If you're including Starship you'd be fitting the sum of two exponentials (according to your model) together, since their launch cadences are quite decoupled as demonstrated by separate launch and integration teams.
Quote from: Robotbeat on 12/08/2023 04:43 pmQuote from: spacenut on 12/08/2023 03:54 pmThe Soviet Unions launches included all of their launch vehicles. Sure the R7 family got the most, but what about other styles of rockets? Has F9/FH family beat the R7 family of rocket launches?Falcon so far this year has beaten the 64 or 65 launch per year maximum that the R7 family had achieved in like the 1970s or early 1980s. If you were going to count all the Soviet union launches in a year, you need to count all US launches in a year and possibly all US plus ally launches (as the Soviet union was nominally (stop laughing!) a bunch of countries).Regardless, next year SpaceX could break the Cold War era total global annual launch record ALONE.they are broken this year only if you count 6 Rocket Lab NZ launches as by usa
Quote from: spacenut on 12/08/2023 03:54 pmThe Soviet Unions launches included all of their launch vehicles. Sure the R7 family got the most, but what about other styles of rockets? Has F9/FH family beat the R7 family of rocket launches?Falcon so far this year has beaten the 64 or 65 launch per year maximum that the R7 family had achieved in like the 1970s or early 1980s. If you were going to count all the Soviet union launches in a year, you need to count all US launches in a year and possibly all US plus ally launches (as the Soviet union was nominally (stop laughing!) a bunch of countries).Regardless, next year SpaceX could break the Cold War era total global annual launch record ALONE.
The Soviet Unions launches included all of their launch vehicles. Sure the R7 family got the most, but what about other styles of rockets? Has F9/FH family beat the R7 family of rocket launches?
Quote from: Robotbeat on 12/12/2023 05:40 amOK, this is dumb to argue about. But if you’re asking how many launches SpaceX does, that 100% includes Starship launches. If you’re talking just about Falcon, we’ll that’s Falcon.The OP whose tally we were all comparing against said it in this thread's very first post, before any Starship flight, and confirmed it about a month ago: both Falcon 9 and Heavy are counted but not Starship If you're including Starship you'd be fitting the sum of two exponentials (according to your model) together, since their launch cadences are quite decoupled as demonstrated by separate launch and integration teams.
OK, this is dumb to argue about. But if you’re asking how many launches SpaceX does, that 100% includes Starship launches. If you’re talking just about Falcon, we’ll that’s Falcon.
Quote from: Robotbeat on 12/12/2023 05:40 amOK, this is dumb to argue about. But if you’re asking how many launches SpaceX does, that 100% includes Starship launches. If you’re talking just about Falcon, we’ll that’s Falcon.I do find it hilarious though that the same crowd who were poo-pooing the possibility of SX ever getting to 100 this year and called this thread stupid, are coming back and arguing the technicalities of a couple launches now that they are within spitting distance of achieving it…
Another OTV-7 delayRolling 365 day total stays at 96 +2YTD total stays at 91 +2
SLC-40 launches:
Maybe we can put biases and last-minute grasping at straws to rest
She noted the company has a goal of 100 Falcon launches this year. “If we can do 100 flights of Falcon this year, I’d love to be able to do 100 flights of Starship next year
https://spacenews.com/shotwell-says-spacex-ready-for-starship-static-fire-test/QuoteShe noted the company has a goal of 100 Falcon launches this year. “If we can do 100 flights of Falcon this year, I’d love to be able to do 100 flights of Starship next year
It's clear the rate has increased over purely linear (which predicted high-80s / low 90s back in 2Q2023) in the last two-to-three months, in a more progressive fashion than prior arguably step-like rate increases. It's also clear visibly exponential single-year fits are more unstable and prone to overshoot than linearizations of the rate in spite of the above, with some models predicting up to 110+ flights. They also fail to smoothly link up with exponential fits from prior years (in other words, reparametrization is needed in any case, as with the simple linear trend), rendering their potential regarding predictive power at the same level than the linear tallies, if not worse because of the bias towards overshoot.
And the linear curve fit of launches to date has not ventured outside of 84 - 85 total launches for the year for at least two months.
As of the 16th of May, they’ve had 32 launches so far. On track for 101.8 launches this year given a gradually but continuously improving launch rate (not saying any particular rate, just saying something about the form of the equation) and 61 launches last year.
Literally he said Falcon flights. Not all flights. And the first one quote tweeted Falcon, too.
Quote from: Chinakpradhan on 12/12/2023 02:08 amQuote from: Robotbeat on 12/08/2023 04:43 pmQuote from: spacenut on 12/08/2023 03:54 pmThe Soviet Unions launches included all of their launch vehicles. Sure the R7 family got the most, but what about other styles of rockets? Has F9/FH family beat the R7 family of rocket launches?Falcon so far this year has beaten the 64 or 65 launch per year maximum that the R7 family had achieved in like the 1970s or early 1980s. If you were going to count all the Soviet union launches in a year, you need to count all US launches in a year and possibly all US plus ally launches (as the Soviet union was nominally (stop laughing!) a bunch of countries).Regardless, next year SpaceX could break the Cold War era total global annual launch record ALONE.they are broken this year only if you count 6 Rocket Lab NZ launches as by usaThose should count towards NZ, IMHO.Electron is launched, serviced, and mostly made there.Neutron should count towards the USA though when it finally launches for those same reasons though.
That's actually the first time I've seen a direct quote from someone in SpaceX leadership that unambiguously states they have a goal of 100 Falcon launches for 2023. Thanks for digging it out.
If the long term trend is an exponential increase then any linear projection from a segment of it will be on the low side, even if there is no noise in the data. It will then start correct itself in the second half of the period in question, but will only get it "right" at the end. There was no way the linear projection back in the first few months of the year could correctly have predicted the high 90s total we are seeing now: it's not just prone or biassed to underpredicting, its basically guaranteed.
QuoteIf the long term trend is an exponential increase then any linear projection from a segment of it will be on the low side, even if there is no noise in the data. It will then start correct itself in the second half of the period in question, but will only get it "right" at the end. There was no way the linear projection back in the first few months of the year could correctly have predicted the high 90s total we are seeing now: it's not just prone or biassed to underpredicting, its basically guaranteed.Any projection will do just that, with larger errors the more noisy the process is in the analyzed timescale, and the larger order its derivative is modelled as - except when using an analytic trend on a process always following that trend with minimal noise. The linear trend will have underestimated the true pace by around 8-10 launches (87-89 projected in April vs say 97 this year), which is about 9-10%, at the same time the exponential fit, with far larger oscillations, was overestimating the eventual true pace by around 12 launches (110+ projected around the same timeframe), which is around 10-11%. You're cherry-picking the quotes, while ignoring the true divide I posted a while back. Not gonna do the research for you again, the quotes are right there in the September timeframe.
Back on topic, with all the recent scrubs it looks like "100 F9/FH" is almost certainly out of reach. The only possible way left is something like this - ...
100 Falcon launches is now relying on at least 2 both of:- 3 in 17 days from SLC-4E - the record here was 8.03 days until the most recent launch broke it (Dec 01-Dec 08 = 6.57 days)- another big new record - basically only one option: 3-day pad turnarounds at SLC-40- 2 in 18 days from LC-39A - the record after a FH is 17.38 days (Jan 15-Feb 02) so another launch on Dec 30/31 would have precedent - 5 in 15 days from SLC-40 - with a 3.93 day pad turnaround record delayed too much
Quote from: eeergo on 12/13/2023 02:33 pmQuoteIf the long term trend is an exponential increase then any linear projection from a segment of it will be on the low side, even if there is no noise in the data. It will then start correct itself in the second half of the period in question, but will only get it "right" at the end. There was no way the linear projection back in the first few months of the year could correctly have predicted the high 90s total we are seeing now: it's not just prone or biassed to underpredicting, its basically guaranteed.Any projection will do just that, with larger errors the more noisy the process is in the analyzed timescale, and the larger order its derivative is modelled as - except when using an analytic trend on a process always following that trend with minimal noise. The linear trend will have underestimated the true pace by around 8-10 launches (87-89 projected in April vs say 97 this year), which is about 9-10%, at the same time the exponential fit, with far larger oscillations, was overestimating the eventual true pace by around 12 launches (110+ projected around the same timeframe), which is around 10-11%. You're cherry-picking the quotes, while ignoring the true divide I posted a while back. Not gonna do the research for you again, the quotes are right there in the September timeframe.No, a curve (correctly) modelling a curve won't, but a line approximating an upwards curve will. That's the whole point.In the attached illustration, the red line is always above the blue line. It has to be, because the blue line is curving upwards. If you are tracking progress against the red line, but actual progress is on the blue, you'll start off looking like you are underperforming.If the underlying trend was linear then the linear modelling would give a better fit. But it isn't.Yes, I cherry-picked quotes, but deliberately, to illustrate the point: back in May the linear trend had been under-estimating by at least a dozen solidly for a couple of months. RB's exponential model was very noisy, projecting all kinds of numbers, but was able to get within maybe 4 or 5 at around the same time. This is entirely expected, because the linear model is expected to underpredict at that time. The linear projection line will be above the curve at all points apart from the start and end, given that it is curving upwards.
I have shown that concept upthread too. Practical issues render it moot: the noise is too large in an annual series, and the exponent too small, to appreciate the difference. Plus, the exponent changes every year: a given year's trend is not predictable with the overall fit (due to, again, noisy periods caused by ramp-ups or slow-downs driven by external factors -such as the current weather delays- or internal to the processing -such as manifest juggling, technical issues as with FH's, or directives to launch more often, as with the marked increase in the past two months). The *multi-year* fit IS predictable and fits better to an exponential -or, soon, sigmoid- but with significant excursions away from it. The *yearly* fit is, again, so noisy that a linear extrapolation or an exponential fit can describe the trend with similar accuracy, yet the purely exponential tends to significantly overshoot more than the linear undershoots. Predictive capabilities of such fits fail short the steeper the functional derivative is.
What, exactly, is in contention here?
With 1 likely Starship flight, then to reach 100 we likely need 5 from the West Coast. There's 4 in the fleet, where can #5 come from?~ Default option: After B1063.15 flies, B1063.16 just has another fast-ish turnaround. (or possibly B1061.17 then B1061.18)~ Might B1082.1 be commissioned, ready to launch in 2023, and go to the West Coast fleet?~ Will they just send one from the East Coast fleet e.g. B1073.12, which used to happen often?<snip>(Note the 4 or 5 consecutive flights with 40-43 day turnarounds, which is short compared to the East Coast fleet. I can imagine a 5th West Coast booster happening sooner rather than later)<snip>Reserve:- B1076, last flown 2023-11-12, EOY -49 days (East Coast, currently landed on ASOG)- B1082, currently in testing/commissioning. Ready for flight this year?- * next West Coast landing to be turned around quickly? (B1063.16 or less likely B1061.18)
