Cats, Weight, Performance and Value

[Intentional Drifter]

Although I have posted some of my thoughts on this issue in the 420 thread, the comments by Dave got me thinking. (CatManDo - by the way, Dave, I love the gif, but I don’t believe you really feel that way.). Of course, one of the big problems with understanding the effects of “weight” on “performance” is the nebulousness of the term “performance”, in that it typically means different things to different people, depending on their expectations for the boat. However, most of us generally include within those expectations the time it takes to get from point A to point B. So, let’s take that as at least one common denominator. Unfortunately, even defining “weight” is complicated by the differing measurement methods used by the manufacturers and how different people outfit their boat. “Light” displacement, “weight when loaded to the waterline”, “maximum” displacement, “full” displacement (and there are probably a few others) are cited by someone or another.

From my perspective, I really don’t care how much a boat weighs, as a unitary variable. I do care, very much, how it performs in getting me from point A to point B. Many choose to examine these using a variety of theoretically related variables and, from them, speculate about the ultimate performance. That is probably a worthwhile exercise, especially for designers and builders. But, for a consumer like me, I am most interested in how a boat performs in the “real world”.

Fortunately, there are some data available. Every year at about the same time, the Atlantic Rally for Cruisers (ARC) leaves from the same place and arrives at the same place. The elapsed time from start to finish is logged. They have a separate multihull class and the data are posted to their website. It is a pretty popular rally and, given the quasi-competitive nature, we can expect that most of the boat crews are doing their best to log fast times. Also, all of the boats go through an inspection they must pass before departure and carry a minimum set of equipment.

I took the data from the ARC for the last three years (’04 - ’06). These are particularly “good” years in that they represent a variety of conditions. 2004 was a particularly slow passage year (mean crossing time = 23379 minutes) with all of the boats encountering light winds. 2005 was a fast year (mean crossing time = 14342 minutes) and 2006 was close to average (mean crossing time = 18315 minutes). So, using these three years together should give us some decent overall estimates of performance in a variety of conditions. (Note: all data are multihulls only and DNF boats are not included.)

A total of 57 multihulls competed and finished during the three years (20 in 2004, 17 in 2005, and 20 in 2006). 20 boats competed during two years, representing a variety of brands and models, ranging from an Island Spirit 400 to a Privilege 585 and including a Dragonfly 1200, several Lagoons and Catanas of different models, an Outremer, a Kelsall Space 52, and a St Francis 50. (Notably absent are Australian cats, probably due to the area. That’s too bad, since if they are superior to the predominantly Euro cats, then we can’t determine that from these data.)

For brands and models that had more than one boat represented in a given year, I analyzed the data in two ways, both separately (in other words, keeping each boat listed by itself) and aggregately (averaging the crossing times for all of the same boats and using only the average crossing time). So, if there were three Lagoon 410’s in a given year, with the aggregate method I averaged the crossing times for the three boats and used that result to represent the Lagoon 410 for that year. As it turned out, the method used did not matter in the results, so the results presented below will use the aggregate method since that compensates for one boat having a particularly skilled captain or favorable winds, or vice-versa.)

Data regarding the elapsed time (uncorrected - remember this is “real world” and I don’t care if they had to motor some to get through a lull, plus, this keeps the correction formulas out of the mix) and boat brand/model were obtained from the official ARC websites for the respective years. Other data about the boats (length at the waterline and displacement) were obtained from the manufacturer’s official data.

A note about the displacement figures quickly became apparent: there is a wide variety of number cited by the manufacturers and finding a single figure that can be compared from one to the other is difficult. The most common figure cited is “light” displacement. However, what each means by this number is not necessarily specified and may vary. So, although this was the only common figure and thus was the one used in the analyses, it is still prudent to treat it with caution.

One additional data point was also collected: current asking price for the boat/model. This was obtained by searching the Yachtworld listings and taking the average of the asking prices boats of the same model that were no more than four years old. This was rounded off to tenths of a million in US dollars.

Some may already be asking, why no data for sail area? Two reasons: (1) Working sail area (believe it or not) is not always listed by the manufacturers; and (2) Given the wide variety and sizes of downwind sails, as well as the amount of time in which they have may have been used, this was viewed as being too unreliable a variable for inclusion. Again, since I am assuming that each crew was approaching the rally with at least a bit of competitive attitude, tempered by the fact that they were doing a transatlantic run, I decided to assume that each crew used their best judgment as to the sails they flew and when they flew them.

Data were analyzed using SSPS (a statistical analysis program with which I have over 25 years of experience). On to the results.

Starting with most obvious hypothesis that lighter displacement boats will have shorter crossing times, I ran a Pearson correlation coefficient. It was: -.08, which is not significant. As a matter of fact, this is a very small number and basically means that there was no relationship between crossing time and displacement. (For the more statistically focused among us, the significance test, 2-tailed, was .59. Correlation coefficients run from -1 to +1, with zero meaning “no relationship”.)

Next, perhaps the crossing time is more related to waterline length. This correlation was also not significant at -.24 (sig, 2-tailed, .12). However, it is at least getting closer to it and in the expected direction (longer boats having slightly, but insignificantly shorter crossing times, thus the minus sign).

I then ran a new variable in which I took the length at the waterline and divided it by the displacement to result in a variable called “Pounds of displacement per foot at the waterline”, since this would better represent not just displacement or length, but how much “boat per foot” they were trying to push through the water. The correlation of this variable with the crossing time was also not significant (Pearson r = -.004, sig., 2-tailed, .98)

Now, does this mean that weight doesn’t matter? No, at least not to me. But, what it does mean to me is that after the designers and builders finish doing their jobs, they design and build a “total” boat. In that context, weight is compensated in such a way that it no longer, by itself, can be used to predict the overall performance of the boat -- even when adjusted for the length at the waterline, at least in “real world” conditions.

However, I still wasn’t satisfied with just correlations. So, I divided the boats up by categories so I could directly compare Light, Average, and Heavy boats, using both the raw displacement numbers and the “pounds per foot” number. I considered “light” boats to be those weighing less than 9 tons, “heavy” boats to be those weighing 14 or more tons, and “average” boats being those that weighed in between. (There were 17 light boats, 17 average boats, and 10 heavy boats.) I then ran an Analysis of Variance (ANOVA) between the three groups, which again was not significant (F = .23, sig. = .79)

However, just using “light”, “heavy” and “average” can be misleading, since smaller boats (like a Lagoon 380 and a Fountaine Pajot 38) that have a low gross weight could also be included with a boat like an Outremer 45. So, the boats were again re-categorized using the “pounds per foot at the waterline” variable. This resulted in 8 boats being categorized as “light”, 10 being categorized as “heavy” and 26 being considered as “average”. Example boats from each category include:

Light

Kelsall Space 52
Outremer 45
Dragonfly 1200
Nautitech 40
Lagoon 410

Average

Lagoon 55
Catana 43
Lagoon 440
St. Francis 50
Catana 471

Heavy

Catana 582
Lagoon 500
Privilege 585
Lagoon 570
Catana 522

When we look at the mean crossing times for each category of the boats (based on “pounds per foot of waterline length”, here are the actual average crossing times (in minutes) for each category:

Light 18002
Average 19474
Heavy 17986

Total 18868 (the average crossing time for all of the boats)

Although it probably doesn’t take an ANOVA to tell you this, here it is, anyway: There were no significant differences in crossing times. (F = .54, Sig. = .59)

Since I also had the pricing data, I was curious about another matter: what is the relationship between performance, comfort, and price? Is there a way to quantify this? After all, there are a number of these boats that are very nice, and very expensive (prices ranged from $250K to $1.6M, with an average price of $760K). Granted, most of the less expensive boats were shorter and/or older models no longer produced (including a Prout 39). So, I computed a Performance/Price Index ( Crossing Time divided by Price, which was further divided by 1000 in order to get the result into smaller numbers). This was further refined by multiplying this by Length at the Waterline as a measure of “comfort” and divided by 100, again to obtain results with smaller numbers. Having done so, here are some of the results. In this type of index, the numbers themselves do not have an absolute meaning, they are only relative and useful for comparison between the boats. Smaller numbers indicate a “higher price” for the performance and comfort, whereas larger numbers could be interpreted as indicating greater “value” for the performance and comfort:

Note: The average Performance/Price/Comfort Index for all boats is 13.8 (SD = 6.6)

Catana 43 13.2
Catana 471 11.8
Catana 472 7.5
Catana 521 5.2
Catana 582 8.1
Dragonfly 1200 16.4
F - P 38 26.0
Island Spirit 400 10.4
Lagoon 380 23.2
Lagoon 410 24.9
Lagoon 440 16.1
Lagoon 500 14.3
Lagoon 55 17.8
Lagoon 570 4.8
Nautitech 40 15.9
Outremer 45 16.4
Privilege 435 17.5
Privilege 495 16.9
Privilege 585 7.5
Prout 39 20.2
Space 52 31.7
St Francis 50 10.4

Boat that rate as better than average in the Performance/Price/Comfort Index include some that may be surprises: The Fountaine-Pajot 38, the Lagoon 380 and 410, the Prout 39 and the Kelsall Space 52. Boats that fall lower than average include some pretty luxurious and expensive boats: the larger Catanas, the Lagoon 570 and the Privilege 585.

Perhaps what these data are saying is that if you equate performance as meaning, primarily, the ability to get from point A to point B and comfort as equating to waterline length, leaving luxury accommodations out of the equation, there are some clearly better values in meeting that expectation. However, if you place a higher value on the luxuries, you will pay much more for it, but don’t expect that to also pay an additional dividend in performance. From the ARC data, it won’t.

You can certainly criticize this analysis and the basis for the criticisms has been made plain, above. It represents a largely European group of boats. It is a transatlantic crossing that is usually (though not always) a downwind run. You can argue that the lack of sa/d numbers is a problem, and, if we could somehow keep that constant across a one to two week period of sailing, I would agree with you. You can argue that there are other, better variables to use. Please do, and while you’re at it, give us the data, too.

As I’ve said, before, I am not a boat designer. Neither have I ever built a boat. But, I’m not stupid, either. I know there are sales pitches-a-plenty out there and more than enough “opinions”. This is an attempt on my part to bring some sense of logic to the discussion that is based on actual data from real world experience and not from a salesman’s promises or some preconceived notions acquired somewhere in the distant past. If you’ve got other data, then please, share it. Not opinions or conclusions, please, but real, able to be reproduced, checked, and analyzed data.

I hope some of you find this interesting and helpful.

ID

“I love the smell of sacred cow in the morning -- it smells like, Bar-B-Q!”

[Intentional Drifter]

P.S. I will gladly share the dataset with anyone who wants. Just send me a PM or email. You will need SPSS, though.

ID

[Evan]

Interesting analysis, but I think your data is too sparse to draw meaningful conclusions. Some crews treat ARC as an all out race and blow out several chutes on passage; others try to minimize discomfort and damage to the boat. These factors will greatly affect passage times. You really need a RACE where all skippers are trying their hardest to win.

Or try "multihull demo days" after the Annapolis show. The fast boats try to sail close to the other boats to show how fast they are; the slow ones sail far away so they don't seem slow. I sailed on four 40' or so cruising cats and it was very obvious which were the slow boats in light winds.

As a naval architect, nothing is more amusing that published displacement figures of yacht builders. ALL of them should be treated as polite fiction. Most mono builders historically treat displacement as "1/2 load condition" - fuel and water tanks 1/2 full, some crew, and gear, food etc for maybe a week. Most multi builders are only giving light ship weights. The EC regulations may change that (or may not).

One example - look at the PDQ 36. It's published displacement never increased, even after many ammenities were added during it's design life.

[Intentional Drifter]

Thank you for your comments. From a statistical perspective, though, a sample of 57 is actually not bad at all. Certainly enough to get sufficient variance such that the statistics used can be done without violating assumptions. Would more be better? Of course. But, where do those data exist? For the purposes of this analysis, which is to look at passagemaking performance, more important than the raw sample size are the conditions under which they were collected. A long-term passage where the actual time used in completing it is actually an excellent situation. It will include a variety of weather conditions, sea states, crew & captain experience, etc.

Short (less than 3 days) race conditions are exactly what we do not want for assessing cruising/passagemaking performance. That is because there is too little variability in the conditions. This is even more of a factor in something like the multihull demo days -- too little variability in conditions, too much opportunity for one or more boats to try and "help" themselves by carrying very little load. For a transatlantic passage, all of the boats are going to have full tanks, lots of provisions, spares, etc. Those are much more like actual cruising conditions, because they are.

Regarding the issue of some crews treating the ARC as an all out race and blowing several chutes in the process vs others who are trying to minimize discomfort, I have no doubt that you are correct. However, I would also expect that those crews who do treat it as an all out race and are willing to threaten the loss of thousands of dollars in chutes would also be ones that have more money, which would also imply more expensive, and bigger, boats. What these data show, though, is that this is not resulting in significantly better passage times. While it may result in one or two "beating" the rest, when you look at what can be reasonably expected as a group - it makes no difference.

Regarding the published displacement figures, I believe you are correct, and say so. However, I compensated for this in one analysis by not using the absolute numbers, but rather the general category of light, heavy and average. When you look at the example boats in such categories, I think most people would agree that, relative to each other, these would be light, heavy and average weight boats. Even then, though, the crossing time differences were not significant.

So, where does this leave us? At this time, there has been an assumption voiced that weight, as a unitary variable, matters. But, I haven't seen data that supports this assumption. When looking at the data I could find, collected in these real world conditions, it is not possible to reject the null hypothesis -- that being that weight, by itself, makes no difference. In this case, not even weight as a function of pounds per foot of waterline length, resulted in significantly faster crossing times.

Logically, it is impossible to prove that something does not exist, we can only prove that it does. If you say that it matters, then it is the responsibility of the person saying so to show it in a way that is testable and can be disproven. At the point at which it is tested and not disproven, then you can say that the null hypothesis is rejected and the proposition is then taken to be correct. Regarding this particular issue, we are not at that point.

Got better, different, more comprehensive data? Do share it, please. I'd love to see it.

ID

[ess105]

ID - thanks for raising this discussion to one based on available facts rather than unsupported oppinion.

[GordMay]

I.D.
I frequently hear the phrase that one “cannot prove or disprove a negative” – but would suggest that, logically, NOTHING EXISTS, until we have contrary proof that it does . Belief without evidence is faith – not knowledge.

Ie: A “null hypothesis”:
I have title to, and the right to sell the Brooklyn Bridge?
Wanna buy it?

Note: The study & practice of logic hasn’t yet been satisfactorily summarized in a single sentence – not mine & not yours . (would that I could)

[Factor]

I am reminded of Disralies famous saying regarding there being the sorts of lies.

[mikereed100]

ID

Another interesting feature that is tucked into your data is that while the average length of the light boats was about 44' and that of the heavy boats 55', their passage times were almost identical.

[ess105]

The phrase "There's Lies, Damn Lies and then there's Statistics" refers to how statistics can be used to deliberately mislead. The phrase in no way implies statistical analysis is bogus.

So Factor - what is it you're trying to say? If there is a flaw in IDs reasoning, state it. Let's have a proper discussion.

[Intentional Drifter]

Actually, Gord, much of what we call "science" is the process of disproving negatives. It is proving negatives that is the problem. You are correct that belief without evidence is faith. I, personally, don't have a problem with that, but let's just keep straight which it is that we are espousing. It has been said, and supposedly demonstrated, that with enough faith, we don't need a boat. I don't have that much faith, so I'll base my decisions on the physicists and engineers. They can demonstrate their findings, and reproduce them for anyone to see.

RE: Disraeli's famous quote: "lies, damn lies, and statistics". A pithy statement voiced by one of the foremost politicians and rhetoriticians of his day. Yes, a competent and unethical person can manipulate data and statistics to infer many things. That is one of the reasons why we have peer review processes in science. Are you inferring that I have been dishonest and manipulative in my analysis? If so, or even if you're just curious, you are welcome to ask for the dataset, the statistical runs, and see for yourself. No secrets. If you think I've fudged the data, miss-applied the stats, or overlooked something important, then feel free to correct me. It would be a service to all of us.

The saying, often attributed to Disraeli (or Twain, although the parentage of it goes much further back than both of them) is intended to tell people to be dubious of statistical findings vs "common sense". The reality is that people do attend to statistical results, all the time in our daily lives, without even thinking about it. I imagine that you do not routinely run red lights, ignore gale warnings, and not change the oil. After all, these things are all based on statistics.

ID

[Intentional Drifter]

MikeReed100 --

Yes, you are correct. While the longer boats were slightly quicker across the pond, it was not enough to reach significance. Now, what this means, statistically, is that the spread of results between the longer and shorter boats was not far enough apart -- even though the average times were "different" -- to say that the difference was not due to chance. So, enough shorter boats were the same as, or quicker than, enough longer boats, so as to render the overall finding unreliable.

However, the finding is also close enough that if I were doing an experimental study, I would include it again -- there may be something to it, but these data cannot confirm that it is there.

However, from a practical standpoint, one must then determine, for oneself, if the actual difference, if it is there, is worth the cost needed to obtain it. That is a personal decision.

ID

[Joli]

Well ID, you have uncovered the great truth. Passage times for similar size boats regardless of the number of hulls will be similar. Hard to believe huh? Considering all the hype about the speed of multi's The reality is a cruisng boat is not a racing boat.

You are also struggling with weight comparisons for light medium and heavy but can't find any meaningful results that indicate the lighter boat will be quicker on a passage. The reason the lighter boat is no quicker then the heavy boat is the SA/D of and empty light boat is much different when loaded for cruising.

I have used this example before: A Conser 40 has a SA/D of 35 or so empty and weighs 10k. Well add 5k for everything and you have a 50% increase in displacement. The SA/D falls through the basement. Take our mono. It weighs 65k with a SA/D of 22. Add that same 5k for cruising gear and our SA/D barely budges.

You did the SA and found this to be true. We raced for years and found this to be true. In the end, a cruisers speed will be dictated by waterline.

We don't Cruise VOR 70's or G Cats. We cruise boats that are more heavily loaded and more conservatively sailed. According to our polars we can make the jump from the Ches. Bay to Sopers Hole in a little over 5 days. I'll tell you right now, we will never do that. We will never push the boat that hard for two reasons, safety and comfort. 12-13 knots is fine when day sailing but it is tiring after a while.

Anyway, it doesn't matter. The trip is part of the adventure and if you make 150 miles a day or 175 miles a day, who cares.

[Factor]

Originally Posted by ess105

The phrase "There's Lies, Damn Lies and then there's Statistics" refers to how statistics can be used to deliberately mislead. The phrase in no way implies statistical analysis is bogus.

So Factor - what is it you're trying to say? If there is a flaw in IDs reasoning, state it. Let's have a proper discussion.

The flaw in the reasoning stems from the assumptions. The analysis assumes that most people were maximising their boat speed, within conservative limits. We dont know this to be the case, we dont knbow for example whether some boats were deliberately slowing up to "stay with the pack" we dont know whether some boats were just doing the pond jump or they were ready for a circumnavigation. We dont know what crossing straegies were employed.

Gee you blokes are getting defemsive on this - all anyone said was that the lagoon 420 is a big heavy comfortable bus, it aint fast but it obviously fits the bill for some people. As for JOLIs suggestion that this proves that multis arent faster than monos, well there is a leap in logic.

And ID what about correcting the draft figures for the seawind 1160 in the related post.

Thats the other problem with this analysis, the underlying data is really suspect. The displacements etc aren't actually known.

[Intentional Drifter]

Hey, Joli, a "great truth"? Ha. A few hours of idle entertainment when it is too cold to go outside. I somehow don't think that Nature would be interested!

Besides, as my moniker says, I'd rather go slower and enjoy. The whole mono vs multi thing to me is silly. I like 'em both, just for our needs, a multi makes more sense.

ID

[Intentional Drifter]

Factor --

RE: The Seawind, for one thing, I didn't post the information for the Seawind in the first place, but for what it's worth, Seawind says it is 3'2". I have no reason not to believe them.

RE: Crossing strategies, you are correct, we don't know if there were boats that were trying to slow themselves or if others were trying to beat the rest. One of the advantages of using larger samples (and, although I'm repeating myself, I haven't seen a larger sample and if you have one, please share it) is that such individual characteristics are compensated for by the decisions of others in the group. That's one of the big advantages of using inferential statistics.

RE: defensiveness about the 420, I beg to differ. The reality is that we don't have any data about how it performs in such a situation. I haven't seen anyone try to hide this or say otherwise. It is simply unknown. Are we hopeful? Of course. My entire endeavor was intended to discover if there was any factual basis in statements, like yours ("it is big heavy comfortable bus, it ain't fast...") to expect that the 420 would be slower doing passages than other multis. I haven't seen such evidence, yet. Do you have some?

RE: Displacement numbers being suspect. I agree and said so; "although this was the only common figure and thus was the one used in the analyses, it is still prudent to treat it with caution." They are what they are and came from where they did. Do you have anything better? If so, I'd be happy to see them and run it again.

ID