top of the column.
For example, in your 4-hour LINK sheet, if you look through the column below, you will see XDEV experienced a MDD (Max Drawdown) of -9.13% and B&H experienced a MDD of -51.87% over the same period.
Dividing the Profits (or losses) modelled in a sheet by the MDD during the same period allows us to calculate a reward to risk ratio and get a sense of the reward received for risk taken and make comparisons between assets in terms of reward to risk, rather than, simply return. This ratio is calculated in cells J7 and K7. This allows us to compare XDEV to B&H, but the ratios may also allow us to compare XDEV's performance across different assets.
As stated above, in itself, it is interesting to compare the timing and depth of XDEV and B&H drawdowns. I also wonder whether the Reward to Risk Ratios will help us in our optimisations to see when an optimisation is "good enough", or, when it is "under-optimised" or "over-optimised", i.e. curve-fitted. Maybe helping us find parameters that are "good enough" and robust and therefore will be effective going forward.
I read this comment online, which is why I included it. I quote, "The other factor I always like to look at is the profit/drawdown ratio. If this is anywhere near 1 [then] I see this as not a good strategy. This indicates that the risk and return are essentially equal, and if you have ended up ahead it is probably due to luck, not by your strategy having a statistical edge. If it is too high (say > 4) then I start to wonder whether it is over-optimised. The drawdown also provides a good guide for the expected future strategy performance, if in real use the strategy exceeds your expected drawdown it may be time to switch it off." https://www.forexfactory.com/showthread.php?t=263380
I am not sure how fruitful the ratio will be. It will be interesting to see if there is some correlation between this metric and other metrics that might indicate when an optimisation is "good enough".
I think your preferred metric is checking whether BUY or SELL signals occur at "extreme" moves in the DEV factor.
Alex
Hi Myst,
This is really helpful. The situation of my numbers not looking like the sheet exactly is already something I have across. For example, after re-optimising a sheet.
The question is what to do then? Do you have any tried and tested answers?
How do you try and get back into balance with what the sheet is proposing?
Alex
I added the Buy or Sell % column to your Bitcoin 4 hour sheet. (Column AC)
That way if for some reason your numbers don't match the sheet exactly you will still know exactly what % of your portfolio should be used for a buy or a sell.
Nevermind.....I just figured it out. I had the wrong cell number referenced. All fixed!
When I plug the Buy and Sell forecast into my 4 hour Bitcoin sheet I am not getting the correct numbers.
Not sure why.
I'll send you the sheet. Take a look when you have time.
No rush....thanks,
Nevermind.....I just figured it out. I had the wrong cell number referenced. All fixed!
When I plug the Buy and Sell forecast into my 4 hour Bitcoin sheet I am not getting the correct numbers.
Not sure why.
I'll send you the sheet. Take a look when you have time.
No rush....thanks,
I have a 4 hour LINK sheet I am going to share with you shortly.
Quick question....still trying to wrap my head around the Draw down columns. How do you arrive at the top 2 cell figures in those two columns? What do they mean?
Disregard the yellow shading in those cells you mentioned. It doesn't mean anything. I just forgot to change them to the same as the rest.
Hi Myst,
It is my pleasure.
I’ll check out the improvement in ROI’s on both.
I am gratified to hear you think we have improved the sheets and I am grateful to hear you think they are honing in on what your vision for X_DEV was meant to be.
I am sure we'll share many more insights.
I look forward to trading them with you!
Alex
Hi Myst,
I received the MSTR and 4-hour BTC and ETH 4-hour sheets. Thank you.
With regards the BTC and ETH 4-hour sheets:
See below:
If you look at the formula in J5, you'll see I multiplied 365 * 6. This is the number of 4-hour periods there are per day. I.e. 365 x 6 to annualise the period return. Doing this gives the correct Annualised Return in J5. In your ETH sheet 253.62% per annum.
The number "3.98" in J7 is a ratio, not a percentage, calculated by dividing the Profit / Loss by the Max Drawdown to express the reward in terms of risk taken, where risk is defined as max drawdown. In the ETH sheet 22.79% (PROFIT) / 5.73% (MDD) = 3.98. Showing a return of 3.98 x the risk during the period.
Is that clear?
I notice some of the 4-hour periods are coloured a darker yellow. Why is this?
Alex
Well done Alex!
Now it's just like the program.
Yea, that is a pretty complicated formula! lol
Did you receive my sheets? MSTR, BTC and Ethereum?
Hi Myst,
I have shared another iteration of the ETH sheet with you in which I have now added BUY Forecast and SELL Forecast columns.
I think I have managed to calculate BUY Forecasts which allow us to set limit orders in advance and incorporate actual CASH VALUE at the time. It was a little complicated.
You will see how, from time to time, the BUY Forecast will seem to be different to the BUY Market Order in column “O”. This is because the value shown in the BUY Forecast Column takes into account the CASH VALUE available at the time, so the value shown in the BUY Forecast column is actually equal to the CASH VALUE column available at the time in instances where the BUY Market Order suggested is greater than actual CASH VALUE. In these instances, the BUY forecast simply equals remaining CASH VALUE. (This works out to equal the amount shown in the next day’s Suggest Shares Bot/Sold column.) When we have run out of CASH and are 100% invested the BUY Forecast column simply shows $0.
Take a look.
Best regards,
Alex
One change I forgot to make on your sheets....
I color code the dev factor. Red for positive numbers which are generally sells and green for negative numbers which are generally buys.
Makes it easier to see if the larger numbers are being hit for signals which adds confidence that the parameters are optimized correctly.
If you want to make that change it's entirely up to you.
I've sent you my 4 hour sheets of Ethereum and Bitcoin.
I've seen incredible improvement in ROI's on both!
I think we have really improved the sheets greatly with your new insights and formula's.
They are feeling like they are honing in on what my vision for X_DEV was always meant to be. For that I am extremely grateful for your efforts!
Looking forward to trading these sheets in the new year!
Alex, I've just sent you my MSTR sheet with all of the updates.
The new formulas added +28% to the ROI!
I'll have to post the data in another post so any viewers that stop by can check it out.
Thanks again,
Really nice job Alex!
Thanks a bunch! Excited to try all the new formulas on my sheets. I'll share as soon as ready.
Hi Myst,
I have shared an ETH sheet with you that incorporates a new formula in the Shares Bot column. Take a look.
Doing this turned out to be a bit complicated. I also had to adjust the formula in the CASH VALUE column - column "V".
Previously, the CASH VALUE column also linked to the CLOSE column - column "E".
Now, the formula in the CASH VALUE column adds or subtracts from the CASH VALUE based on the correct BUY or SELL amount. Take a look.
In the ETH sheet I have shared with you, by correcting the CASH VALUE calculation from row to row, there is a dramatic performance improvement. I have yet to test this on other plays.
Experiment with the sheet and please confirm this is the case as well.
Aside from the above, the sheet includes your most recent SELL Market Order and BUY Market Order formulas.
The sheet also incorporates the “1-day lag” version of the formulas in the SELL Market Order and BUY Market Order columns. In other words, following the way XDEV is designed to work, it uses the previous day's XDEV factor in order to allow limit orders to be set in advance.
The above is certainly progress. However, at this stage, we still lack the functionality to set Limit Orders in advance.
We need to replicated the following on the sheet:
As shown in the picture, each row will have to show the amount to BUY and SELL, covering both eventualities in advance, to allow us to set limits orders.
I will try and add 2 new columns to the sheet and title them BUY Forecast and SELL Forecast and have them show the BUY or SELL amounts to set as limit orders.
Entering the OHLC information into the spreadsheet the next day will confirm which of the two options was hit - or none at all - and the existing calculations on the sheet will update the sheet in the usual way.
However, as in the picture above, we need to calculate both options in advance in each row in order to set limit orders in advance.
Kind regards,
Alex
Whether the results improve or not, it is the way the system should work so please send me a sheet with the new formulas as soon as you can. I'd like to work with those as I optimize my other sheets.
The 1 day lag is also the way the system is designed to work so I have made that change permanent. So far the results have been mostly positive.
Thanks for your patience!
Hi Myst,
It is not at all clear it will improve the results, but logically, sure, it needs to be changed. I’ll update the formulas to reflect the change.
I am look forward to feedback from you as to whether the “1-day lag” version of the spreadsheet is working for you.
I think it is good that both of us carry out the amendments and compare the result. The outcome is more likely to be objective. Two heads are better than one!
No need to apologise for the delay. I understand it is Christmas.
Just update me as soon as you are satisfied with your results.
Alex