The Offshore Voyaging Reference Site

Anchor Tests—The Good, The Bad, and The Downright Silly


I’m frequently asked what my opinion is of a given anchor, often of a new pattern.

And over the years I have developed a pat answer:

Sorry, I don’t opine on anchors unless I, or people I know well and trust, have used the anchor several hundred times in many different places over a period of several years.

At this point, almost invariably, the questioner adds that his or her interest was provoked by the anchor in question’s great performance in a test.

That would seem logical…except for the fact that most, maybe all, anchor tests are flawed and therefore should not be used, at least not in isolation, to determine what anchor we voyagers should use to keep our boats and loved ones safe.

Login to continue reading (scroll down)

Inline Feedbacks
View all comments

My professional career started with the test and evaluation of some very expensive stuff, satellites and the rockets to launch them. Test cost often reached 50% of the developmental cost. Unfortunately, some of the test ethic we applied in those days is no longer affordable. We designed tests along two axes — test to destruction and test to develop the perimeters of a performance envelope.

Most, OK all, of the anchor test results I have seen in 41 years of sailing have conflated the two testing regimes and have resulted in three fallacies, intended or not. Fallacy 1 = test conditions accurately reflect operational use. Fallacy 2 = enough testing has been done to produce statistically significant results. Fallacy 3 = test designs capture environmental variability. These fallacies are either derivative of tester bias or external ($$$) constraints on test design, execution, and interpretation.

Your case about soft mud is a good example. What is soft mud? What is the statistical distribution of particle sizes, what is the ratio of water to solids, are there lubricious deposits in the mud, etc, etc… We scooped up two equal measures of mud from a dive site several years ago (in preparation for hiding a weapon for recovery divers to search for). The first can weighed 20% more than the second can. The samples had been taken 2 feet apart. Overall, we found a 30% variability across the 1/2 hectare site. The site survey chart was labelled Sft Md over about 3 hectare.

We have come to rely on our SPADE a great deal more than any other anchor we have had, but we still won’t trust it during transitional events. We still set anchor watches and use anchor drag tools.


Thanks, John,
Anecdotally, after one season of turtle grass in the Abacos we knew our anchor was destined for replacement on our return to the Chesapeake. Having anchored off Galesville, MD, we were later warned of the derecho heading for the DC area and had just enough time to run out more chain and add the second snubber. Winds were predicted to be 70 mph.

The first gust that hit us, registered 89 mph and nearby they saw 100 mph. The 70+ wind lasted for 15 minutes. We expected to be on the beach before it was all over, but amazingly the anchor, so irritatingly untrustworthy in the Bahamas, held. But Janet commented she didn’t think we were in the same spot. The GPS track showed our (power set) anchor on 175 feet of chain in 8 feet of water over hard mud had dragged 245 feet before stopping. We reset the alarm and racked out in total darkness.

The following morning on retrieval, we discovered that the anchor held because it was jammed in the throat of a knock-off seaplane anchor attached to a derelict boat that had sunk near our favorite spot during our absence.

We bought the SPADE several days later when the power finally came back on in Annapolis.


Great post, we all need to be reminded that modelling and testing are only ever an approximation or simplification of what will happen in the big, wild world. On another point the inuit have recently been found to have a specific genetic modification that allows them to tolerate their high fat diet with impunity- the rest of us may find it less healthy. The French appear to eat less than their anglo saxon cousins. (avoiding transfats is probably wise, though)Again , more research required and we’ll probably find that it’s a little more complicated than we first thought.


Concur with your thoughts on spade.
Using S120 on Malo 39 for last 5 years in UK and Medditeraen .
Previous CQR and Bruce just don’t compare.
Any suggestions foR a lovely 25kg Stainless steel Bruce sitting at home in Garage (UK)
Fair winds

Ed finn

When one reviews all the published research literature and testing on the new generation anchors versus the more traditional styles, (epidemiology?)
and take all the data into perspective, then the summation of all the test data agrees with the real world experience. The new style -concave- anchors are much much better then the old styles.
So, practical experience agrees with the science and testing. Whoopee!

Regarding that stainless steel Bruce anchor you have at home, I suggest it would make a nice lawn ornament.
In a few years it will be as outdated as those huge cast iron fisherman and Navy anchors that institutions and municipal governments love to use for ornaments.
Put that SS Bruce on the lawn I say, and get out in front of the curve!

Marc Dacey

I find anchor tests useful in the same way courtroom art is useful. They capture a moment in time, but they aren’t really evidence. That said, they do represent a start point in that they tend to weed out the less-useful anchor designs and *can* (but sometimes do not) point at anchor styles of good utility that aren’t always top of mind. Five years ago, I might have bought a Delta. Now I am, due not only to this site, but to testing I’ve seen, leaning toward the Spade and I like what I’ve seen of the Sarca Excel as well, save that it’s very difficult to obtain in Canada, whereas the Spade is not.

The fact that I know very good sailors who still swear by their Bruces and CQRs suggests to me that even a currently sub-par anchor can still hold given good technique and a proper watch. In that spirit, I do wonder if more effective anchors merely allow worse technique to be survivable, much as airbags keep bad or careless drivers alive to drive again another day.

Brian Sheehan

Obviously, anchor testing is a contentious topic and yet we will always learn something from it….if nothing else, the tester’s analysis and reaction to the results.

Case in point: A Finnish boating magazine recently conducted a pull test on a brick hard gravel beach, and then on a softer beach using a four wheel drive 4 x 4 truck.

Unsurprisingly, our Fortress anchor, with its too large flukes had trouble penetrating the hard gravel beach, however in the softer soil it buried so deeply that afterwards they had to dig it out with shovels.

The conclusion they drew was that the Fortress was a risk to use in a softer (and more common) soil for anchoring, since in their words, “few people want to dive with the shovel in their hand prior to departure!”

That’s the first time in 18 years of observing anchor tests that I saw a model get scolded for performing too well…..

Brian Sheehan


I handle all customer comments and inquires for our company, and by far the most frequent complaint I receive is retrieval after a heavy blow. As one might imagine, breaking free an anchor with two large flukes and a massive surface area can be very challenging once the sea bottom is tightly compacted and compressed against those two flukes.

Earlier this year, a sailor in the Caribbean wondered after a similar situation whether the Fortress was a “sacrificial anchor,” as he ended up cutting his line. Although loss is not covered by warranty, I sent him a new anchor.

Additionally, a short while ago I received a call from an 87-ft USCG patrol boat in the Gulf of Mexico, and they wanted to buy a new 47-lb FX-85, their primary anchor aboard this vessel. The “Coastie” told me that the seas were too rough for their patrols, and so they hung on this anchor for three days, and when it came time to retrieve the anchor, they could not break it free, and so they cut the line.

During the Chesapeake Bay soft mud testing, we lost a 21-lb FX-37 after the wire rope snapped at 3,500 lbs. We were directly above the anchor and we were trying to slowly break it free, and the winch operator estimated that it was buried 13 feet into the mud. The 81-ft research vessel shifted slightly from the waves of a passing boat, and the wire rope slipped between the horizontal and vertical rollers on the custom fairlead, and so there was metal on metal and the rope broke, sending everyone scattering on the aft deck.

I think the key here is PATIENCE, and a similar recovery method, as you described, is used where you stay in a position directly above the anchor with a tight line, and then let the movement of the boat slowly break the anchor free.

On a separate note, we drill a hole into the Crown (center piece) of the anchor, to which you can install a breakaway (trip) line or better still, a secondary line with a buoy that will bob along the surface, although few boaters want to take the time and trouble to install one and retrieve their anchor this way.

Marc Dacey

Brian’s point about triplines to the crown (or the equivalent circa 90 degrees from the fluke angle) attachment point is well made. Unless I have the sort of foresight to make a waypoint where I drop the hook, why would we not buoy an expensive anchor of any type? For one thing, it shows fairly closely (if you’ve make the trip line length close to the depth) where one’s hook is, which is going to suggest visually where your chain is likely to be laying, thus alerting other boats anchoring where to avoid. The “easier retrieval” aspect is just gravy. Even fifty feet of Spectra or Amsteel (a good choice for this job, I would think) is a fraction of the cost of a new anchor.

Eric Klem

Hi Marc,

In general, I am not a fan of trip lines and try to avoid them if at all possible. I have seen many problems with them including a boat drifting over the anchor in a windshift, catching the trip line on the prop making the crew unwilling to motor and pulling the anchor out in the process. In something like 2000 anchor sets, I don’t remember ever having to abandon an anchor and I have never needed the trip line the few times it has been rigged. That doesn’t mean that I haven’t had issues with retrieving anchors but it has always been something that I can sort out sooner or later. Anchors that set quickly are much better as they cover less ground decreasing your chance of hooking a wire, chain, log, etc. If you do use an anchor buoy, make sure to put a “fuse” in the line as it could be a real mess if someone wrapped your amsteel up in their prop. I might go with a line that sinks so that you can have a fuse 5′ under the surface so that you can reach the line below the fuse by hauling up but other boats props will never come close. If the concern is a deeply buried anchor and not obstructions on the bottom, you might consider running the trip line along your chain with very light lashings that you could break if you needed to.


Jim Trefethen

Another winner article John.

I had an opportunity to do some practical anchor testing years ago when I was caught by Hurricane Bob in Lewis Cove in Linekin Bay aboard a 40-foot powerboat named Duchess. The only three anchors I had were a 55-pound CQR, a 30-pound Danforth HT, and a 30 pound Luke fisherman. I set these in a diamond pattern with as much rode as I had then joined my family in an evacuation shelter for the duration. When the weather cleared the next day, Duchess had moved about fifty yards closer to shore, she had clocked the compass about twenty times completely snarling the three anchor rodes (two rope and one chain) to the extent they were unsalvageable. Two of the three anchors had pulled free and where incorporated in the snarl, and the third was all that held the boat.

The question, 0f course, is which of the three anchors was the hero that saved the day (and my boat)? Given that the bottom off Linekin Bay (just north of Boothbay Harbor) is an amalgam of rocks and sand and kelp, the answer is obviously the fisherman. The other two are useless under those bottom conditions even in moderate weather.

Does this prove that the Luke Fisherman is the best anchor to use in a hurricane? Of course not. It proves that some of us are lucky buggers who get more breaks than we deserve.

Jim Trefethen

Eric Klem

Hi John,

Well put. Like you, I am an avid reader of anchor testing but I understand the limitations of the data. I would love to see someone figure out a good way to test setting and veering capabilities of anchors in really tough bottoms but this would likely be way too expensive and people would invariably complain about the setting method.

To Chris’ point about using a single test to try to fully characterize an anchor, I think that the variability in the test results is staggering and shows that these factors have not been separated out. Looking at the raw data from the Sail test a few years ago, the standard deviations were often greater than 50% of the average holding power which means on some sets you would drag in a gale and with others you could be fine in a hurricane all in the same area.

Unfortunately, real world test results can be problematic as well. Many people are not familiar with the relationship between windspeed, waveheight and the forces involved. I would consider this a windy year for us and out of 70+ nights, we have had 1 night of storm force conditions, 3 nights of gale force and a few thunderstorms/fronts where it blew hard for a few minutes. While we have been extremely impressed with our anchor, most of the impression is based on power setting and occasional weather events. Additionally, the bottoms around the world are very different so it is easy to believe that an anchor does very well only to find that it doesn’t work in a different area. Therefore it takes a large number of people who anchor a lot reporting on their experiences to figure out the smaller difference.


Drew Frye

I recently finished a large anchor testing project.
a. Consistent bottoms scale-up, layered bottoms don’t.
b. It’s bloody hard work.
c. Getting good data is really difficult. When testers show ALL of the runs, the variability is often +/- 50%. Of course, some anchors are more consistent than others, and we like those!
d. When in doubt, go for a swim. There is something going on down there you don’t understand. Particularly true with 2-anchor rigs; lots of false information out there.

Brian Sheehan


Nice piece of work with your anchor test story in the December 2015 issue of Practical Sailor.

Well done,

Drew Frye

a. That was the dingy anchor review, which I only did because I had to develop baseline data for the Claw, Guardian, and Mantus (representatives for 3 design schools). There will be several more, all focused on 2-anchor systems,
b. All of the shift/veer data was cut for size. Basically, all of them reset well except for the Guardian. In that case, it did track well with changes that were gradual (shifts in a rising wind), just not reversals.

Actually, anchors scale surprisingly well in consistent, non-layered bottoms, so that is all that I worked with using small anchors. Pretty accurate over a range of 1000 to one, actually, and the US Navy uses 100:1 factors in initial fleet mooring design. But in the real world, anchor dragging is what happens when there is something bad about the bottom.

The whole purpose of the test program was not to test anchors per se (very tough, and I’m not going to beat that dead horse), but rather to look at trends when different 2-anchor rigging methods are used, since a lot of people have trouble with this. I could drag 2-pound anchors, but there is simply no way any reasonable equipment is going to budge two 35-pound anchors well-set in sand! The small anchors allowed for some easy myth busting (tandem anchors are nearly always worse than either anchor alone) and pointed in some interesting directions (open-hawse with 2 different anchors works best if the hawse is asymmetrical).

Then I went full scale. What I learned small transferred, plus some new lessons were learned. My motivation was the low holding values reported in the Solomons Island Tests (yes, folks with good anchors drag there, because the mud is barely soup), since that is my home turf. In fact, their data for that location squared well with my experiences, both before and during testing.

There will also be an article on snubbers, because I had to develop anchor rode tension data to translate conclusions to reality.

My 3 most vivid experiences, since each is important:
* I anchored my 34-foot cat on two 2-pound anchors in 20-30 knots in the open one night (I had a big anchor deployed with a lot of slack). My way of storm testing a 2-anchor rig. They dug and shifted, but did not fail. The little 2-pound guardian was actually quite difficult to pull up!
* Testing a 12-pound Northill as the tandem for a 35-pound Manson Supreme in medium sand. It lifted the Manson right out of the bottom and turned it into a kelet. Useless.
* Testing the same rig on small rocks and oyster it was the best rig. If one slipped off a rock in a shift, the chain would guide the other into a rock.
* Loosing a strain gauge to Netune, when a wave hit while anchored on non-stretch line. Huge force.

Eric Klem

Hi Drew,

I recently read the posts on your website about tandem anchoring and found them very interesting. Your results make intuitive sense and they show just how dangerous many people’s assumptions are with regards to spreading loads.

It will be interesting to see what you have come up with for snubbers. One of the big questions in my mind is how much stretch is appropriate. Because of the simplicity of the models that people have created, it is assumed that more stretch is always better but I suspect that in situations with badly behaved boats and limited fetch, loads may be lower (and the boat motion nicer) with more limited stretch.


Conor Smith

Great article as always.

I am a Ph.D student in Miami, FL studying physical oceanography and we use laboratory tests in our wave tank facilities to try to simulate some aspects of observable phenomena in the ocean.
In one of our experiments we were comparing designs of ocean drifters and studying their behavior in ocean environments. One thing we have observed and try to stick by is the 1/3 rule, that anytime you scale anything smaller than 1/3, the difficulty of scaling accurately is much much more complicated because not all processes(wind/wave forces/inertia) scale nicely.
In short, I agree that these scaled anchor tests should be viewed very dubiously! Not to mention that the sand boxes are DRY! How is that even closely related to the ocean?
Keep it up!


Brian Sheehan


If you google “Robert Taylor Navy Civil Engineering Laboratory” you will find anchor design and soil mechanics research that was done by Bob Taylor, who consulted for us during the 2014 Chesapeake Bay testing.

Bob has spent more than 45 years in the field and he has infinitely more knowledge than anyone on the pleasure craft side of the fence. Our late company founder kept a small library of his work, and here’s a link to his “Advanced Anchoring & Mooring Study” that you might find to be of interest:

Regarding anchor testing on dry sand, whether on a beach or in a box, that is totally practical, assuming of course that you are actually planning on anchoring there.

Best of luck with your studies.



Hi John
I replaced my old Bruce and purchased a 44 lb Spade.

Can anyone help me with three questions?

1. On my bow roller, the Spade is self-launching, so that there is nothing to hold the anchor aboard except the windless. I need to install a chain stopper or a hook snubber, correct?

2. While sailing, what is the best way to hold the anchor in place? Shall I simply lash the anchor down to the roller?… or install a bow roller that does not self-launch the Spade?

3. Finally, the end of the Spade’s shank brushes against the side of the fuller drum as it goes overboard. New bow roller or raise the roller?



Just wanted to share my real-world experience with the Spade : it’s amazing. Since we have it we have probably anchored a hundred times, and only once did it not set properly (the seabed was posidonia and rocks). When I say it didn’t set properly, I mean that it dragged when we applied 2500 rpm of pull on our chain ; it was probably ok for the night.

I can’t count the number of boats we’ve seen dragging around us, to the point that it has become the main thing we worry about while anchored.

Anyway, one more piece of anecdoctical evidence towards Spade !
Brann, s/v Lucy


Hello John,
Thank you for all of great advice and insight in your web site. I have however one observation regarding this subject matter – (albeit I intend to now buy a spade anchor to replace my CQR because I am persuaded it will be an improvement for me) – I note you are not a fan of the CQR at all, whilst Miles and Beryl Smeeton I gather used a CQR on Tzu Hang during all of their travels without apparent complaint – which I venture represents quite a practical test of an old favourite for some – just saying 🙂 With thanks


Thank you John and understood – I can see that had the Smeatons had the opportunity to read your very good article they would have bought a Spade also I expect (had the technology been invented by then). Thanks for this great resource – I learn something new most days !

Charles Starke MD

SV Panope Anchor Test “Video 100”

Heavy anchors scale 1-5 is best
Sarca Excel 48 lb 4.3
Rocna Vulcan 55 lb. 3.9
Mantus M1 45 lb. 3.9
Spade S60 45 lb. 3.8
Super Sarca 47 lb. 3.8
Ultra 46 lb 3.7
Manson S. 45 lb 3.4
CQR 45 lb 3.3
Delta 45 lb 3.1
Bruce 45 lb 3.1
Rocna 45 lb 2.8

Charles Starke MD

In reviewing the concluding table from SV Panope’s evaluation of anchor usage, I think certain factors are incorrectly weighted in his average to get a final reading. A simple average of different factors is used to get a judgment of anchor value.
1- Tip weight is much more important in anchor effectiveness and is probably the main factor in ease and quickness of setting. This is weighted equally to other factors that have no bearing on anchor effectiveness, and is probably the most important factor for anchor function and behavior. The Spade anchor has the highest tip weight of any anchor, and is therefore under-rated unfairly.
2- Galvanization life and ease of regalvanizing have no bearing on anchor effectiveness.
3- Cost of anchor has no bearing on anchor effectiveness, and is a small value compared to the value of a boat.
I believe all these factors mistakenly underweight the final value ascribed to the Spade anchor.
Best wishes,
Charles L Starke MD FACP

Charles Starke MD

Dear John
I agree. This last video, #100, is a compilation of results, but has no anchor test video. A subtext refers to the anchor video of a particular anchor to be able to view your anchor in action.

The different tests are summarized and the input numeral from one to five are detailed for each anchor and test.
But non-functioning judgements are weighted in equally to function.

Perhaps function average should be a separate judgement number than overall value to the owner’s pocketbook.
I would weight tip weight much higher than 5 since this is extremely important to function. And I personally don’t count galvanizing ease or quality highly valued since I expect my Spade to last my lifetime, especially since there is always paint!

But the video and the singular work of SV Panope is valuable, amazing and should be valued by the whole yachting world.
Best wishes,
Charles L Starke MD FACP
SV Dawnpiper

Eric Klem

Hi Charles and John,

This type of comparison is regularly used in comparison of design concepts and is known as a Pugh Matrix. Typically, there will be a weight applied to each column in order of importance. It is a useful tool for discussing trade-offs but it should not be used as anything more than a communication tool, it should not give you an answer without you giving it a gut check. I will often conditionally format the table so that in addition to seeing scores, we can see general trends of where concepts do well and where they don’t. Getting column weightings can be particularly tricky and I like a process known as the analytical hierarchy process where you compare 2 attributes against each other at a time and go through all combinations; it will even tell you if you are consistent or not.

There are however several pitfalls with doing Pugh Matrices. Some of them are:

  • It is impossible to fully describe needs with a Pugh, for example you may need a non-linear scaling or there may be an absolute minimum for some requirement. This comes back to doing a gut check and not taking results literally.
  • Getting the right column categories can be really tricky and it is hard to fully describe the needs.
  • Doing a 1-5 type scoring is tricky to keep consistent inside columns and particularly tricky to keep consistent across columns (people tend not to scale evenly which messes with your weightings). Some people like a 1,3,9 scoring to really accentuate good attributes but I don’t.
  • You have to watch double dipping on your columns. For example, if you had columns both for tip weight and performance, that would be double dipping, either you quantify performance and score on that or you quantify proxies for performance like tip weight, angle, shape, etc and use those.

What I see in this video is that Steve has provided a compilation of all of his data which is very important. Having worked with data a lot, I can assure you that if you don’t put it all in one place, people’s memories are shockingly bad and they tend to misremember data and therefore draw the wrong conclusions. Whenever taking a lot of data, some form of summary table of key metrics is always really helpful. I know that Steve adds up the scores but I personally ignored that. If you were cruising in bottoms similar to his, then you might come up with your own weightings and actually come up with comparative scores based on this table. I tend to agree with both of your thoughts that the performance categories would be heavily weighted and things like cost and galvanizing would have very small weightings. For those of us whose bottoms are not great matches, it is useful information but I personally am not actually adding it up. It would be great if Steve can keep expanding his testing to include additional bottoms such as really hard sand, heavy weeds, Chesapeake style silt, etc. There is a lot of good information in his videos and I really applaud the thoughtfulness and effort he has put into this hobby project (my understanding is that he really isn’t taking in enough money right now to make it anything else) and I think it is great he has a summary table but I also am ignoring his rollup scores.


Charles Starke MD

Hi Eric
I agree 100%. You have made a great analysis. Thanks
Best wishes
SV Dawnpiper

Rob Gill

Hi Charles,

We have a SARCA Excel, the anchor that came out top in the Panope anchor ranking. And I have to say, primarily because in NZ, the Spade is soooo… expensive that it made my eyes water so I couldn’t find my wallet.

But at the start of every cruise, I swear I’m going to swap out our anchor and buy a Spade. Until I remember that I need AT LEAST 5:1 scope to get the “damned” SARCA anchor to set. Don’t even bother trying to set on 4.9:1 (at 2500 rpm).

But when I viewed more of the links in the video, it demonstrates for instance that the Spade under-performs relatively in a boulder bottom, and also in soft mud. And due to the hollow and open shank of the Spade, he questions if it could be subject to unknown and hidden corrosion and shank failure. He adds his Spade started rusting quickly around the lead tip. Then he marks the Spade down for the difficulty of re-galvanising to mitigate the above issue (which is probably erroneous because the shank can be removed from the head). And remember the reviewer owns a Spade as his main bower, so is probably unbiased?

And he equally marks down the ROCNA due to the hollow roll-bar and visible manufacturing defects (and canes it for the re-set issues). But he also starts the video saying that the tests only apply in his waters and substrates – that different anchors may outperform in different waters – deep kelp?

But a question for you; if the SARCA Excel performs at least as well as the Spade overall (as tested in multiple substrates ), then why aren’t price, ease of re-galvanising (read longevity), and overall construction, relevant test criteria for rating/choosing an anchor?

I can attest our EXCEL after a real blow is dug in so deep, you are lucky to see the shackle (if you dive down), let alone the shank.

Thanks again for the link – nice to get some confidence back in our anchor choice!


Charles Starke MD

Hi Rob & John
Your comments and questions are all good. The price of the Spade at the time I bought it on sale was only slightly higher than a Rocna. And my decision was colored by experience like John. My Rocna on a new boat had recently dragged in a tidal area. A Spade on a previous boat had dug in so deep into Northwest Rockland Harbor, Maine, behind the breakwater, that the bow was pulled down to the water, and I motored around for one-half hour, until it came up clean and we popped up! I sold the 88 lb Rocna and bought a 99 lb Spade, and an aluminum Spade as a back-up anchor, the same size as a 120 lb galvanized. Dawnpiper is a 47’ Trintella, 40,000 lb.
I don’t think I heard how good the Sarca Excel was until after my purchases.
I also discounted in my decision the slightly higher cost compared with the value of the boat, the risk of galvanization failure in my lifetime because of availability of paint, and the risk to the shank because of beam-like strength and construction.
Both Spade and Sarca Excel anchors seem to be excellent. Previous experience may be the big determinant for both John and me.
Best wishes,
Charles L Starke MD FACP
SV Dawnpiper

Marc Dacey

After Season One of using our 30 kilo SPADE, I can’t fault it in terms of holding our steel 12 metre cutter to the seabed. Our only concern is that it buries so deeply we have to take our time breaking free and we have, in any soft bottom conditions, brought up a great deal of muck with the anchor, leading to some interesting contortions at the bow trying to knock off gobbets of mud before we get the anchor on the roller.

But the proof is in the pudding, so to speak.