Every penetrating tool ends in a decision. Most of the time that decision is inherited rather than made.
The conical point is what the icepick came with. It arrived from the ice trade already finished, and the men who later found other uses for it were not in a position to be fussy about tip geometry. They wanted something that worked and something that could be explained away if found. The cone did both. What went unasked, for the better part of a century, was whether it was the right shape.
Two things made that question easy to avoid. The cone has structural advantages that invite no complaint: its geometry distributes mass evenly around the axis in the manner of a convex grind — the strongest blade geometry — which makes it resistant to lateral force and difficult to chip or deform under off-axis load. And it is the natural product of a lathe. Taper a round bar, refine the taper, polish. The icepick point was always going to be a cone because that is what a lathe produces without extra decisions. Ease of manufacture and structural robustness together are a strong argument for not asking questions.
I asked. The chisel tip was the answer, and it changed most of what I make.
The Cone Everyone Inherited
A conical point is radially symmetric. It tapers to a single apex. Every cross-section along its length is a circle. That symmetry is the source of both its virtues and its limits — worth being precise about which is which.
Mechanically, a cone penetrates by displacement. It forces material aside radially. In elastic tissue — skin, muscle, fascia — that material is pushed apart rather than parted. The fibres stretch around the advancing shaft. They do not separate. They move.
This matters more than it sounds. Elastic tissue does not stay where you put it. The moment forward pressure stops, it recoils. The displaced fibres close back against the shaft and grip it along its full inserted length. Entry is comparatively easy. Recovery is not.
Anyone who has driven a round punch into dense material and had to work it back out understands this without needing the vocabulary. The tool goes in cleanly. It comes back reluctantly. The reluctance scales with depth. The chisel tip addresses that last property directly.
What Changes When the Point Becomes a Plane
A chisel tip replaces the cone with two converging planes terminating in a straight leading line. The point is no longer a point in the geometric sense. It is a short line.
That line does something the cone cannot: it parts material along a defined axis instead of pushing it outward in every direction at once. The resulting opening is not a round hole stretched into existence. It is a slot.
Entry
At the moment of contact the chisel presents a much smaller initial bearing area than a cone of equivalent diameter — a line rather than a growing circle. Penetration begins at that line and only afterwards becomes a wedging action along the converging planes. In practice this means the tip finds purchase on hard or angled surfaces where a cone would skate.
Extraction
This is where the difference stops being academic.
Because the chisel has parted material rather than displaced it, the tissue surrounding the shaft has been structurally separated. It does not recoil into contact with the same force, because there is less stored elastic tension holding it there. The opening is a slot with parted margins rather than a stretched aperture straining to close.
Extraction resistance drops accordingly. The tool comes back the way it went in.
I want to be careful about the strength of that claim. I am not offering a measured coefficient. What I am offering is the consistent result of making both geometries for fifteen years and listening to people who have used them in conditions I would not choose. The pattern does not vary.
An Additional Capability
There is a secondary behaviour that emerges from the geometry rather than being designed into it. A chisel tip, presented at the correct angle, can scrape and define rather than simply penetrate — in the same way that a wood chisel used laterally in a mortise pares and cleans the surface rather than splitting the grain. This is not the primary purpose of any tool I build. But it is a consequence of having a geometry that produces a defined plane rather than a displaced hole, and it is worth noting as part of what the form makes possible.
Penetration Is Not Cutting
Here is the part that most discussions of this subject get wrong before they begin, and it is worth stating plainly because everything else depends on it.
I do not sharpen these tools. They do not cut. They are not intended to.
On the pry tools I leave between 0.5 and 1 mm of thickness at the tip. That is not a compromise, it is a requirement: the tool has to work as a lever and a driver, and any sharpened apex would roll or chip the first time it took a lateral load. It is measurably blunt, and anyone with a caliper can confirm it.
On the spikes and the karambits the two chisel planes converge to a line, and on a cylindrical spike it is the intersection of the shaft with the angled cut. I take that convergence as far as it goes and then polish it. I do not hone it. I make no claim that it cuts paper, and I would not make that claim even if it sometimes did, because it would be difficult to demonstrate consistently at fine sections and I am not willing to tell a buyer something I cannot stand behind.
What does the work is not sharpness. It is a convergence that is well defined and well polished, combined with the force a reverse or hammer grip permits you to put behind it.
That distinction is the whole argument. Penetrating and cutting are different mechanical operations. Cutting severs material laterally along an edge. Penetration parts it ahead of an advancing point. A tool can be excellent at the second while being incapable of the first, and designing for the second frees you from every constraint the first imposes — maintenance, fragility, and a good deal of law.
So the range divides into two grinds doing two jobs. The pry tools are blunt by requirement. The spikes carry a polished convergence with no sharpened apex. Neither claims to cut, and neither needs to.
What the Forensic Record Shows
There is independent documentation of this, and it comes from people with no interest in my conclusions.
Forensic pathology has a small but real literature on injuries produced by screwdrivers — a blunt, unsharpened, converged implement, which is to say the closest mass-produced analogue to what I make. Roger Byard’s Patterned injuries from screwdrivers, published in Forensic Science, Medicine and Pathology in 2022, sits alongside Croft’s earlier study in the Journal of Forensic Sciences and more recent work by Reyes Albaladejo in the same journal.
Two findings matter here.
The first is that tip geometry is legible after the fact. A cruciform driver leaves a cruciate pattern reproducing its own cross-section, consistently at around five millimetres square. A flat driver leaves something entirely different. The geometry of the tip is recorded so faithfully that a pathologist can work backwards from the pattern to the shape of the implement.
I have a personal observation that confirms this from the making side. Ice Pry models finished with a minimal secondary bevel at the convergence — what I call a microfilo — leave a distinctive H-shaped footprint in cardboard and EVA foam. The vertical bar of the H is the slot the leading line parts; the horizontal bars are the impression of the two faces as the tool advances. The geometry writes itself into the material. The cross-section of the tool is readable in what it leaves behind, which is exactly what Byard’s literature describes from the other end of the process.
I know of no stronger way to state the thesis of this article. Tip geometry is not a styling decision that slightly modifies performance. It is the determining variable, and it is determinative enough to be read off afterwards by someone who was not there.
The second finding is the one that speaks directly to what I build. A flat, slotted tip produces roughly rectangular openings with no tissue bridging, with margins subtly split, classified in that literature as a hybrid of sharp and blunt force.
Tissue bridging is the signature of blunt tearing — the strands that survive when material is ripped rather than parted. Its absence means the material separated cleanly. And it separated cleanly under a tool with no sharpened edge whatsoever.
That is the mechanism I described above, documented by people studying something else entirely. A defined, converged, polished line parts material without needing to cut it.
A separate observation — which I offer as working hypothesis rather than established mechanism — concerns the edges I leave near the tip. I finish them without chamfer, at ninety degrees, so they continue separating material as the tool advances beyond the leading line. My interpretation is that this produces a wound character distinct from a clean incision: the slot is defined, but the margins are worked rather than merely parted. Whether this translates into any practical difference in effect is not something I can demonstrate in a laboratory. I note it because it is part of how I build, and because the forensic literature’s hybrid classification — neither purely sharp nor purely blunt — suggests the mechanism is real, even if its consequences are not fully mapped.
The honest caveat: a hardware screwdriver is not one of my tools. Different steel, different finish, different section, no attention paid to the convergence at all. The analogy is close — far closer than the surgical-needle literature people usually reach for — but it remains an analogy, and I will not pretend it is a test of my work.
Byard also notes, in passing, that screwdrivers are surprisingly rarely used in fatal assaults. The literature exists because the cases are unusual enough to be worth writing up. I mention it because I would rather give you the inconvenient half of a source than have you find it yourself.
There is a second, institutional leg to this. The United States National Institute of Justice publishes Standard-0115.00, governing the stab resistance of personal body armor. It does not treat penetrating threats as one category. It separates edged blade threats from spike threats and tests them under distinct protocols, because a point and an edge defeat protective material by different mechanisms, and armor that stops one does not necessarily stop the other.
An institution charged with keeping people alive reached the same conclusion I did, from the opposite direction. Point and edge are not variations on a theme. They are different problems.
The Argument Nobody Makes
Everything above concerns behaviour at the moment of use, which is the part people enjoy discussing. The advantage that actually determines whether a tool is still worth carrying in a decade is duller than that.
A conical point cannot be restored by hand in any meaningful sense. To bring it back you must remove material evenly around a full circumference while maintaining concentricity and a consistent included angle. That is a task for a fixture and a patient operator. In the field it is not a task at all. What happens instead is that the point degrades — dropped on concrete, corroded, worked against something harder than it — and stays degraded, because restoring it properly is beyond what the owner can do and beneath what they will pay someone else to do.
A converging flat is one plane. Any flat stone brings it back. So does the unglazed base of a ceramic mug, which is what most people actually have to hand. Restoring it requires no fixture, no skill beyond holding a consistent angle, and no decision more complex than which way up to hold the tool. Note what this is not: it is not sharpening. It is returning a polished convergence to its intended shape.
A tool that can be maintained is a tool that is still a tool in twenty years. A tool that cannot is an object that used to work. The geometry that makes the difference is not the one that performs best on the first day. It is the one that can be brought back on the four thousandth.
Where the Cone Still Wins
If this reads as though the chisel is simply better, I have written it badly. There are three conditions under which I would still choose a cone.
The cone requires no indexing. It is symmetric, so it behaves identically regardless of how it is held. A converged flat has an orientation, and orientation is one more thing to get right at a moment when fine motor control has already left the building. For a user with no training base and no intention of acquiring one, that symmetry is worth more than any extraction advantage. It is also why several of my designs address indexing directly rather than working around it — the blind-index geometry behind the Blink Grip exists because orientation seemed to me the wrong thing to leave to chance.
That said, this advantage deserves qualification. At the shaft diameters I work with — typically three to ten millimetres in the Voodoo range — the geometric offset between the tool’s centreline and the leading line of a chisel tip is small enough to be practically irrelevant. Under stress, when fine motor control has already degraded, the additional orientation required is a fraction of the variation already present in the grip. I made the same observation when writing about chisel grinds in tactical knives. The indexing advantage of the cone is real at the drawing board. Whether it survives contact with the conditions it is supposed to address is a different question.
The cone resists lateral load equally from all directions. Material is distributed evenly around the axis, so a bending force from any angle meets the same resistance. A converged flat is asymmetric and has a strong plane and a weak one. Against unpredictable off-axis loading, the cone is the more forgiving structure.
And against genuinely hard barriers, the mass sitting immediately behind a conical apex resists deformation better than the thinner section behind a converged line. If the anticipated use involves striking bone or rigid material at unfavourable angles, that difference is not theoretical — it is exactly why the pry tools keep their half millimetre.
The chisel is the better geometry for most of what I design. It is not the better geometry in every case, and I have never claimed otherwise.
A Note on Priority
On 3 October 2011 I published a tool called the Ice Pry, described then in my own words as a lever, a wedge, a screwdriver or a punch. Posts that followed extended that description to include dagger and push dagger — which is exactly how I had conceived the tool’s potential from the start, and how I was already using it. That record remains online and remains dated. It is the first public appearance of a flat converged tip applied to a tool built to penetrate, and it is the origin of a geometry that now runs through the Voodoo series, the Ice Pry family, the Compact Pry Knife, the Goliath Tarantula and the Micro Edge line.
I want to state the claim narrowly, because narrow claims are the only kind worth making.
I did not invent the chisel point. It is older than any of us, and claiming it would be both unprovable and untrue. What I am stating is that its application to dedicated penetrating tools appears in my published, dated work from 2011, and that I was making them before I began publishing.
The geometry has become more common since. I take that as confirmation rather than as grievance — a design nobody copies is usually a design nobody needed.
What travels is the shape. The reasoning stays here.
— A.N. Nash | Bladetricks