Same here, Sailor and Engineer, thank you Mozzy. And good you gave some insight on these graphs cos some of them were not clear for me when you showed them previously.
As a Wind Data engineer (for wind turbines not sailboats unfortunately hehe) and an passionate sailor, I love these videos and your explanations. You make it simple and understandable for all. I also believe more can be extracted so I hope I will find the time in the next days to look at the data myself. Forze Luna Rossa, Buon Vento! ⛵
@@richardmyers6054 that's pretty cool! Curious to know the field. Discovered this channel too late, but better late than never :) Getting super excited for the semi's!
Hi ! sorry to bother you, I´m a soon-to-be great data analayst :-) and I´ve started to extract and clean the raw data, but I have so many questions ... do you mind if I ask you if you spent some time on it ? Thanks!
Thank you so much for doing this. I had been struggling to understand this while you had been talking the last few episodes. Really appreciate it, and great job with it!
All I've seen at this point is the thumbnail, and I've already liked the video. Sure,.it's a Mozzy vid. It's gonna be good. But most importantly: I couldn't agree more. Data is the key to everything. And data has been painfully absent from the cup coverage through official channels
It's deeply disappointing that we're seeing a large amount of real-time data being withheld from viewers. We almost never see the VMGs, which are the only relevant speeds. They don't show how much hydraulic pressure the boats have. We're not seeing any ladder lines. This is quite depressing, because the owners obviously think we are morons, and we can't be trusted to understand what's presented as real time data. I find it quite insulting, and deeply frustrating. I'm loving the racing, but I could be loving it SO MUCH MORE!
With Virtual Eye, accesible on the AC website, you can see all the matches. there you can see some data in real time. I personaly dont watch them live. I check some phases afterwards.
Hey Mozzy, thanks! Good detail to better understand the performance and stats Would be great to see this spread out on an excel with KPI headers, EG; Upwind speed per knot, Average speed loss per tack/gybe, average upwind angle, etc Keep up the great work mate - Go the Kiwis!!!
Thanks so much for sharing the links and the explanation of the various plots. The Race Viewer is a lot of fun and I wish I had more time to mess with the raw data!
Hi Mozzy, as a graduated Aeronautical Engineer, working in the Oil&gas field indeed😅 as a non-destructive testing engineer, I want to express all my respect and apprecition for your contents. I have been following you since your first AC videos. You are like a friend yo me now. Many thanks mate. Cheers
Another great video which gets the grey matter going. The part on tracking the tacks and gybes is particularly interesting. One aspect of these manauvers which is probably also important are the interial impacts in terms of potential kinetic energy wasted and also how much it is used to load and unload the foils during the crossover. This would also impact location of mass in the design and as such is going to be difficult to adjust for except through how the boat is steered and trimmed during a manoeuvre. Maybe the subject for another geeky video?. Keep up the great work, much appreciated.
Terrific video. I had reasoned out most of this from your previous discussions of the graphs, but it's good to have a definitive summary. Thank you! I would have simply described distance gained as "the integral of VMG", but maybe that's too techy? Then the "distance lost" metric in Nord becomes the integral of steady-state VMG during the duration of the tack (i.e. what would have happened if you hadn't tacked) minus the integral of the VMG through the tack. For the comparison to be valid you have to look at the entire time that it takes to get back to steady-state VMG. Your point at 10:30 boils down to the fact that Ineos isn't bavk anywhere near steady-state VMG when the graph ends, so you don't see the additional ground they lose during the acceleration after the tack.
I feel like 20 seconds ought to be enough to be back to steady state. They're only on a tack for like a minute at a time. Honestly I think this graph is a lot more useful to the race engineers than 'distance lost' as what you're looking for here is the finer points of how to make that line as straight and steep as possible.
@@weatheranddarkness I think we've already seen plenty of instances where it takes long than that to get up to speed. Accelerating an AC75 seems to be a complex process, since the apparent wind and therefore required sail trim etc changes drastically as you pick up speed. Put another way there's a significnat sequence of "modes" that they have to transition through to get to steady state.
@@patrickchase5614 True, true but on the whole, in a race, it seems like the average time back to 'steady state' is well shy of 20 seconds. If you're pushing 20 seconds you're probably reaching just to stay on the foils. If you can point me to one of these graphs that paints a different picture I stand to be corrected.
@@weatheranddarkness Taking >20 sec to accelerate is exactly what the graphs that Mozzy discussed in this video at 6:40 or so are showing. The VMG graph on the left side shows that both Ineos and Prada are still accelerating (and neither have recovered to their entry speed) when the graph cuts off at +20 sec. In other words, both AC75s consistently needed more than 20 seconds to get back up to speed in _the very race that this video uses as an example_ . Did you actually watch the video and listen to Mozzy’s observations before you replied here?
Although the last time I sailed was some 40+ years ago and certainly are no sailing geek I thoroughly enjoy the videos; fascinating how complex and 'intertwined' most aspects of the boat are and how to get the maximum performance from it. Much appreciated....go team NZ.
BIG TIP : The "time spent at, for example, VMG upwind" plots are better understood when you turn them through 90 degrees. Then you see a more familiar statistical distribution! And I assume the mirrored look is port/starboard?
@@weatheranddarkness I had just thought that it was a style of distribution to mirror it left/right. They seem perfectly symmetrical, at least as far as I can tell, so either they sail exactly the same numbers on both tacks or it's just a way to represent the data. Let me known if you see any asymmetry that I'm missing. Maybe Mozzy can comment. He's the man!
Great video. Even with the data shortage, there's still so much to dive into, but yes those can't angles would be interesting to unpack why we're seeing some of these differences.
Thanks Tom, while I understand most of the metrics that you talk about, I haven't got how the graphs apply, I look forward to you revisiting this again in more induvial detail. Watching you sail and describe sailing the AC40was nearly as exciting for the viewers as it was for you,.so before they pack up and go home Mozzy Sails need to sail a AC75.
Thanks for the explanations. Have been looking at these after seeing them referenced on the Sailing Anarchy forums but hadn't ralised the second page race viewer was there -doh! Looking forward to your analysis once you get more data, but also curious about your comments on the difference in leeway. Some races/boats were showing big differences.... Thanks Mozzy :)
Years ago when using calculated TW values from on board instruments (Ockham) we had to use "double smoothed" values to get coherent values and trends. This was implicit in the development of "Wally". Nice to hear you mention VMC/VMG.
The tacking plot that would be useful is an integration of VMG data (=distance) zeroed to the expected VMG outside a manoeuvre (probably average upwind VMG over the race distance excluding tacks, or an average from immediately before the plotted manoeuvre, or even an extrapolated expected vmg). Then you'll see cumulative distance lost, including from the bear away to build speed, without worrying about plotting too much pre-tack time.
exactly! I have made a few of these plots and tool like Njord Analytics which I show does this calculation but doesn't show it graphically, only reports the cumulative distance lost
These boats are so sensitive to wind speed that the tack loss should be calculated vs straight line vmg at a given wind speed. You do a good old linear regression and use the TWS at the moment of the tack. Could go even further taking into consideration how long ago was the previous tack. Seems to take quite some time to get the last 2-3 knots of boat speed.
Great info. I think the derived plots of dirty air would be interesting to analyse by tack, as much of Ineos' success is in covering the LR tacks to hurt as much as possible. It would be interesting to seee how good and accurate at inflicting this pain they are once they are nose in front and if it becomes less effective at distance.
Lots of talk about which boat Inios should pick, less about which helm they should face. Given that the boat that leads at the first cross has a real advantage, when push comes to shove, I would rather not be facing Slingsby or Spithill. So maybe let them slug it out in the semi's and then you only have to face one of them in the final.
Excellent data and related discussion. When these boats are often travelling 20m/s it makes a 100m lead seem almost insignificant, quite unlike displacement sailing.
Looking at Day 3 of LV Semi-Finals, it is clear that both Alinghi and Patriot are comfortable in the 6.5knt Wind Range. Both Luna Rossa and Britannia looked vulnerable. Britannia and Luna Rossa can out-gauge Alinghi and Patriot in upwind medium wind conditions. Both LR and Britannia look very fast in the reaching conditions So today, the dice are cast! Heavy wind or Light wind will determine whether we go into Thursday! I believe that TNZ already knows this, so they will have looked at all the permutations of the wind cycles for October.
I thank you for , stating from the start , what you limitations on collecting Data ,,, I will say your insights to the data you present and the understanding of Ac 75 is the best on the internet, thnak you.
Yeah, this is great. It's very difficult to quantify performance differentials during a regatta; unlike say a couple of F1 cars where it's far more visible to see those strengths or weaknesses. These details expose the good stuff.
Listening to the onboard audio, it sounds to me that team NZ are constantly using hydraulics, mostly in short bursts, almost once a second. This suggests some kind of automation, and also, if they’re not running out of oil, a more efficient setup than, say, American Magic.
Thank you Mozzy for these explanations. Another question which is nagging me is "what is the strict definition of a 75 m penalty ?" ; How in practical terms is this penalty cleared by the boat ? And how does the jury evaluates compliance of the boat and declares the penalty cleared ? A 360° turn was self-explanatory, but I don't understand the 75 m rule.
The frequency graphs should all have the same area (the cumulative frequency is one), but instead they've been all stretched out of scale to match the same maximum frequecy
the areas often don't match up due to the filtering I mentioned. I.e. teams spend different amount of times straight line sailing. This can be because of dropping off the foil or doing more tacks etc.
@@MozzySailsIn that case the frequency should be relative to the total sampled time, net of tacks and jibes and any other cut, which should be very easy to implement. And if you look at the plots it is clear that they've been normalized for the same maximum frequency, which hints to a trivial conceptual mistake
I know this vid wasn’t about the boats per se but an interesting thing I saw is ineos apparent wind angle is basically identical upwind and downwind. Normally we wouldn’t expect that to be good for vmg no? To me it suggests they designed for a very narrow groove with their aero features and sail design and may be part of why they seem to struggle with high modes
I was thinking that the distribution plots were quite interesting in their multimodal appearance with clear dwelling at certain speeds. I'd love a bit more detail on why that is. I'm a definite geek, and I spent a good amount of time last night pausing and zooming in to those plots, lol. I wonder what other data they are collecting (and not sharing!)? Is there any opportunity to see inside the cockpits and understand the user interfaces? As ever, great content.
Thanks Tom. Complex, and must be a hell of a job to compile all of that into a race plan. So what? Less manoeuvres are good for Ineos even if you make your opponent tack, vmg trumps vmc in close quarters. Ineos must win the first drive on the beat like they have been doing, NZ must have some funky data in the leeway for high mode that we can either copy or combat. We still beat LR twice, that’s important. I hope we pick LR in the Semis, I can’t help but feel AM are our boogie man.
Analysis is more complicated than at first glance. Assuming tacking is no option (this is often the case due to boundary/layline constraints and match racing considerations) sailing in high mode in headers and fast mode in lifts is optimal. Having to build speed (low mode) when on a header is particularly damaging as is being forced into high mode on lift. TWS is also a pretty rough indicator as the amount of turbulence in the wind field powering the sails is going to affect the amount of power that can be extracted. Wind speed is measured at a single point and that's an approximation of the power supply. In a match race there is always going to be data bias in favour of the leading boat that gets to sail in less turbulent air and also gets to dictate the angles sailed (to some extent). What would be useful is to pit two identical boats against each other so as to quantify the performance bias that's attributable to being in the lead. This would be different in different conditions. This would quantify the handicap of a bad start - and the amount of gain needed for a pass.
Excellent. Might individual boats have different performance of port and starboard tacks and gybes? Would be interesting to separate the data potentially. I suspect our boat does have a difference, but I don't yet sail and AC75.
Mozzy, could you please get this nerdy about cant angle?? I know the data is missing to talk about the numbers, but the effects of different cant angles and how it affects the boat. I think I'm missing a detail or two that you surely know.
Very interesting: I wish I had a wind pressure map to check what a Tack is really about in the data graphs, compared to the competitor ship so that would be clear when a ship is actually gaining due to design, rather then wind whim. Unfortunately, the commentators of the races rarely offer us these data when relevant.
Fantastic. If you can calculate the average VMG for each tack, can you then determine the time required for each maneuver to reestablish that speed? Then sum that time of all maneuvers for the race to compare the two boats?
The bit about VMG beyond leyline is thought provoking, underlaying and/or getting a knock being the other end of the spectrum of error. I think the coverage of the races could bring voice to the wind a bit more thoughtfully, because the races are otherwise mysterious. Perhaps you be could be the voice-of-the-wind Mozzy?
Mozzy, thanks for your videos. They really help explain what’s going on. I would really like a video explaining the cyclors and hydraulic systems. How does it work? What is being controlled? …rudder, foil arms, foil trailing edge, jib Cunningham, main Cunningham, jib sheet, main sheet, kicker, outhaul and I’m sure there are lots of other things being adjusted. So when the cyclors are generating huge wattage, where is it going? Is it being stored somehow for use at the next manoeuvre, or is it more real time energy so they have to go for it for major mark rounding or double tacks. Some insight would be great. Thanks.
Also, I'm disappointed that the media has taken a step down for this cup. AC36 had full race replays on the back of each boat with just onboard comms. I know there is a virtual eye replay on the official AC37 cup, but If I see 1 more 'in boat' close-up when there is a crucial inter-boat change in aspect I'm going to put a fist through my screen. Missing Ken Read, hating Steven and longing for the AC36 media format. Rant over
Hey Mozzy. Great video! Any comment on the lack of data during the broadcasts? Do you know why they aren't showing it? I see so many comments about this. Maybe your famous enough at this point to get it changed for all of us sailing geeks and nerds? :)
Why is Upwind VMG bimodal for ITA? I would expect that to happen for VMC if the buoys were not square, but not for VMG, unless they are using different foils on each tack. Another theory could be an increase on the wind from one upwind leg to another.
Great explanation. In summary tho' what the data seems to show is that INEOS gains ground from each tack as it doesn't seem to require an acceleration before the tack regardless of the similar speeds between the yachts afterwards... Keep it coming Tom.
I think INEOS gains ground in tacks in the last race because they are going faster to begin with. The tacks actually cost them proportionally more than they do Luna Rossa. That's why I think a meters lost calculation would be simpler for majority to understand than this distance gain plot
@@MozzySails I agree, I just didn't explain myself well. However it is counter intuitive as most would assume increasing speed before the tack was the ideal practice - which just makes it all the more interesting!
I have a video this friday with Athena Youth and Women talking race strategy, it's kind of focused to fleet racing but in similar boats so should tell you a decent amount. If I have time I will do one for match racing
Wonder-full data analysis, thank you! You should be doing the commentary. Close ups of a foil are a waste. In the data analyses, were there any probability plots to see if the data were normally distributed?
Great explanation by the way Mozzie. At risk of sounding like my old manager, what I'm looking for is that one page rounded summation of all the data (all boats) that goes along the lines of... The data shows that at the end of the 2nd round robins, team nzl have better low wind speed vmg followed just by team LR etc... Maybe a bar chart listing all the important performance parameters and a level of goodness for each team. Does anyone know if such a thing exists??
Fantastic, thanks so much for this 😊 Those buoys making the start and each ‘end’ of a leg have puzzled me. They move don’t they when the Umpire feels the conditions merit that? My question is how are they moved? I have all sorts of theories and all seem a little outlandish 😅
They are essentially small robotic boats which hold the position they are told to. They are not anchored. They are maintaining position by active thrust against wind, waves and current
eh... saying INEOS is faster with the data commented in this video is a huge gamble. We'll see but i remember somebody mentioning a 7-0 three years ago.
Hi Mozzy, I have a question. Can an AC75 upwind VMG exceed the true wind speed? I have heard people say that these boats could sail upstream in a flowing estuary in zero wind. In my brain, that means that they would need to achieve an upwind VMG that exceeds the true wind speed. Can you help me understand if the estuary scenario is really possible? Love your videos, really insightful. Thanks.
Really it's whether Ben wants Luna Rossa first, & Patriot second, or the other way around. I think Ben continues to improve sailing against Patriot first, then Jimmy second. The other way round Ben might peak too soon & cool off before facing ETNZ.
If INEOS played their last updates in this round robin 2, the Italians and the Americans didn't it. So i will not being so comfortable to face LR in semi-final, knowing that they've still have to make that upgrades.
@@solinvictus1234 I think we might be at the end of competition where consistency & confidence are more valuable than speculative tweaks. SailGP has proved how good Slingsby is, but they (likely!) haven't got it nailed enough on Patriot to proceed past anyone but Alinghi on current form. INEOS surprised me with their ability to improve, but just how much is conditions that suited them at the right time, otherwise Luna Rossa would be continuing to make it look like their contest. ITA v UK would likely be the most interesting final, but I could see it going either way.
@@blinkybagger8342 All can win, all can lose. Regarding LR all i can say is that they still have to make updates, wispers from Barcelona instead talked of a failed update by INEOS during RR1 and a re-update in this RR2 (and certainly the hoat was doing great both in weak and string wind, it was not the case in RR1). Also LR after the issue had vs Alinghi officially Protested with the regatta judges due the malfuncion of a standard part of the arms mechanism (the one that NZ give to all) cause due the malfuncion they was almost hitting two civilian boats at the harbor and it could have been a disaster. So i don't know how much "at full" LR was vs INEOS. However semis will say us all.
@@solinvictus1234 It seems 3 years between drinks was too long and that there is more development (& entertainment) that could be squeezed out of this generation of boats. The reliability of supplied components should have been the entry criteria to the stages where it counts. If history of the the cup is anything to go by, we're due some protracted legal activity to resolve a dispute, rather than more time on the water...perhaps that is a positive impact of SailGP's existence & the cup becoming an industry. Too many people may go hungry/elsewhere if the circus isn't bringing in a crowd.
One thing I find a little hard to grasp from these is how to really see the boat performance. The VMG is one thing but how can you compare it if the TWS is so different? From that upwind data GBR had about 0.25 knots higher median TWS but less than that in VMG which seems to me like ITA are making the most of the wind. Would it possible to plot the VMG as a multiple of TWS?
General AC75 sailing q for you, Mozzy - why are the trim movements on the mainsail so dramatic? I get the overshooting to help the turn, but many of the upwind trims are far more than you see in any other boat, and I can't believe it's massive apparent wind shifts? Any ideas?
you go through gusts and lulls very quick. AWA doesn't change hugely, so entry angles don't need much trim but power available does and so the position of the leech need rapid movement, so I think that it what you are seeing with the large and quick traveller movements
As an economist I love a graph… but I’m still fabulously confused! We need to create a scatter graph of the viewer’s overall complete understanding of these videos. 😂
Thanks mate, On the TWA graphs the waist in the bulge clearly shows the modes ITA use before tacking, a build mode then tack showing up as this little dent in the bulge, My question is what is the scale and units of the width of each speed, time at that speed in seconds? so scale?
Interesting stuff, been looking at it too for some races. Just a question, shouldn’t the blob diagrams total area be closer for both boats? Since total time and distance sailed are not that different.
This is due to the filtering required to extract straight line foiling data only. Turns, none foiling time, maneuvers are taken out with varying amounts of data left for each team
![Blox Fruits Dragon Rework Update [Full Stream]](http://i.ytimg.com/vi/EqDAp8udhm0/mqdefault.jpg)
![Seungmin "그렇게, 천천히, 우리(As we are)" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/kAzmhLHePqU/mqdefault.jpg)
As an engineer. I think I can speak for all of us, we love this detail. Thankyou mozzy
Great content mate
Yes, exactly the same. We engineers absolutely love this!!
Same here, Sailor and Engineer, thank you Mozzy. And good you gave some insight on these graphs cos some of them were not clear for me when you showed them previously.
As a Wind Data engineer (for wind turbines not sailboats unfortunately hehe) and an passionate sailor, I love these videos and your explanations. You make it simple and understandable for all.
I also believe more can be extracted so I hope I will find the time in the next days to look at the data myself.
Forze Luna Rossa,
Buon Vento!
⛵
Mozzy is a wind turbine engineer too I think!
@@richardmyers6054 that's pretty cool! Curious to know the field. Discovered this channel too late, but better late than never :)
Getting super excited for the semi's!
Hi ! sorry to bother you, I´m a soon-to-be great data analayst :-) and I´ve started to extract and clean the raw data, but I have so many questions ... do you mind if I ask you if you spent some time on it ? Thanks!
Absolutely no idea what you’re talking about but I loved every bit of it once again
Thank you for that explanation...much needed for this non sailor. Really appreciate all that goes into the analysis, the commentary and the videos.
Love your work, been building excitement levels for 2 years. Appreciate analysis with data- enthusiastic TNZ fan
You are making the cup so much more interesting. They should have you on the official commentary team!
Thank you so much for doing this. I had been struggling to understand this while you had been talking the last few episodes. Really appreciate it, and great job with it!
Mozzy, stumbled upon your channel accidentally while watching race replays. I love the explaining you do and especially data!
All I've seen at this point is the thumbnail, and I've already liked the video.
Sure,.it's a Mozzy vid. It's gonna be good.
But most importantly: I couldn't agree more. Data is the key to everything. And data has been painfully absent from the cup coverage through official channels
It's deeply disappointing that we're seeing a large amount of real-time data being withheld from viewers.
We almost never see the VMGs, which are the only relevant speeds.
They don't show how much hydraulic pressure the boats have.
We're not seeing any ladder lines.
This is quite depressing, because the owners obviously think we are morons, and we can't be trusted to understand what's presented as real time data.
I find it quite insulting, and deeply frustrating.
I'm loving the racing, but I could be loving it SO MUCH MORE!
Probably a lot of it is that they don't want the other teams to know.
So I’m not the only one muttering “shut up Steven and show us the VMG” while watching 🤣
With Virtual Eye, accesible on the AC website, you can see all the matches. there you can see some data in real time. I personaly dont watch them live. I check some phases afterwards.
@@NICOLAS25478 thanks for the reminder I had forgotten that was available
Hey Mozzy, thanks! Good detail to better understand the performance and stats
Would be great to see this spread out on an excel with KPI headers, EG; Upwind speed per knot, Average speed loss per tack/gybe, average upwind angle, etc
Keep up the great work mate - Go the Kiwis!!!
hey really enjoyed this one mozzy! Was spending some time trying to decode the graphs myself and you helped me fill in some of the blanks too!
Thanks so much for sharing the links and the explanation of the various plots. The Race Viewer is a lot of fun and I wish I had more time to mess with the raw data!
Hi Mozzy, as a graduated Aeronautical Engineer, working in the Oil&gas field indeed😅 as a non-destructive testing engineer, I want to express all my respect and apprecition for your contents. I have been following you since your first AC videos. You are like a friend yo me now. Many thanks mate. Cheers
Another great video which gets the grey matter going. The part on tracking the tacks and gybes is particularly interesting. One aspect of these manauvers which is probably also important are the interial impacts in terms of potential kinetic energy wasted and also how much it is used to load and unload the foils during the crossover. This would also impact location of mass in the design and as such is going to be difficult to adjust for except through how the boat is steered and trimmed during a manoeuvre. Maybe the subject for another geeky video?. Keep up the great work, much appreciated.
Terrific video. I had reasoned out most of this from your previous discussions of the graphs, but it's good to have a definitive summary. Thank you!
I would have simply described distance gained as "the integral of VMG", but maybe that's too techy? Then the "distance lost" metric in Nord becomes the integral of steady-state VMG during the duration of the tack (i.e. what would have happened if you hadn't tacked) minus the integral of the VMG through the tack. For the comparison to be valid you have to look at the entire time that it takes to get back to steady-state VMG. Your point at 10:30 boils down to the fact that Ineos isn't bavk anywhere near steady-state VMG when the graph ends, so you don't see the additional ground they lose during the acceleration after the tack.
I feel like 20 seconds ought to be enough to be back to steady state. They're only on a tack for like a minute at a time. Honestly I think this graph is a lot more useful to the race engineers than 'distance lost' as what you're looking for here is the finer points of how to make that line as straight and steep as possible.
yeah you nailed that description
@@weatheranddarkness I think we've already seen plenty of instances where it takes long than that to get up to speed. Accelerating an AC75 seems to be a complex process, since the apparent wind and therefore required sail trim etc changes drastically as you pick up speed. Put another way there's a significnat sequence of "modes" that they have to transition through to get to steady state.
@@patrickchase5614 True, true but on the whole, in a race, it seems like the average time back to 'steady state' is well shy of 20 seconds. If you're pushing 20 seconds you're probably reaching just to stay on the foils. If you can point me to one of these graphs that paints a different picture I stand to be corrected.
@@weatheranddarkness Taking >20 sec to accelerate is exactly what the graphs that Mozzy discussed in this video at 6:40 or so are showing.
The VMG graph on the left side shows that both Ineos and Prada are still accelerating (and neither have recovered to their entry speed) when the graph cuts off at +20 sec. In other words, both AC75s consistently needed more than 20 seconds to get back up to speed in _the very race that this video uses as an example_ .
Did you actually watch the video and listen to Mozzy’s observations before you replied here?
Although the last time I sailed was some 40+ years ago and certainly are no sailing geek I thoroughly enjoy the videos; fascinating how complex and 'intertwined' most aspects of the boat are and how to get the maximum performance from it. Much appreciated....go team NZ.
Thanks for negotiating the date for us, and for explaining how it all works. This adds so much to the experience of watching the cup!
Another great ‘deep dive’.. thanks Tom, love your work
BIG TIP : The "time spent at, for example, VMG upwind" plots are better understood when you turn them through 90 degrees. Then you see a more familiar statistical distribution! And I assume the mirrored look is port/starboard?
I think that's the point of setting them vertical
@@weatheranddarkness I had just thought that it was a style of distribution to mirror it left/right. They seem perfectly symmetrical, at least as far as I can tell, so either they sail exactly the same numbers on both tacks or it's just a way to represent the data. Let me known if you see any asymmetry that I'm missing. Maybe Mozzy can comment. He's the man!
Great video. Even with the data shortage, there's still so much to dive into, but yes those can't angles would be interesting to unpack why we're seeing some of these differences.
Be great to see tacks, VMG etc port vs starboard to find the better helm!
Considering how much trimming is required to balance the boat through turns you're evaluating not just helm but also trimmer together.
Thanks Mozzy, appreciate the step through and the link to the site with data - Jimmy
Thanks Tom, while I understand most of the metrics that you talk about, I haven't got how the graphs apply, I look forward to you revisiting this again in more induvial detail. Watching you sail and describe sailing the AC40was nearly as exciting for the viewers as it was for you,.so before they pack up and go home Mozzy Sails need to sail a AC75.
You are on fire mate ! :-D these last few days a video per day !! WONDERFUL ! 😀
Awesomeness insight and next level data to compare the technical aspects rather than simply spectate the sailing decisions.
Thanks for the explanations. Have been looking at these after seeing them referenced on the Sailing Anarchy forums but hadn't ralised the second page race viewer was there -doh!
Looking forward to your analysis once you get more data, but also curious about your comments on the difference in leeway. Some races/boats were showing big differences.... Thanks Mozzy :)
Keep up the good work covering the gap between professional sailors and the rest of us👏👏👏👏
Years ago when using calculated TW values from on board instruments (Ockham) we had to use "double smoothed" values to get coherent values and trends. This was implicit in the development of "Wally". Nice to hear you mention VMC/VMG.
The tacking plot that would be useful is an integration of VMG data (=distance) zeroed to the expected VMG outside a manoeuvre (probably average upwind VMG over the race distance excluding tacks, or an average from immediately before the plotted manoeuvre, or even an extrapolated expected vmg). Then you'll see cumulative distance lost, including from the bear away to build speed, without worrying about plotting too much pre-tack time.
exactly! I have made a few of these plots and tool like Njord Analytics which I show does this calculation but doesn't show it graphically, only reports the cumulative distance lost
These boats are so sensitive to wind speed that the tack loss should be calculated vs straight line vmg at a given wind speed. You do a good old linear regression and use the TWS at the moment of the tack. Could go even further taking into consideration how long ago was the previous tack. Seems to take quite some time to get the last 2-3 knots of boat speed.
Great info. I think the derived plots of dirty air would be interesting to analyse by tack, as much of Ineos' success is in covering the LR tacks to hurt as much as possible. It would be interesting to seee how good and accurate at inflicting this pain they are once they are nose in front and if it becomes less effective at distance.
Great video. Need to get AC to give you all the data!
Perfect. Just what I needed.
Lots of talk about which boat Inios should pick, less about which helm they should face. Given that the boat that leads at the first cross has a real advantage, when push comes to shove, I would rather not be facing Slingsby or Spithill. So maybe let them slug it out in the semi's and then you only have to face one of them in the final.
Get race hard. You have to beat everyone to win the Americas Cup. Take LR they have a weakness at the moment. No real advantage to take ARBR.
Thank you for explaining the data plots
Excellent data and related discussion. When these boats are often travelling 20m/s it makes a 100m lead seem almost insignificant, quite unlike displacement sailing.
Looking at Day 3 of LV Semi-Finals, it is clear that both Alinghi and Patriot are comfortable in the 6.5knt Wind Range. Both Luna Rossa and Britannia looked vulnerable.
Britannia and Luna Rossa can out-gauge Alinghi and Patriot in upwind medium wind conditions. Both LR and Britannia look very fast in the reaching conditions
So today, the dice are cast! Heavy wind or Light wind will determine whether we go into Thursday!
I believe that TNZ already knows this, so they will have looked at all the permutations of the wind cycles for October.
Another awesome video , thanks mate :)
Awesome work , that data extraction is great to look at.
A lunchtime sandwich at work has never done so much for my sailing!
Very insightful, thanks for the explanation.
Would be great if you were one of the commentators, if not this cycle maybe the next one?
Great data! that Lego set is really cool 👌
I thank you for , stating from the start , what you limitations on collecting Data ,,, I will say your insights to the data you present and the understanding of Ac 75 is the best on the internet, thnak you.
Was curious last time where to find the "race viewer' ..many thanks for the great explanation and resources
Yeah, this is great. It's very difficult to quantify performance differentials during a regatta; unlike say a couple of F1 cars where it's far more visible to see those strengths or weaknesses. These details expose the good stuff.
Nerd fest - love it! Lots way above my pay grade but working through the examples :-) Every day is a school day. Thanks for another great video.
Listening to the onboard audio, it sounds to me that team NZ are constantly using hydraulics, mostly in short bursts, almost once a second. This suggests some kind of automation, and also, if they’re not running out of oil, a more efficient setup than, say, American Magic.
Thanks! Don’t consider my self a geek,but cool to see where all the data come from
Thx, really interesting
Epic video! Learned so much!
Thanks Mozzy, interesting stuff.
Thank you Mozzy for these explanations.
Another question which is nagging me is "what is the strict definition of a 75 m penalty ?" ; How in practical terms is this penalty cleared by the boat ? And how does the jury evaluates compliance of the boat and declares the penalty cleared ?
A 360° turn was self-explanatory, but I don't understand the 75 m rule.
What I'd really like to see is what is happening when ETNZ goes highmode and passes those boats..
The frequency graphs should all have the same area (the cumulative frequency is one), but instead they've been all stretched out of scale to match the same maximum frequecy
the areas often don't match up due to the filtering I mentioned. I.e. teams spend different amount of times straight line sailing. This can be because of dropping off the foil or doing more tacks etc.
@@MozzySailsIn that case the frequency should be relative to the total sampled time, net of tacks and jibes and any other cut, which should be very easy to implement. And if you look at the plots it is clear that they've been normalized for the same maximum frequency, which hints to a trivial conceptual mistake
Quindi la barca italiana è ancora la favorita, ed ha perso perché ha sbagliato le partenze?
@@andrealeone4784 I wouldn't say favourite. INEOS is now slightly fast, but not by a huge amount. Luna Rossa can certainly win with better tactics
Grazie mozzy, continua così! I tuoi approfondimenti sono fantastici. Aspetto sempre con trepidazione la notifica che ne annuncia uno nuovo. Grazie💯
I know this vid wasn’t about the boats per se but an interesting thing I saw is ineos apparent wind angle is basically identical upwind and downwind. Normally we wouldn’t expect that to be good for vmg no? To me it suggests they designed for a very narrow groove with their aero features and sail design and may be part of why they seem to struggle with high modes
Really helpful punditry! Thanks.
Love your work
I was thinking that the distribution plots were quite interesting in their multimodal appearance with clear dwelling at certain speeds. I'd love a bit more detail on why that is.
I'm a definite geek, and I spent a good amount of time last night pausing and zooming in to those plots, lol.
I wonder what other data they are collecting (and not sharing!)?
Is there any opportunity to see inside the cockpits and understand the user interfaces?
As ever, great content.
Well said!!
Thanks Tom. Complex, and must be a hell of a job to compile all of that into a race plan.
So what? Less manoeuvres are good for Ineos even if you make your opponent tack, vmg trumps vmc in close quarters. Ineos must win the first drive on the beat like they have been doing,
NZ must have some funky data in the leeway for high mode that we can either copy or combat.
We still beat LR twice, that’s important. I hope we pick LR in the Semis, I can’t help but feel AM are our boogie man.
Analysis is more complicated than at first glance. Assuming tacking is no option (this is often the case due to boundary/layline constraints and match racing considerations) sailing in high mode in headers and fast mode in lifts is optimal. Having to build speed (low mode) when on a header is particularly damaging as is being forced into high mode on lift.
TWS is also a pretty rough indicator as the amount of turbulence in the wind field powering the sails is going to affect the amount of power that can be extracted. Wind speed is measured at a single point and that's an approximation of the power supply.
In a match race there is always going to be data bias in favour of the leading boat that gets to sail in less turbulent air and also gets to dictate the angles sailed (to some extent).
What would be useful is to pit two identical boats against each other so as to quantify the performance bias that's attributable to being in the lead. This would be different in different conditions. This would quantify the handicap of a bad start - and the amount of gain needed for a pass.
All excellent points that I fully agree woth
Excellent. Might individual boats have different performance of port and starboard tacks and gybes? Would be interesting to separate the data potentially. I suspect our boat does have a difference, but I don't yet sail and AC75.
Mozzy, could you please get this nerdy about cant angle??
I know the data is missing to talk about the numbers, but the effects of different cant angles and how it affects the boat. I think I'm missing a detail or two that you surely know.
Very interesting: I wish I had a wind pressure map to check what a Tack is really about in the data graphs, compared to the competitor ship so that would be clear when a ship is actually gaining due to design, rather then wind whim.
Unfortunately, the commentators of the races rarely offer us these data when relevant.
I love it!! And the Lego model.
would it be possible to compare how the ac75s compare to the f50s and the ac50s?
Thank you!!!
Fantastic. If you can calculate the average VMG for each tack, can you then determine the time required for each maneuver to reestablish that speed? Then sum that time of all maneuvers for the race to compare the two boats?
The bit about VMG beyond leyline is thought provoking, underlaying and/or getting a knock being the other end of the spectrum of error. I think the coverage of the races could bring voice to the wind a bit more thoughtfully, because the races are otherwise mysterious. Perhaps you be could be the voice-of-the-wind Mozzy?
Very interesting I begun to understand offer the second watch
Does this Lego model withstand a 6m drop with minimal damage?
I think in scale it would be pretty robust! Sadly I don't have one myself to test 😞
Mozzy, thanks for your videos. They really help explain what’s going on.
I would really like a video explaining the cyclors and hydraulic systems. How does it work? What is being controlled? …rudder, foil arms, foil trailing edge, jib Cunningham, main Cunningham, jib sheet, main sheet, kicker, outhaul and I’m sure there are lots of other things being adjusted. So when the cyclors are generating huge wattage, where is it going? Is it being stored somehow for use at the next manoeuvre, or is it more real time energy so they have to go for it for major mark rounding or double tacks. Some insight would be great. Thanks.
ruclips.net/video/2LVjCzgBBpI/видео.html
@@MozzySails Thank you. I’d missed that old video. I think that covers it!
Also, I'm disappointed that the media has taken a step down for this cup. AC36 had full race replays on the back of each boat with just onboard comms. I know there is a virtual eye replay on the official AC37 cup, but If I see 1 more 'in boat' close-up when there is a crucial inter-boat change in aspect I'm going to put a fist through my screen. Missing Ken Read, hating Steven and longing for the AC36 media format. Rant over
Hey Mozzy. Great video! Any comment on the lack of data during the broadcasts? Do you know why they aren't showing it? I see so many comments about this. Maybe your famous enough at this point to get it changed for all of us sailing geeks and nerds? :)
Why is Upwind VMG bimodal for ITA? I would expect that to happen for VMC if the buoys were not square, but not for VMG, unless they are using different foils on each tack. Another theory could be an increase on the wind from one upwind leg to another.
Sea state in BCN can be very different tack to tack
Great explanation. In summary tho' what the data seems to show is that INEOS gains ground from each tack as it doesn't seem to require an acceleration before the tack regardless of the similar speeds between the yachts afterwards... Keep it coming Tom.
I think INEOS gains ground in tacks in the last race because they are going faster to begin with. The tacks actually cost them proportionally more than they do Luna Rossa. That's why I think a meters lost calculation would be simpler for majority to understand than this distance gain plot
@@MozzySails I agree, I just didn't explain myself well. However it is counter intuitive as most would assume increasing speed before the tack was the ideal practice - which just makes it all the more interesting!
Any chance of an idiots guide to understanding the strategy of the race, especially the start?
I second this as someone who got into the cup though the engineering being interesting a strategy explanation would be very useful
I have a video this friday with Athena Youth and Women talking race strategy, it's kind of focused to fleet racing but in similar boats so should tell you a decent amount. If I have time I will do one for match racing
Love to know how etnz is going
Wonder-full data analysis, thank you! You should be doing the commentary. Close ups of a foil are a waste. In the data analyses, were there any probability plots to see if the data were normally distributed?
DO we know if there is some open soure python notebook about those extractions / cleaning / analysis ?
Could the French train with kiwis before the match? After all, they got their design package from them.
Great explanation by the way Mozzie.
At risk of sounding like my old manager, what I'm looking for is that one page rounded summation of all the data (all boats) that goes along the lines of... The data shows that at the end of the 2nd round robins, team nzl have better low wind speed vmg followed just by team LR etc... Maybe a bar chart listing all the important performance parameters and a level of goodness for each team.
Does anyone know if such a thing exists??
Is there any chance these can be made into a one off design Christmas Bauble set?
Fantastic, thanks so much for this 😊 Those buoys making the start and each ‘end’ of a leg have puzzled me. They move don’t they when the Umpire feels the conditions merit that? My question is how are they moved? I have all sorts of theories and all seem a little outlandish 😅
They are essentially small robotic boats which hold the position they are told to.
They are not anchored. They are maintaining position by active thrust against wind, waves and current
eh... saying INEOS is faster with the data commented in this video is a huge gamble. We'll see but i remember somebody mentioning a 7-0 three years ago.
Hi Mozzy, I have a question. Can an AC75 upwind VMG exceed the true wind speed? I have heard people say that these boats could sail upstream in a flowing estuary in zero wind. In my brain, that means that they would need to achieve an upwind VMG that exceeds the true wind speed. Can you help me understand if the estuary scenario is really possible? Love your videos, really insightful. Thanks.
Do you think you would be any good at a live commentary during the races? I do!!
Come on Ben... Pick Jimmy and make our day
Really it's whether Ben wants Luna Rossa first, & Patriot second, or the other way around. I think Ben continues to improve sailing against Patriot first, then Jimmy second. The other way round Ben might peak too soon & cool off before facing ETNZ.
If INEOS played their last updates in this round robin 2, the Italians and the Americans didn't it. So i will not being so comfortable to face LR in semi-final, knowing that they've still have to make that upgrades.
@@solinvictus1234 I think we might be at the end of competition where consistency & confidence are more valuable than speculative tweaks. SailGP has proved how good Slingsby is, but they (likely!) haven't got it nailed enough on Patriot to proceed past anyone but Alinghi on current form. INEOS surprised me with their ability to improve, but just how much is conditions that suited them at the right time, otherwise Luna Rossa would be continuing to make it look like their contest. ITA v UK would likely be the most interesting final, but I could see it going either way.
@@blinkybagger8342 All can win, all can lose. Regarding LR all i can say is that they still have to make updates, wispers from Barcelona instead talked of a failed update by INEOS during RR1 and a re-update in this RR2 (and certainly the hoat was doing great both in weak and string wind, it was not the case in RR1). Also LR after the issue had vs Alinghi officially Protested with the regatta judges due the malfuncion of a standard part of the arms mechanism (the one that NZ give to all) cause due the malfuncion they was almost hitting two civilian boats at the harbor and it could have been a disaster.
So i don't know how much "at full" LR was vs INEOS.
However semis will say us all.
@@solinvictus1234 It seems 3 years between drinks was too long and that there is more development (& entertainment) that could be squeezed out of this generation of boats. The reliability of supplied components should have been the entry criteria to the stages where it counts. If history of the the cup is anything to go by, we're due some protracted legal activity to resolve a dispute, rather than more time on the water...perhaps that is a positive impact of SailGP's existence & the cup becoming an industry. Too many people may go hungry/elsewhere if the circus isn't bringing in a crowd.
One thing I find a little hard to grasp from these is how to really see the boat performance. The VMG is one thing but how can you compare it if the TWS is so different? From that upwind data GBR had about 0.25 knots higher median TWS but less than that in VMG which seems to me like ITA are making the most of the wind. Would it possible to plot the VMG as a multiple of TWS?
When are AC going to contract you for the series
Data ARE King
I love using this cliche because it's wrong and pisses off any true data scientist even though it's advocating for their work
@@MozzySails and then there’s the continuous maelstrom over myriad versus myriad of. 😂👏🏻👏🏻👏🏻
General AC75 sailing q for you, Mozzy - why are the trim movements on the mainsail so dramatic? I get the overshooting to help the turn, but many of the upwind trims are far more than you see in any other boat, and I can't believe it's massive apparent wind shifts? Any ideas?
you go through gusts and lulls very quick. AWA doesn't change hugely, so entry angles don't need much trim but power available does and so the position of the leech need rapid movement, so I think that it what you are seeing with the large and quick traveller movements
Is live tracking available to the public, please?
Does anyone if someone does this kind of analysis for F1. I have looked at Mark Hughes stuff and it is pretty simplistic by comparison.
Ben ...please pick LRP and make things out of the box and down the road!!
As an economist I love a graph… but I’m still fabulously confused! We need to create a scatter graph of the viewer’s overall complete understanding of these videos. 😂
Thanks mate, On the TWA graphs the waist in the bulge clearly shows the modes ITA use before tacking, a build mode then tack showing up as this little dent in the bulge, My question is what is the scale and units of the width of each speed, time at that speed in seconds? so scale?
I think the plots are actually a 'probability' rather than time, but you can think of it as time, so width of bulge means more time.
@@MozzySails looks to be how often it was at that speed for say a 5 second sample time. Or whatever sample time the records use.
@@vhg130 sample rate is 1hz
thx
Interesting stuff, been looking at it too for some races. Just a question, shouldn’t the blob diagrams total area be closer for both boats? Since total time and distance sailed are not that different.
This is due to the filtering required to extract straight line foiling data only. Turns, none foiling time, maneuvers are taken out with varying amounts of data left for each team
@@MozzySails Ah clear, that makes total sense.
great job...even the *.ac37 files are actually unreadable😣
BOOM!