We’ve just rolled out a new update to make you decide where to fly in VATSIM easier and faster.
Major Update
New Flight Control Estimation Page
You can now request the expected ATC coverage for any origin/destination pair — not just your recommendations! Try it out here (must be logged in).
When you submit a request, you’ll get:
The expected control time for your route
The probability of each controller along the route becoming or staying online
Tons of extra details all presented in a friendly easy to digest way.
This gives you full visibility into how likely your route will have ATC coverage.
Minor Updates
Control Details Button: Every recommended flight now includes a handy “Control Details” button that takes you straight to the new estimation page — it’ll even trigger a request automatically!
Helpful Tips: Small pop-up tips will now appear at the top when requesting flight recommendations or control estimations (ephemeral but useful!).
Want to see it in action?
Check out this short video where I walk through all the new features:
What’s Next
We plan to incorporate the following:
Real flight routes in the candidate recommendation pool.
Ability to select in the UI what kind of candidates you want to rank (e.g. only rank real routes and/or popular flights).
ATC coverage forecast (similar to bookings but for all controllers! based on our predictions)
Autogenerate daily briefing.
…and more
Your feedback means a lot — feel free to share any thoughts or suggestions!
Hey Alex, I really love your tool as the major target for me when flying online, is a maximum of ATC coverage and VATSIM Dispatcher greatly helps with planning to have ATC on both arrival and departure.
I’ve just tried the flight control estimation from LFPG to LOWS and the UI is nice and intuitive.
Just wanted to get how the calculation of the probability works: How does VATSIM Dispatcher estimate a 14% probability of LOWS_APP being staffed in this particular scenario?
Hi Reinhard! Thanks for the feedback and question.
If I understood correctly you wonder why is 14% rather than 0% given that no clear activity (booking/events or past online session) has happened or will happen for that position LOWS_APP. If that’s the case, the model might be boosting the probability a little bit because of “things happening around this position”! (e.g., as you shown there the previous flight segment seemed to be controlled (not sure if by CTR or something else, but it is). Our model ingest not just information about the current position but also what is going on around it. For example, if there is an event in EDDM (or other relevant related positions), that might boost the probability of LOWS being online. In other words, it does not depend only on what happened (or will happen) in LOWS_APP exclusively. Good observation!
I’m actually adding some more improvement to the model, including for predictions close to the request time, i.e., prediction_ts ~ request_ts (in these cases current status of network, and more importantly, whether there is a controller online, might be more relevant than past interactions). Here is a sneak peek at how we measure performance for our models (and the improvements over time – purple is a very simple baseline)
In any case, please keep your feedback coming! Is greatly appreciate it and helps us improve. If you notice anything off (which it clearly might happen), please share with us with data points like this! It helps identify where the model might be lacking.
Hi Alex, thanks a lot for your insights!
Events and bookings in surrounding FIRs surely have an impact on other ATC sessions and it perfectly makes sense this way as it is something I’ve noticed too, that foreign ATCs tend to support one FIR’s event with ATC.
I’ve never really flown in the US, because of the lack of scheduled ATC sessions (nothing more depressing than flying 10hrs and having no one to welcome you on arrival).
Curious to try vatsim dispatcher for flights from EU to US and vice versa, when Bluebird’s 757 is released and I’ll finally do long hauls again.
Hi,
What’s the source for the names of the positions? For example, I was checking out this for a flight around EGLL, and we have the two positions labelled EGLL:GND and EGLL:GND - presumably because EGLL has three separate ground splits available, and recorded separately. Is this intended behaviour, because it seems rather unhelpful to pilots?!
Similarly for the enroute position above, there are three of those… which is odd because there is only one EGTT_S_CTR position, yet it is recorded three times. Again, is this recording top-down coverage of the sector in some sort of weird way?
Hope this is the correct place to report this
@1847294
Each bar represent a flight/control segment (~15min or so, don’t recall exact number). For each bar we show the most likely controller (in your case EGLL:GND is the top one in both segments you see). You can be under a control center for minutes and the probability of them being online changes with time (so that’s why makes sense to break it down into timed segments).
The labels refer to boundaries. There are multiple (complex) reasons why we do it that way but let me unpack it somewhat. In the end the only thing that matters is whether you’ll be in a controlled airspace at a given time (which is not trivial to find and the best and most accurate way we have to do that right now is by using their polygon). There are multiple sectors/polygons (including the “EGLL area”) and callsigns vary a lot, depending on the traffic amount. e.g., in an event you might (hypothetically) see something like LON_S1_CTR, LON_S2_CTR and since all have the LON_S prefix, they would have “the same (geometric) boundary/sector”, that is EGTT-S. In that case is better to predict on the boundary than the actual callsign (which will change a lot). In the end, whether that geometry is controlled by 1, 2 or 3 people is irrelevant 'cause you will have control. You just care whether any of those “sectors” will be online, irrespective of who will ultimately be in charge – to determine who to contact you still need to follow VATSIM protocol and check who is actually online when you decide to fly
In the flight below. The first flight segment could be controlled by any of those “facilities” (GND, TWR, APP, etc). We use boundary:facility (format).
Major Update
New Flight Control Estimation Page
Minor Updates
Control Details Button: Every recommended flight now includes a handy “Control Details” button that takes you straight to the new estimation page — it’ll even trigger a request automatically!
Helpful Tips: Small pop-up tips will now appear at the top when requesting flight recommendations or control estimations (ephemeral but useful!).
Want to see it in action?
What’s Next
(and the improvements over time – purple is a very simple baseline)
