Yang Zhang volatility dashboard

This thread here...reminded me of something that I meant to get round to, and as Moodybot posted a panel that's well-coded (I vibe-code, let's not pretend I wrote this) - I tinkered with it to provide something I needed. The trouble with using %s is the variance in volatility, but the idea of extracting efficiency and working out what's cheapest to trade is a good one.

so to this... maybe useful for others. It's a bit rough around the edges I know but it does the job for now.

This panel shows you the usual standard fare; you can change account size, local currency, risk and SL to give you lot size and value per pip. The UI (cog) doesn't work btw - you need to adjust through indicator settings. I know, I know...

But it does show you predicted volatility - the high-low range of the day - starting from midnight London (always).
ATR is the most basic method, but to rev it up I'm using Yang-Zhang model
It won't blow your socks off, but it's pretty well established and beats using an ATR.

YZ Vol is telling you what the range is likely to be in pips from London midnight.
Realise is how far it's travelled today
% real is % of expected travel
Efficiency is the proportion of spread to expected range. Ie how much of your trade will be eaten up by costs were you to get the low and high of the (expected) range. It uses MT5 spreads so not massively useful but a rough ballpark. It's then benchmarked against EU as 1 - so you can see NZDCHF is 60 times more expensive than EU, and it's cheaper to trade NZDUSD and USDCHF as separate legs than the cross.
Also shows you position size according to risk/SL.

And click the YZ Vol figures and it'll tell you how accurate its predictions are.

Anyways... hope this is useful to someone and thanks to the threadstarter and MoodyBot for their graft.

If OP ok I'll place an Mt4 version of the YZ Dashboard.

Moody

Man you are prolific. Thanks MB

I can see this is thrillingly popular with the FF crowd but my God is this useful.

This is my first and last update on this as I've got to merge my other dashboards with some magic jiu jiu and MT5 won't cut it. It's worth pointing out for vibe coders like me if you start with good code (MoodyBots in this case), it makes subsequent changes using AI a breeze. These changes took about an hour.

I've added two currency meters, one that tracks since midnight London, and the other the last 60 minutes. And a special calendar.

Why special? It shows news t0 and t-1 (today + tomorrow) and highlights news exactly 1440 minutes away, plus or minus a buffer in minutes (set in settings). If Friday, looks to Monday. And why is tomorrow's news today's problem?

There lies the edge. It comes from TWAP, and much as I use a lot of esoteria, this one is firmly in the algo camp. Size orders are - usually - promised in VWAP (this causes algos to bounce off VWAP lines after a move outside of VWAP parameters). But, generally they get delivered in TWAP, usually calibrated to a 24-hour cycle because it's neater. In order to create a position before news, you don't wait until the release, you work it in, slowly.

So fecking what? Look closely. Euro news today: 8:15 (& 8:30/9). What time did EU bull move start yesterday? And what time the USDCAD (due 1:30pm today) plunge yesterday? Tomorrow's 6am UY? Not always so perfect of course, see GBP news at 9:30 today kicked off 9am yesterday. Tomorrow's 8am EU at 7:50. Best to look at the whole FX spectrum and zero in on the first major and use a buffer. It defaults to 60 minutes pre and post, but in practice 30 minutes is about right.

Anyway, that'll do. Crack on and enjoy.

Thank you for this resource, but how do I enter trades with this system?

It's not a trade dashboard, it's an info dashboard if that's what you mean. Or if you mean a full trading system - it's down to you. I've got a completely different use for the vol figures than most (an industry benchmark) but for anyone using ATR to manage trades, this is at least 25% more accurate at predicting the range, so it's an instant improvement
For the news trades, you need to look and study and figure out if you'll trade reversals from the time slot, or breakouts from the moment in time. It's the sort of thing a good quant could test in an hour or two, so you might want to find an enthusiast to quantify it. I use it as a warning signal for increased volatility and chance of reversal, but as part of a bigger system.
As I said, I created it for my use but I'm sure some of you can get something out of it too.

That's a great contribution triphop, thanks for sharing.

Do you determine some kind of value area around the VWAP, in effect ahead of the actual release?
IIRC you use prime fx pool volume data, have you also seen directional clues around that time, similar to futures book (icebergs, aggressive orders etc)?

you're welcome. Re VWAP I don't use it tbh but appreciate its importance. I saw BlueSteele's thread and the usual naysayers telling everyone how it's not applicable so just adding some weight to why. Futures vol is logical and sane it's just i derive value from a method that's a bit more out there. But still maths. Horses for courses.

re fx pool volume data -yes in short. Almost every single move (I call them legs, moves of a few hours) - every day - is dominated by one bank and received by 2 or 3 others. All the others are still market making just they're generally delta neutral. When one takes off (ie, taking one side only, increasing in frequency) it's likely to continue, and especially so around one of these time zones.

I finally got round to converting your Dash into MT4, the news filter still uses a csv file which has to be downloaded once per week but its functional.

Running 24hrs before the move ....... now thats an edge.

I'm not going to post it on here, it's too complicated to explain the setup etc, but for those who put the work in, it's worth taking the time.

Moody.

Top work. You deserve to get something decent out it for the coding hours you put in alone.

For anyone using it, although I said I'd not post any further updates, after wrestling with APIs and trying to make a framework with my noddy coding skills, I realised I can live with MT5 for a bit longer, so big upgrade to a full scale FX dashboard in next couple of weeks.

A quant - which I am absolutely not - told me the YZ implementation was an abomination. I thought it was pretty decent, and it does outperform ATRs. But YZ should only be used on historical vol not the forecast. We all knew that right? Right. An HAR-RV model coming up next anyways for the geeks, and for traders- it's basically the most accurate forecast of volatility for FX there is (that's out in public).

I'll keep an eye on your thread, if there are any Quants operating on here, be nice to hear your thoughts.

M.

The big Kahuna - full FX dashboard - update

So, as mentioned this early dashboard sparked an idea - to replace as much of Eikon as humanly possible. Eikon - which is Bloomberg's rival that they're no longer worried about - has the worst UI I've ever encountered and I've lived with it for years. Truly abysmal. Now is the time to not only replace, but improve it.

I'll stick a feature list on the new dash when I post it in another thread but for now, here's the problems I was looking to solve. The main aim is to get a very quick read on the whole market, then to the majors, then to the pair itself. Three stages. With no fluff, and accurate metrics. Risk, volatility and direction.

1) I want to know quickly where money is flowing between bonds, equities, commods and currencies. I also don't want arbitrary timeframes like 4 hours, I need to know by session - asia / London / US
2) I want a read on which currencies are moving most across the currency universe but every retail strength meter I've seen uses DXY weightings which are wildly inaccurate and about 25 years out of date.
3) I want to know my jump risk on any trade, ie the chances of price gapping through my stop and how much damage it would do if it happens
4) I want a best-in-class forecast of the volatility - estimated high/lows and open to close - for the upcoming trading day. I want historical accuracy reports to ensure it's performing as expected.
5) I want to know what high impact news is coming up today and 24hrs later and have timezone aware additions of the cut, roll and fix
6) I want to know current timezone/sessions for equities and currencies without thinking and have it update with DST
7) I'd like a pip calculator available at all times displaying size a/c to SL & risk values
8) I'd like to view all relevant information for a pair by a pair without having to sift through irrelevant news.
9) For each pair I want to see related assets, how they're moving and their coefficient (how correlated to the underlying they are)
10) I want a countdown to the next high impact news event for any pair.
11) I want a visual representation of volatility for each pair of how much of the expected move is complete
12) I want this dash to be robust, cache properly, and not hammer my machine and or degrade violently when it can't fetch data
…. and three more for the next version:
13) I want volatility predicted by hour using intraday seasonality.
14) I want to launch a chart with all predicted vols pre-marked out, news now and next
15) I'd like an initialisation screen to detail progress in all the complex calculations and data retrieval.

An expensive and difficult set of requirements. I realise this is all me/me/me but I'd be amazed if the vast majority of traders wouldn't find this relatively useful. It's a hell of a lot cheaper than buying a BBG terminal that's for sure.

I've got 1-12 working broadly as expected bar separating out high-low/open close on the vol figures but some of the metrics are out. Work in progress... Here's a couple of screenshots for now.

Most, including me, wouldn’t have a scoobie on how to interpret the complexity of your dash.
Hope you get it completed the way you want it to function.
AI has now reached a point where just about anyone can turn out a very capable piece of software, all they have to do is try but what you are
doing, that’s a whole new level.

I’ve been jumping between ChatGpt and Claude with the latter now my go to, I’m finding it more capable and easier to interface with than Chat.

Keep up the great work

Ha thanks man don't forget it's got a MoodyBot engine running at its core. GPT reckoned a reasonable estimate for the whole dash was £100k, and if the rest goes as planned - probably 4 days work total. Wild.

Agree with you on Sonnet over GPT. GPT writes great specs though and excellent maths logic, so I spec there, then use Kiro for the implementation which calls on Sonnet plus god knows what else.

Generally been amazing though occasionally acts like a Tellytubby when you ask it to delete a rectangle ("I've deleted the entire file you're welcome!") so not quite plug and play yet...

Keep you posted

And we're on. Mission and half of 15000+ lines of code but I think it's worth it
https://www.forexfactory.com/thread/...-dashboard-mt5