Toughness is a blade's resistance to chipping and fracture — its willingness to bend or dent instead of cracking when it hits bone, staples, or a knot in a log. It is the property people most often get wrong, because the intuitive story ("harder, more wear-resistant steel = more brittle") is only half true.
We calibrate a toughness model against Charpy V-notch impact measurements and express the result on a 1–10 scale. The calibration correlates 0.96 with measured Charpy energy. More useful than the score itself is what the model had to know to get there.
Carbide type beats carbide volume
Every knife steel is a soft-ish iron matrix with hard carbide particles embedded in it. Those carbides give edge retention — but they are also where cracks start. The naive model says "more carbide = more brittle." The data says the shape and coherence of the carbide matters more than the total volume:
- Vanadium carbide (VC) — roughly round, and coherent with the surrounding matrix. Cracks do not find an easy path around it. Mild toughness penalty.
- Chromium carbide (Cr7C3) — angular and incoherent. Its sharp corners concentrate stress and act as crack-initiation sites. Severe toughness penalty.
This is why two steels with similar total carbide content can have wildly different toughness. It is baked directly into the model's fitted coefficients:
| Feature | Coefficient | What it says |
|---|---|---|
| stainless_x_cvf | -0.85 | high carbide volume in a stainless matrix is the worst combination |
| vc_vol (vanadium carbide) | -0.84 | hurts toughness — but this is the "good" carbide |
| crc_vol (chromium carbide) | -0.30 | angular Cr carbides hurt on top of everything else |
| pm_tool (powder metallurgy) | +0.58 | powder processing meaningfully raises toughness |
| baseline (intercept) | 6.11 | a "clean" low-carbide steel starts here |
Powder metallurgy is a toughness cheat code
The single largest positive term is powder metallurgy. Conventionally cast steel solidifies slowly, so carbides grow large and cluster. Powder metallurgy (PM/CPM) freezes a fine powder before carbides can coarsen, giving a uniform dispersion of small carbides. Small, evenly spread carbides mean fewer large stress concentrators — so a PM steel is far tougher than the same composition cast conventionally.
You can see it in the rankings. The toughest steels the model scores are almost all powder tool steels:
| Steel | Toughness | Type |
|---|---|---|
| CPM 1V | 8.4 | PM tool steel |
| CPM 3V | 8.0 | PM tool steel |
| CPM Rex 45 | 8.0 | PM high-speed steel |
| CPM M4 | 7.8 | PM high-speed steel |
| Vanadis 4 Extra | 7.4 | PM tool steel |
Stainless is a toughness tax
The worst combination in the whole dataset is stainless matrix + high carbide volume. To be stainless, a steel needs a lot of chromium in the matrix; push carbon up to also get wear resistance, and much of that carbon forms exactly the angular chromium carbides that wreck toughness. That tension is why high-wear stainless steels sit near the bottom for toughness:
| Steel | Toughness | Note |
|---|---|---|
| ZDP-189 | 0.5 | very high C + high Cr stainless |
| CPM S125V | 1.4 | extreme carbide, stainless |
| M398 | 1.7 | high-vanadium stainless |
| M390 / CPM 20CV | 3.1 | popular high-wear stainless |
| D2 | 3.4 | semi-stainless, coarse conventional carbides |
Note D2's low score despite being a conventional (non-PM) steel: it combines a lot of chromium carbide with coarse conventional processing, so it gets hit twice.
MagnaCut, and why it made news
The reason CPM MagnaCut generated so much excitement is visible in the model: it scores 6.6 toughness while still being solidly stainless (6.3 corrosion). Historically you picked one — a tough non-stainless tool steel like 3V, or a stainless steel that chips. MagnaCut's composition was designed so that vanadium and niobium tie up carbon into the "good" round carbides, leaving enough chromium in the matrix for stainlessness without forming a mass of the angular chromium carbides that would tank toughness. It is a direct, deliberate exploitation of the carbide-partition chemistry described above.
The takeaway
If you want a tough blade:
- Prefer powder metallurgy — it is the biggest single lever.
- Understand that stainless + high wear resistance costs toughness; pick two of the three.
- Do not read a high carbide count as automatically brittle — vanadium-carbide steels tolerate more carbide than chromium-carbide steels of the same volume.
- For genuine hard use (batoning, chopping), the tool steels — 3V, CPM 1V, 4V — are tough for a reason, at the cost of corrosion resistance.