Why Sharp Edge Resin Dice Cost More Than Rounded (A Factory Breakdown)

Roughly once a week a buyer sends us a version of the same message: "Your sharp edge quote came back higher than your rounded quote — but it's resin either way, isn't it?" It is resin either way. That is not where the money goes. The gap between a sharp edge set and a rounded set is almost entirely a routing difference — the sequence of operations each die has to survive before it can be packed — and once you lay the two routings side by side, the price sheet, the lead times and the defect rates all stop being mysterious. What follows comes from the process sheets we work to on the floor at YSDICE, step for step, with the numbers our owner gave us rather than the ones that would look best in a brochure.

1. The cost driver is the routing, not the resin

Start with the material, because that is where most buyers assume the answer lives. A 16mm polyhedral die weighs a few grams. A complete 7-piece set is a small handful of resin, a trace of pigment and a smear of number paint. Sharp edge sets are poured in epoxy; rounded sets are poured in unsaturated polyester. The two chemistries do not cost the same per kilogram — but spread across seven small dice, the material delta lands somewhere in the range of loose change. It has never once explained a quote gap a buyer could actually feel.

What does explain it is bench minutes per die, and — much more importantly — whether those minutes are spent on one die at a time or on several hundred at a time. That single distinction is the whole article. Rounded finishing is batch work: a barrel full of dice tumbles for hours while an operator does something else entirely, and the labour cost divides across everything in the barrel. Sharp edge finishing is piece work, and worse than that, it is face work: a D20 has twenty faces, and each one is lapped flat individually against a levelling plate by a person who is watching the edge lines while they do it.

Multiply that out. A 200-set sharp edge order is 1,400 dice. Between the D4, D6, D8, two D10s, D12 and D20 in each set, that is 14,000 individual surfaces, and every one of them is flatted and polished by hand because the only machine that would do it faster would also destroy the product. That is the number the quote is really pricing.

2. The sharp edge routing: fourteen steps, most of them by hand

Here is the epoxy routing as it appears on our sheet, with what each step is actually protecting against. Nothing here is exotic; it is simply that none of it can be skipped or batched.

1. Design confirmation. Face layout, numeral style, edge treatment and set composition are locked before anything is cut. Changes after this point mean a new master.
2. High-precision master. 3D printed or CNC machined, then polished face by face. Every flaw in the master is copied into every die that will ever come off the mold, so this is the one step where hours spent are hours saved a thousand times over.
3. Silicone mold cast from the master. The mold has to reproduce corners and edge lines cleanly, because — as we will keep saying — this is where the geometry comes from.
4. A/B epoxy weighed on calibrated scales. Ratio is not negotiable and it is not eyeballed. You can never add extra hardener to speed curing; doing so produces a die that is soft, tacky or brittle, and it will not show up until days later.
5. Colour and vacuum degassing. The container needs real headroom, because the mix climbs dramatically under vacuum. A container filled to the brim is a mess and a lost batch.
6. Slow pour from the lowest point of the mold. The resin is introduced at the bottom and allowed to rise naturally up the cavity. It is never poured from height — a falling stream folds air straight back into a mix you just spent time degassing.
7. Pressure curing. The mold goes into a pressure vessel so that whatever residual bubbles remain are compressed until they are optically invisible. This is one of the reasons sharp edge production is fast in bulk, and we will come back to it.
8. Demold and post-cure. The D4, D8 and D20 corners are released first, individually, and only then is the whole die lifted out. Demolding too early, or pulling a die straight up out of a mold that is still green, is how you round a corner before it has ever met a grinding plate.
9. Sprue and flash removal, local flatting at the gate face. The gate leaves a witness mark that has to be brought back level with the rest of that face — not below it.
10. Face-by-face flat lapping on a levelling plate. The heart of the process. Equal removal per face, checked constantly. Take too much off one face and the die goes lopsided, the edge lines bend visibly, the corners blunt, and the centre of mass shifts — which is a balance problem, not merely a cosmetic one.
11. Ultra-fine polish with plastic polishing compound. Aggressive tumbling is forbidden at this stage. A barrel would produce a beautiful gloss in twenty minutes and destroy the entire reason the customer bought sharp edge dice.
12. Number paint fill. Every trace of grinding dust and polish wax has to be cleaned out of the number channels first, or the paint sits on contamination and lifts later.
13. QC. Geometry, edge lines, clarity, colour, numeral fill and finished dimensions.
14. 7-piece set matching and packing. Seven shapes that look like they were born together, not seven dice that happen to share a pigment.

3. The rounded routing: fourteen steps, most of them in a barrel

Now the unsaturated polyester routing. The step count is identical — fourteen — which is exactly why "number of steps" is a useless proxy for cost. Look instead at which steps are per-piece and which are per-batch.

1. Design confirmation. Same discipline, same lock point.
2. Silicone mold preparation. Dust, moisture and cured residue from the previous run are all removed. Any of the three prints itself onto the next casting.
3. Resin colouring, with MEKP curing agent added last and at low speed. In systems that are not pre-accelerated, the accelerator must be fully dispersed first. Adding hardener quickly or early gives you streaks, hot spots and a cure that races the mixer.
4. Vacuum degassing. Same principle as the epoxy line, same need for headroom.
5. Casting and inclusion placement. Two-tone and gradient effects are built with split or layered pours. Inclusion pieces — animals, figures, machined cores — are set once the resin has reached the right viscosity, then topped up. Too thin and the piece sinks or drifts; too thick and it sits proud and traps air underneath.
6. Room-temperature cure. The mold is kept level, ambient temperature is controlled, and batch size and exotherm are managed deliberately. Get this wrong and you get under-cure, hot spots, shrinkage, cracking or colour drift — sometimes all five in the same tray.
7. Demold and post-cure.
8. Sprue removal and screening. Chips, cracks, severe bubbles and off-centre inclusions are pulled out here, before any finishing labour is invested. This is the cheapest place to lose a die, and we lose most of them here on purpose.
9. Coarse tumbling. This is where the rounded profile is actually created. Not in the mold — in the barrel. The dice go in with hard corners and come out soft-edged.
10. Medium and fine grinding. Media is stepped down progressively. Over-grinding is a real failure mode: dimensions shrink out of tolerance and the engraved numerals thin out until they will not hold paint properly.
11. Fine polish, with a wash between every single stage. One coarse grain carried forward into a fine barrel will scratch an otherwise finished face, and you will not notice until the polish is on.
12. Number paint fill.
13. Cleaning and QC.
14. 7-piece set matching and packing.

4. The one sentence that explains the entire price sheet

Our internal process document puts the difference in a single line, and it is worth quoting because everything else on this page is downstream of it.

The rounded die's soft edges are formed mainly by multi-stage tumbling; the sharp die's corners and edge lines are formed mainly by the high-precision mold, and post-processing must be face-by-face flat lapping plus fine polishing.— YSDICE process sheet, resin dice routing

Read that twice, because it inverts the intuition most buyers arrive with. People assume "sharp" means an extra machining operation was added on top. It is the opposite. The sharp edge die comes out of the mold already correct, and every subsequent operation exists to preserve that geometry while making the surface optically clear. The rounded die comes out of the mold with corners it is not supposed to have, and the finishing process exists to remove them.

That is why the expensive process is the one with no batch step in it. The moment you introduce a barrel to a sharp edge die you have manufactured a rounded die at a sharp edge price. So the barrel is off the table, and the labour cannot be divided across a batch. It is divided across nothing. It is per face, per die, per set, every time.

5. What goes wrong on each line — and who pays for it

Both routings have characteristic failure modes, and understanding them tells you what you are really buying when you compare two quotes.

The symptoms those produce are tackiness, bubbles, cracking, yellowing and geometric distortion. Note how many of them are invisible at demold and only announce themselves days later — a set that yellows or goes tacky in week two has already absorbed its entire labour cost before anybody knew it was scrap. That risk is priced in, and it should be.

The rounded line fails differently. Because the cure happens at room temperature without a pressure vessel, the process is far more sensitive to the weather inside the workshop: humidity, ambient temperature and batch size all push the exotherm around. Under-cure gives you dice that grind gummy. Hot spots give you internal stress that cracks weeks later. Shrinkage moves the dimensions. And colour drift between trays is the quiet one — it does not fail QC on any individual die, it fails at set matching, when you discover that the D4s from Tuesday do not sit convincingly beside the D20s from Wednesday.

Then there is the finishing risk unique to the barrel. Over-grinding does not announce itself either. It takes the dice progressively out of dimension and thins the numerals until the paint fill looks starved. That is a judgement call made by an operator watching a running barrel, and it is one of several places in this routing where experience is the process control.

6. The defect rates we actually run

Very few factories will put these numbers in writing, which is precisely why they are worth putting in writing. They come directly from the owner of the factory, and they are measured, not estimated.

Product	Defect rate	Sets cast per 100 good sets
Rounded resin, no inclusion (plain, solid, swirl, glitter)	about 5%	about 105
Rounded resin with an inclusion or machined core	20–30%	125–140

Five percent on plain rounded dice is a healthy, boring number for a cast process — a few bubbles, a few chips, the occasional colour outlier. Twenty to thirty percent on inclusion dice is a different category of problem, and the honest explanation is not that the parts are bad. At YSDICE we machine the cores ourselves to tight tolerances, so the fit inside the cavity is fine. The variable is placement. A bench worker sets each inclusion by hand, into resin sitting at a specific viscosity for a window of a few minutes, aiming for a piece that will still read as centred through six or twenty transparent faces once the die is finished. Whether that lands is a function of that worker's experience, and no amount of tooling removes the judgement from it.

Two further wrinkles make inclusion defects expensive rather than merely frequent. First, most of the scrap is caught early, at the screening step before the barrel — that is the cheap place to lose a die, which is why we screen hard there. But a second wave appears during grinding: a bubble sitting just below the surface opens into a visible pit as the media takes material away, and a core that looked acceptable through a frosted as-cast surface reveals itself as off-centre once the faces polish clear. Those dice have already consumed most of their finishing labour by the time they are condemned.

Second, dice are sold in sevens. A scrapped D20 does not cost you one die — it can strand six perfectly good dice on the bench until the next casting cycle produces a matching partner. That is a cost problem and a schedule problem, and it is the single best argument for over-casting deliberately rather than casting to order and hoping.

7. Worked example: what a 20–30% defect rate does to a 200-set order

Abstract percentages do not help anyone build a budget, so here is the arithmetic on a realistic order. The cost figures below are illustrative index numbers, not a quote — your real price depends on size, colour, inclusion complexity and packaging — but the ratios are exactly what we see on the floor.

Say a plain rounded 7-piece set costs $4.00 to put through the line: resin, pigment, mold occupancy, tumbling media, paint, labour and packing. An inclusion set costs more before any scrap is counted, because the core has to be machined and placed — call it $5.20 through the line. Now apply the yields.

Scenario	Defect rate	Sets cast to deliver 200	Line cost	Cost per delivered set
Plain rounded	5%	211	211 × $4.00 = $844	$4.22
Inclusion, good run	20%	250	250 × $5.20 = $1,300	$6.50
Inclusion, hard run	30%	286	286 × $5.20 = $1,487	$7.44

Three things fall out of that table. The first is the size of the jump: an inclusion set is not "a bit more" than a plain rounded set, it is 54% to 76% more before anybody adds a margin, and the majority of that increase is scrap rather than the core itself. The second is the width of the band. The difference between a 20% run and a 30% run is $0.94 per set — $188 on this order — and it is decided by casting conditions and bench judgement, not by anything the customer chose. That is why we quote inclusion work against a range internally and then hold the agreed price externally: absorbing that variance is our job, not yours.

The third is the raw volume. At 30%, delivering 1,400 good dice means casting a little over 2,000 — roughly 600 dice scrapped to fill a 200-set order. The shop rule of thumb we work to is blunt and easy to remember: for every 100 good sets, plan to cast 125 to 140.

For sharp edge epoxy the arithmetic works differently again. The pressure vessel takes most of the bubble risk out of the equation, so scrap rates sit well below inclusion work — but the labour per surviving die is high and irreducible, and a die condemned at the lapping stage has already consumed nearly all of its bench time. Sharp edge is expensive because of what it costs to make right, not because of how often it goes wrong. Inclusion rounded is expensive because of how often it goes wrong. Those are two completely different pricing mechanisms wearing the same word: "resin".

8. Lead times: why the "simpler" process takes longer

Here is where buyers are most often caught out. Sharp edge dice are the more labour-intensive product by a wide margin, and they are also the faster product to deliver in bulk.

Material	Light custom sample	Light custom bulk	Mold (full custom)	Full custom sample	Full custom bulk
Metal (zinc alloy)	3–5 d	10–15 d	20–25 d	3–5 d	20–25 d
Sharp edge (epoxy)	3–5 d	10–15 d	20–25 d	3–5 d	15–20 d
Rounded (unsaturated polyester)	7–10 d	25–30 d	15–20 d	7–10 d	25–30 d
Acrylic	5–7 d	20–25 d	20–25 d	—	—
Silicone	5–7 d	20–25 d	20–25 d	—	—
Gemstone	5–7 d	15–20 d	no mold required	—	—
PU dice tray	5–7 d	15–20 d	light custom only	—	—

Sharp edge bulk runs 10–15 days on light customisation and 15–20 days on a full custom mold. Rounded bulk runs 25–30 days either way. Roughly double, for the process with fewer hand operations. There are four reasons, and they are all structural rather than organisational.

Curing cannot be compressed. Sharp edge epoxy goes into a pressure vessel, which shortens the useful cure window and eliminates most bubble rework. Rounded polyester cures at room temperature, and that cure is governed by exotherm — which means batch sizes have to stay modest on purpose. You cannot pour one enormous batch to save a week; that is precisely how you produce hot spots, shrinkage and cracking. The calendar is set by chemistry you are not permitted to rush.

The finishing ladder is serial, and the barrel is a shared resource. Coarse tumble, medium grind, fine grind, fine polish — four stages, each measured in hours, with a full wash between every one so no coarse grain travels forward. Those stages happen in sequence and cannot overlap for the same batch. Sharp edge lapping, by contrast, is parallel: put four more trained people on the bench and the job finishes sooner. Rounded finishing does not respond to adding people. It responds only to adding barrels and adding days.

You are producing 125–140% of the order. Every one of those over-cast sets goes through every stage of the ladder too. A 25–40% overproduction is a 25–40% larger job in every barrel, at every wash and at every inspection point.

Set matching is the last gate. After tumbling, appearance and dimensions have drifted slightly across trays, and seven shapes have to be matched into a set that looks deliberate. If one shape comes up short — and the D4, with the least material at its points, and the D20, with the most surfaces on which to fail, are the usual suspects — you cannot simply substitute. You cast again, cure again, tumble again. That single risk is why we hold the rounded window at 25–30 days rather than promising 18 and apologising on day 24.

The same logic explains the sample stage, which is where the difference is most visible to a buyer. A sharp edge sample takes 3–5 days because a silicone mold can be cast off a printed master quickly and pressure-cured immediately. A rounded sample takes 7–10 days because the geometry is the finishing — an untumbled rounded sample tells you nothing at all about the product you will eventually receive, so the sample has to run the entire multi-stage ladder at production settings. There is no honest short version.

9. What we write down for every batch

Both routings sit on top of the same batch record, because a defect rate you cannot explain is a defect rate you cannot reduce. For every batch we log the raw-material lot, the ambient temperature and humidity in the workshop, the mix ratio, the degassing result, the mold cleanliness check, the inclusion moisture content, the cure state, the appearance consistency and the finished dimensions.

That list looks bureaucratic right up until the day a customer reorders eleven months later and the colour has to match. Then it is the only thing standing between "the same blue" and "a blue somebody mixed from memory". It is also how we can tell you with any confidence that inclusion work runs at 20–30% rather than guessing at a comfortable-sounding figure — the records are where that number came from in the first place.

YSDICE — also known as Dongguan Yushun Hardware Co., Ltd. — is a registered dice manufacturer in Chang'an, Dongguan, and every one of those records lives on our own floor rather than at a subcontractor. That matters for exactly this reason: when a run behaves differently from the last one, we can walk over and look at what changed instead of forwarding a question to somebody else's factory and waiting three days for a shrug.

10. Material documentation, described accurately

Buyers in the toy and games channel need paperwork, and the paperwork here is specific, so let us be precise about what exists and what it covers.

Our supplier holds an SGS report on the unsaturated polyester raw material, tested to EN 71-3:2019+A2:2024, Category III, with all 19 regulated elements reported as not detected. There is a separate RoHS report covering 31 substances on the number-fill paint. Both reports are held by the material suppliers, and we can forward them by email on request.

What those documents are is raw-material and input testing. What they are not is a product-level certificate for a finished die, and we will not describe them as one. If your channel requires finished-product testing to EN 71 or ASTM F963 on the specific SKU you are ordering, that is a separate test on a separate sample, and it should be scoped and budgeted at the quotation stage rather than discovered at customs. Dongguan Yushun Hardware Co., Ltd. will arrange it happily; we simply will not pretend the supplier reports already did it.

11. Which one should you buy?

With the routings, the yields and the calendars all on the table, the choice becomes a lot less about taste and a lot more about what the product actually has to do.

If your priority is…	Choose	Why
Collector appeal, limited drops, photography	Sharp edge epoxy	Flat faces and crisp edge lines are what the camera and the collector are buying. Nothing else looks like it.
Lowest unit cost	Plain rounded polyester	About 5% defect rate, batch finishing, no per-face labour. This is the efficient product.
Fastest bulk delivery	Sharp edge epoxy	10–15 days light custom, 15–20 full custom, because pressure curing and parallel bench work both compress the schedule.
Animals, figures, machined cores, novelty inclusions	Rounded polyester	It is the only routing that supports inclusion placement. Budget for 20–30% yield loss and a 25–30 day window.
Two-tone, gradients, layered colour	Rounded polyester	Split and layered pours are native to this process.
Heavy daily table play, gifts, event giveaways	Rounded polyester	Soft edges roll comfortably, survive being dropped, and cost less per set to replace.
Toy-channel scrutiny on material inputs	Rounded polyester	The EN 71-3 Category III supplier report sits on this material.
A fixed launch date you cannot move	Sharp edge epoxy	The shorter, more predictable production window is worth real money when a slip costs you a shipping wave.

A pattern worth naming: the two products are not really competitors, they are different jobs. A great many of the campaigns we quote at YSDICE end up running both — a sharp edge core set as the collector tier, and a rounded inclusion set as the novelty or deluxe add-on — precisely because each one is unbeatable at something the other cannot do at all. Quoting them together also lets us stage the calendar properly: the rounded line starts first because it needs 25–30 days, the sharp edge line starts later, and both land in the same week for packing.

12. Sharp edge vs rounded cost: FAQ

Is sharp edge resin actually a more expensive material? Epoxy and unsaturated polyester do differ in price per kilogram, but across a 7-piece set the material difference is trivial. The quote gap is labour and yield, essentially all of it.

Could you tumble sharp edge dice to save money? No — and that is not a policy position, it is a definition. Tumbling is what creates a rounded profile. Putting a sharp edge die in a barrel produces a rounded die. The corners are the product.

Why is my rounded sample taking longer than my sharp edge sample? Because a rounded die's geometry is created by the finishing ladder, so the sample has to run the whole ladder at production settings in order to be representative. Sharp edge samples come off a mold that already has the correct shape, so 3–5 days is achievable. Rounded is 7–10.

Can you guarantee my inclusion order will hit 20% rather than 30% scrap? No, and any factory that guarantees it is guessing. What we do instead is quote against the band, over-cast by 25–40% from the start, and hold the agreed unit price whichever end of the band the run lands on.

Does a higher defect rate mean lower quality dice reach me? It means the opposite, provided the screening is real. A 5% rate on inclusion work would tell you that nobody is inspecting. The dice we scrap are the reason the dice we ship look the way they do.

Which format holds up better on a collector shelf? Sharp edge, comfortably — flat faces stay flat and edge lines are what buyers photograph. Rounded sets are for playing with, and they wear beautifully because there is no sharp geometry left to lose.

Do both formats support custom numerals and logos? Yes, on both light customisation and full custom molds. On rounded dice, remember that over-grinding thins numerals — which is one more reason the finishing operator's judgement is worth paying for.

What is the single biggest mistake buyers make here? Comparing a sharp edge unit price against a rounded unit price as though they are the same product bought at two prices. They are two different manufacturing routes with different yields, different calendars and different failure modes. Compare them on the job you need done, then let the routing decide the number.