Show me the Data!
First Data Collection - 50 Pound Scale
Two fencers were used to collect the initial data. Both men are White Scarves (and Pelicans) with about 25 years of SCA rapier experience. Master Llwyd is 6'2" and 240 pounds. Master Giacomo is also not a small man. Both are also previous Atlantian KRMs. Both are authorized in Spear on the armored field, although Llwyd has not put on heavy armor in some years.
In order to hopefully show a pattern, a sequence of weapons were tested: a plastic foil (as used for youth rapier), a 40" epee, a 35" schlagger, a 45" Darkwood Rapier, a Darkwood Two-Handed Sword, and an Alchem rapier spear. Each fencer struck the machine three times with each weapon. Any blows that visually struck well above or below the target center (marked with the attachment screw head) were ignored as those blows would magnify/reduce the measured force due to the change in moment arm. For the swords, we attempted to deliver blows at combat speed and force. For the two handed sword and the spear, several styles of attack were used and we were able to top out the 50 pound scale without delivering combat speed and force.
The raw data is collected in this spreadsheet. Note that for many of the 2 Handed and most of the Spear tests, the force exceeded the 50 pound capacity of the scale, so 50 pounds was what was recorded. A higher capacity scale was yet needed to capture the force from these weapons. This data is not included in the later results using the 100 pound scale.
A video was shot of showing the machine and our process. The new 50 pound scale was broken after the main data collection during the filming of the video. The scale's slider was non-functional (sliding partially back towards zero as force was releaved) when inspected after filming was complete. I suspect that was also true during the last portion of the video after we partially repaired things.
And via YouTube:
Summary of Results
Here's a quick summary of the results. I normalized the results so that the 35" Schlagger was the nominal "average" value and am reporting the percentage above or below that nominal for each weapon:
- Plastic foil: 43% of 35" Schlagger Force
- 40" Epee: 97% of 35" Schlagger Force
- 35" Schlagger: 100% of 35" Schlagger Force (duh)
- 45" Darkwood: 99% of 35" Schlagger Force
- 2 Handed Darkwood used 1 handed: 116%
- 2 Handed Darkwood used 2 handed: 125%
- 2 Handed Darkwood harpooning: 149%
- Spear w/ 2 fixed hands: 125%
- Spear w/ sliding hands: 151% **
- Spear at controlled combat form: 154% **
Discussion of Results
All of the information above is as factual as possible. This section starts to include opinion and analysis.
Before I go any further, I want to say that the fact that we measured the new Spears to deliver more than 150% of the schlagger and rapier combat speed forces while not being thrown at combat speed should seriously concern you. But, it is not sufficient evidence to ban them and forget them. What it is sufficient evidence of is that more careful data collection and experiments need to be performed and understood before we accept these weapons as safe.
Here are some valid quibbles you might have with our test that I want to acknowledge. The recorded test blows with the spear were the first time that Llwyd has ever thrown shots with one. Giacomo has not used one in any kind of actual combat. This data is also a very small sample using two fairly strong men. Finally, we're measuring peak force of a blow rather than the total energy of the shot.
At least three plausible reasons for higher force levels with the two handed weapons were proposed on the KRM mailing list: buckling end conditions (see this earlier article), more power from using two arms (and possibly hips and other body mechanics), and longer throw distances. This test only really accounted for the first idea as we did not throw any shots that would include the body mechanics and throw distances; those would likely add even more force to the shots.
Ideas for additional experiments
I will seek to broaden my data samples after aquiring a higher end scale. This will probably include setting up the machine at Pennsic. I hope, in particular, to collect data from skilled spearmen from the West and Caid (those Kingdoms that have the most experience with them).
I would encourage others to build their own machines and collect their own data (cost is in the $75-$100 range with $30 of that being the 100 pound scale). See the "tips" page for some guidance on that subject. As a Ph.D Structural Engineer/Rocket Scientist I understand verification, validation, and peer review (peer means something slightly different in my modern career) well and would embrace additional study of the subject. I am happy to discuss and advise as requested.
Other types of experiments could be run that might give additional insight. Critical buckling loads for different blades can be measured with a bathroom scale (push down until the blade snaps out of straight and record the force it took to cause that. Be consistant and report on how the blade was being held when you did this as the end conditions can radically change your results.)
Second Data Collection - 100 Pound Scale
Memorial Day weekend is one of Atlantia's biggest events, Ruby Joust. So, I brought the new Mark IV-A machine, my new Alchem spear (that arrived 2 days before the event) and went in search of a wider sample of fencers. I ended up recording 195 data points from 11 different people of varying size, strength, gender, and skill levels. After collecting data from fencers, I moved onto collecting data from some armored fighters to represent what cross-over fighters might do. I was sucessful in collecting data from one Knight, but the first unbelt to hit the machine broke it, so we were done for the event.
I was unable to locate a suitable picnic table, so I attached the machine to a counter in one of the on-site buildings. This ended up forcing most fencers to strike slightly upward, as if hitting their opponent in the face.
The combined impact of 250-ish hits (including those that hit high or low and were not recorded) caused the machine to slowly shift. I reset the hold-down clamps once, but eventually they were enough off center than a spear strike broke two of the (cheap!) clamps and released the machine from the counter. This in turn sheared the main deck at the attachment to the extension arm. The 5x5" target of thin plywood also showed some thin cracks on the face. I am very impressed that it has survived perhaps 500 hits over the full development and testing of the machine.
Summary of Second Data Results
The second round of tests were similar to the first. Since I was after volume of results we did not collect data for plastic foils or for epees. For the heavy rapier data I let people use their own sword or my 45" Darkwood heavy rapier.
For the 2-Handed sword testing we did three types of strikes. My Darkwood two-handed sword was used for all of the data collection. For the first round, the fencer held the sword one-handed as if it were a normal rapier. For the second round, they held the sword with both hands for the full strike. For the third round, the shot started two-handed, but the forward hand was released in mid-shot and only the hand near the pommel remained as the shot landed. This "harpooning" simulates a low-control melee shot that is attempting to extend range.
We also collected data for three types of spear strikes. All data was collected with my days-old Alchem spear with 5 foot shaft. The first type of hit was with two hands holding the spear in a fixed, static location. Only arm motion was used on the shot. The second type of hit allowed the fighter to slide their forward hand aft on the spear as they struck (pool-cueing). Again, this was only arm motion. The third type of shot tried to simulate a controlled melee-combat strike with the fighter taking a forward step (short lunge) as he/she attacked.
All shots with all weapons were attempting to recreate normal blows that the fighter would use in combat. None were "throw as hard as you possibly can". None simulated slipping, or being bumped, or the opponent moving into the shot, or any of the other dozens of things that happen frequently and increase the power of a shot.
The raw data from the second data collection session is in this spreadsheet. If you are interested in generating more statistics on the data, please have at!
The new 100 pound scale was indeed needed, but we did not exceed its range. We had a fair number of hits that exceeded 60 pounds and our highest force was 83 pounds.
Here is the quick summary of the new results normalized to the Heavy Rapier average force:
Even though the average values for each weapon and attack style showed a clear trend, there is a fair amount of scatter in the data, as shown in the histograms below.
Two Handed Sword Result Histogram
As with my previous discussion on the first data, this section will contain my impressions and opinions in addition to the raw data.
These results are fairly similar, but not identical to the first data set. Some of the quibbles from the first set of data collection are also applicable here. None of the fighters had ever held or used in combat the rapier-pike prior to testing. Few had used a 2-Handed sword even in single combat. But, the sample was much wider and included much smaller people, women, fairly new fighters, and a Knight.
The 2-Handed results are very close to those that we produced in the first test run. To me, there is a clear increasing trend. But, despite the mean values and the histogram peaks increasing, there is substantial overlap in the range of data for the three cases, casting some uncertainty on the trends. These results make me think that it could be possible to address the 2-Handed sword problem with technique rather than a ban. If we ban "harpooning" rather than the sword itself, the power level could come down significantly. Would such a ban on a technique be enforceable? And if so, would it be sucessful?
The spear results are significantly higher than previously reported. The three cases show a bit more spread than do the 2-Handed results, but there is still significant overlap in the range of results. The old results noted that we were topping out the 50 pound scale with almost every hit. Now that we can go to 100 pounds we can see just how much more force we were really generating. And doing more than 250% of the heavy rapier average with a stepping, controlled, combat shot scares me! When you add in adrenaline, a moving opponent, a tired cross-over fighter with minimal training, and other melee factors, strikes may easily exceed our laboratory results.
Let me phrase it less delicately: Atlantians may well be thugs, but if spears are allowed on the melee field, we will have them and you will be taking hits from us with them. Does that frighten you?
Having said that, I had a very nice talk with Master Aedan who pointed out that there is a substantial disconnect between these results and the two-years of test combats run in Caid and the West. We agreed that at the moment we're comparing apples and kittens. We're just not talking the same language in our discussions on the KRMs list. We all need to work to address that problem.
To that end, this page has extensively documented through text, photo, and video my approach to testing these weapons. I would LOVE for someone else to build their own machine and duplicate my tests to confirm or cast doubt on my results. That is what science is all about.
It would also be very much appreciated if those who feel they know how to use spears safely would post some video of them doing it. The text descriptions just aren't conveying what it is that I may be missing.
And, I'm hopeful that we can manage to have some kind of demonstration workshop at Pennsic to bring the east coast folks up to speed so that we're not speaking from ignorance but from the position of having a partial clue. But, we'll have to wait on the BoD and the SRM for guidance on such a gathering.
Some thoughts on Energy
Here is where basic spring/mass equations do apply to the rapier weapon strike problem. They apply to the scale that we're using to measure force, as it is just a large spring. Hooke's Law describes the force needed to stretch a spring by a desired distance:
F = k X
where F is the force, X is the distance, and k is the spring constant of the particular spring in question.
For our 100 pound scale, the printed tic marks go from 0 pounds to 100 pounds in the space of four inches. This indicates that our spring constant, k, is 100 pounds / 4 inches or k = 25 lb/in.
Now the kinetic energy of the weapon strike gets converted into the potential energy of the stretched spring in the scale. The equation for the potential energy of a spring is:
PE = k X^2 / 2
if we use Hooke's Law to replace X with F we get
PE = F^2 / (2 k)
So, we can use this last equation and the value of k for our spring to directly convert the measured machine force to the energy that was contained in the strike. Note that the force term is squared. Thus, a strike where the force was twice as high had four times the energy. This can produce some scary, but perhaps not all that insightful, results such as a strike that is measured at 250% of a reference force contains 625% of the reference energy.
Note that if you want to add a column on my result spreadsheet to generate energy values for the strikes, you'll probably want to do some unit conversions. Just plugging in k=25 to the previous equation will give you energy values in inch-pounds. Divide that by 12 to get foot-pounds. Then you can use a conversion factor chart like the one here to pick the factor you need to get to an energy unit that you understand. For instance, if you multiply the foot-pound result by 1.3558 you'll get energy in joules. Or if you multiply by 3.766E-7 you'll get kilowatt-hours (kWh).
Pennsic 43 Data Collection - 100 Pound Scale
I spent more than 12 hours over 4 days collecting data on the rapier field at Pennsic 43. I talked to scores of people, showed off the machine to the SEM, SRM, some DSRMs, many KRMs and a few KEMs as well.
I'd like to note a few things about the Pennsic data collection.
First, one inconsistancy in the results that is a case of Kingdom anthropology leading to an unintential gaming of the system. Many of the Eastern fencers delivered their heavy rapier blows standing straight, with fixed feet, and only arm extension. They assured me that they were hitting at combat power even though their delivery looked nothing like what you'd see in an actual fight. What they were unintentially doing, however, was gaming the testing rig. If they've thrown a few practice shots to set their range such that they're only going to deflect the target half an inch or less, they've set their maximum measurable force to 12.5 pounds or less since both the target and the spring register 25 pounds per inch of deflection. Following this realization, I insisted that later fencers assume an actual fighting stance and throw a blow such that it has the potential to blow through the target but does not due to the fencer's skill and control. So, due to this cultural practice, some of the East's heavy rapier results may be inappropriately low.
Second, each fencer used their own weapons if appropriate ones were available. Otherwise they used mine. This was to try to represent what is really happening on the field better.
Third, I encouraged each fencer to throw a few ranging shots before we started collecting data to get a feel for how the machine felt when they hit it. During one of these ranging shot periods, we saw a 90 pound spear strike. That data does not appear in the results, but I wanted to note that it happened. And if it happened once among 71 fencers throwing 3 recorded shots each, how many times per melee will it happen when there are dozens of fencers throwing scores of shots?
The new spreadsheet contains results from both the Pennsic and the Ruby Joust data collections. There are 1257 data points representing 71 fencers from 7 Kingdoms with 3 weapon types under a total of 7 shot conditions. You can download it here. A minor update was made 8/13/2014 to add 18 pieces of missed data.
Here is a quick summary of the results:
And here are the same results in terms of energy:
The spreadsheet contains all the raw data plus some simple statistical analyses. As before, I encourage people to mine the data for additional insights and results. Also included in the data is a breakdown on average force by Kingdom for each of the 7 shot types we studied. Keep in mind the possible inadvertant gaming of the machine noted above when examining these results. Finally, for fun, there is a "TQ" column on the raw data page in the spreadsheet. TQ is "Thug Quotient" and is a comparison of that person's average force values to the the overall averages. Values over 1 are hitting harder than average. Values less than 1 are below the average.
Discussion of results
The new numbers are nearly all lower than those for previous data sets. The 2-Handed Harpooning average force ratio decreased from 152% to 132% of the heavy rapier average. The combat strike with the spear's average force ratio decreased from 255% to 194%. A nearly doubling of average force is still a concern to me.
But, the average only tells part of the story; the peak forces and how frequently they occur is perhaps more important. With the larger sample our maximum recorded forces all went up. Some additional thought should probably be applied to understanding the likelyhood of the very hard strikes based on this sample. And some research on how much force or energy it takes to crush a mask or break a rib would give us the critical insight into how safe our armor standards are for these new weapons - remember they were designed to stop a broken foil, not a 90 pound blunt impact from a 5/8" diameter rubber tip.
I'll probably add more discussion in the future, but I want to get the data out to people to start thinking about.
Alternate Alchem Blade Results
Alchem was helpful and provided a less thick/less stiff blade for testing. I did not get a chance to mount the blade at Pennsic, but I did get to test it recently at our local practice. The normal stock Alchem blade is just over 1/8" thick when measured with my micrometer. The new blade is just over 1/10" thick. Thus far we've had 4 people throw 6 controlled combat shots each at the machine. So we have 24 data points (we have 177 controlled combat data points with the stock blade). The results for the two blades thus far are very similar:
The current, very small, sample does not show a noticable change in the force data. I will collect more data with the blade as the opportunity arrises.
Spear Safety Discussion
So, the big question from the spear data is "are they safe?" There are several possible failure scenarios that I want to try to examine. They are: concussion, rib breakage, and mask failure. Lord Gawin Kappler (Joe Shaffer) is a Ph.D. candidate in Neurobiology. He has put together some interesting research and data on the medical issues which can be found here.
Another serious concern that the machine data cannot address is off-axis contact such as from beats, parries, disengages. It would be easy to hit your teammates with the haft as you recover from throwing a strike. SCA melee fields are chaotic places and it is impossible to continuously have 100% awareness of everyone and everything around you. Nor can you have 100% control of your own weapon as people try to move it around so they can attack you.
Gawin has continued his literature search related to rapier concussions and has published several more articles on the subject:
2016 World Tour Results
To support the SCA rapier rules revision committee, the machine was sent on a World Tour around the society in 2016. Here are some summary charts of the results so far. For the first time we collected data on "hard" hits where the fencer was simulating someone stepping into their shot in such a way that they were unable to pull the shot. At Pennsic 45 we also collected data on Cut & Thrust percussive blow forces. Update 8/26/16: Based on some feedback from our West-An Tir war volunteer scientist I moved some of those results from the Heavy Rapier columns to other places. That has changed the results charts below. I have also updated the Excel file linked below that has all of the results.
The first figure compares some results from each data collection session:
The second figure combines all of the data collected to date:
The third figure sorts the heavy rapier results by Kingdom. Note that for many Kingdoms we have data from only 1-3 people (each person provides 3 data points) so that very small sample is an unreliable guess as to how the Kingdom actually fights. The "Ratio" column shows the ratio of the Kingdom's average to the Overall Society average. For instance, the data for Atlantia is 10% higher than for the Society as a whole. Let me make it clear that this particular chart is just for fun! The sample sizes for most of the Kingdoms is way too small to be meaningful.
Finally, if you want to look at the data and ask different questions using it, please download and have at it. If you come up with any great insights please let me know so I can report on them as well!
New Alchem Spear
During the last day of testing I was handed a new Alchem spear. It used an ash shovel handle and a saflex blade. It was significantly lighter than the old design. I only had time to run a few tests with myself as a subject. I hit the machine again with the old spear and then with the new one. My average with the classic flat tang blade was 38lb. With a vertically held blade on the new spear my average was 27lb and with a horizontally held blade it was 24lb. I like this new design much better!
Discussion of results
The higher average force for the testing at West-An Tir war (also reflected in the by Kingdom results for the West and An-Tir) concerns me a little. Since we had shipped my machine out for their use, there isn't any error from having a different measuring machine. So, either those two Kingdoms genuinely hit about 10% harder than the rest of the Society or there was a difference in the testing process. It's also possible that both factors had an impact. We are discussing doing testing at Great Western War that will hopefully give us a broader sampling of how the western Kingdoms play.
Otherwise, the new results are consistent with the old results for the same testing conditions. We had a few new tests that were run that are also interesting:
- Average "hard" blow force: 164% of average normal blow
- Average normal percussive blow force: 117% of average normal HR thrust
- Average "hard" percussive blow: (with a very small sample size) 188% of normal HR thrust
The new percussive cut power tests went fairly well. The configuration of the machine and table were not well arranged for lefties; they ended up throwing false edge blows most of the time. The original oak target was cut in half by a blow that was at least 82 pounds (see the saga page for pictures). That blow represented a proposed "military" level of cut and thrust intended to be thrown against significantly higher armor than the current minimum for C&T. A new plywood target was manufactured and used for the rest of the testing. Even with the polar fleece padding (two layers on the striking surface), damage to the top ply of the plywood was noticable on both the left and right sides of the target.