TRIAL OF SWORDMANCY: 640K STOCK BILLS A DAY | ENDFIELD HUB

Everyone agrees you should stop drawing at 10 Battle Points in Trial of Swordmancy. Fine. But here is the question almost nobody answers with an actual number: played as well as it can possibly be played, how many Wuling Stock Bills does the mode pay per day? And how close can you get to that ceiling without running a calculator next to your game? We sat down and modeled the whole thing as an expected-value problem, simulated millions of days, and came back with two numbers. Played perfectly, Trial of Swordmancy is worth roughly 640,000 Wuling Stock Bills a day. Played with a dead-simple rule you can execute half-asleep, it is worth roughly 618,000. That 22K gap is the entire price of not carrying a spreadsheet into your dailies, and for most players it is a price worth paying.

TL;DR - Key Points

• Perfect play earns ~640,000 Wuling Stock Bills per day. That is the expected value of a fresh daily state across the 3 Rewarded Trials, computed exactly, not eyeballed.
• A simple by-hand strategy earns ~618,000 per day, about 22K less (roughly a 3.4% tax for not consulting a program every draw).
• The simple rule: keep a hand of 9 or 10, otherwise draw if you can or reroll if you have rerolls left. Out of rerolls, lower the bar to accept 8, 9, or 10. Always activate Rewards Doubling on 2 of your 3 trials.
• Rerolls and Doubling charges are shared across all 3 trials, not per-trial. The “game” you are optimizing is the whole day, not a single hand.
• The optimal strategy is a dynamic-programming table. Every reachable game state has an expected value; you always pick the decision (draw, keep, reroll, double) with the highest one.
• The current card pool has 10,980 reachable game states. The simple strategy makes the same call as the optimal table about 85% of the time.
• The card pool rotates every 3 days. The math does not change, only the inputs, so these numbers hold up across rotations we tested.

The Problem Is the Day, Not the Hand

The mistake most strategy talk makes is optimizing a single Trial in isolation. Draw to 10, do not bust, collect. That is correct within a hand, and our Dataplate draw strategy guide walks through that bust math in detail. But it quietly ignores the two resources that actually make this a puzzle: rerolls and Rewards Doubling charges.

Both of those are daily resources shared across all 3 Rewarded Trials. You do not get fresh rerolls for each hand, and you only have two doubling charges to spread over three fights. That changes everything. The right play on Trial 1 depends on how many rerolls you expect to want on Trials 2 and 3, and whether you have already spent your doubles. A decision that is locally optimal for one hand can be globally wrong for the day.

So the unit we optimize is not “a draw” or even “a Trial.” It is the full day of 3 Rewarded Trials, with the reroll and doubling budgets carried across all of them. Get that framing right and the rest is bookkeeping.

If you have never touched the mode, start with our Trial of Swordmancy mode overview. This post assumes you already know what a Dataplate is and why 10 is the magic number.

Two Numbers: 640K Perfect, 618K Easy

Here is the headline result, side by side.

The optimal number is not a simulation average. It falls straight out of the math: it is the expected value of a pristine, start-of-day state, read directly off the table (more on how that table is built below). The simple number is a simulation average, taken over millions of generated days, because the simple rule is easy enough to just run and measure.

The takeaway is the gap. 22,000 Stock Bills a day is the entire reward for perfect play over good-enough play. Whether that is worth carrying a calculating program into your dailies is a personal call. For us it is not. We will happily forfeit 3.4% to keep the mode a two-minute habit instead of a homework assignment.

The Simple Strategy You Can Run by Hand

This is the one you actually want. It looks only at two things: the sum of your current hand and how many rerolls you have left. That is it. No card counting, no per-state lookup.

In plain words:

• With rerolls in the bank, only a 9 or 10 is good enough. Anything else, you try to improve it: draw another card if the deck still lets you, and if you cannot draw, spend a reroll to start the hand over.
• Once your rerolls are gone, drop your standard to 8. You no longer have a safety net, so an 8 is now worth keeping rather than gambling a draw that might bust you.
• Always activate Rewards Doubling on 2 of your 3 trials. The simplest version is “double the first two.” More on why the order does not actually matter below.

A small nuance the model surfaces: in the rare cases where the cards push you past 10, landing on exactly 20 or 21 is also an acceptable result rather than a wasted hand. In practice you are aiming for 9 or 10 every time; treat the high landings as a quirk of certain draws, not a goal to chase. If you want the full picture of what happens above 10, our 21-point trim breakdown covers the overflow bands.

That rule set, run over millions of simulated days, averages ~618,000 Stock Bills. It is within spitting distance of perfect, and you never once need to look anything up.

A Note on “Rerolls” Versus Forfeits

Quick terminology bridge, because the in-game menus and the community use slightly different words. The “rerolls” in this model are your daily redraw budget: the free mulligans the mode hands you to escape a bad hand. You get a limited number of them per day, and crucially they are shared across all 3 Rewarded Trials, not refreshed per hand. When the model says “reroll if you have any,” it means spend one of those daily redraws to throw the current hand back and try again.

Because the budget is shared, every reroll you burn on Trial 1 is one you will not have on Trial 3. That is exactly why the simple strategy gets more permissive once you are out of them (accepting an 8 instead of holding out for a 9). The model is just formalizing the instinct every experienced player already has: spend your mulligans early when you can afford to be picky, and take what you are given once the safety net is gone.

Why Doubling Order Does Not Matter (Mostly)

You have two Rewards Doubling charges and three trials. You should always spend both; an unused doubling charge is pure waste. The only real question is which two trials get them.

Here is the part that trips people up. Our daily-farming guide recommends scouting Trial 1 with no doubling, then doubling Trials 2 and 3. That advice exists for a human reason: you might not know the freshly rotated deck yet, and you would rather double a hand after you have seen what the deck can do.

But from a pure expected-value standpoint, the order is irrelevant. Every Trial draws from the same known deck. The right-side panel shows you the deck contents at all times, so there is no hidden information to “scout” in the math. Doubling Trial 1 and Trial 2 has the identical expected value to doubling Trials 2 and 3. The simple strategy therefore says “double the first two” purely because it is the easiest rule to remember. If you prefer the scout-first habit, double the last two instead. The Stock Bills come out the same.

The one thing that is not optional: spend both charges, every single day. A doubled clear is the single biggest lever on your daily total, and leaving one in the tank is the most expensive mistake the simple strategy can still make.

The Perfect Strategy: Expected Value, Briefly

Here is the idea without drowning in notation. For any game state, there is an expected value (EV) of Stock Bills you will eventually earn from that point onward, assuming you play optimally from there. A “game state” bundles everything that matters: your current hand, what is left in the deck, how many rerolls remain, how many doubling charges remain, and which Trial you are on.

There are only four kinds of decision you ever make: draw, keep, reroll, double. Each one transforms the current state into a new state (or a distribution of new states, in the case of a random draw). The optimal decision in any state is simply the one that leads to the highest expected value.

That definition is recursive, the EV of a state depends on the EV of the states it leads to, but it bottoms out at clean base cases (a finished Trial pays a known amount; a busted hand pays what it pays). That recursive-with-base-cases shape is the textbook setup for dynamic programming: you compute the EV of every reachable state once, store it in a table, and never recompute it.

Once that table exists, “optimal play” is almost boring. At every decision point, you look up the EV of each available action and take the largest. Repeat until the day’s three trials are done. There is no cleverness left at runtime; all the work is baked into the table.

Reading the Answer Straight Off the Table

The elegant part: because the table holds the EV of every state, including the pristine start-of-day state with full rerolls, full doubling charges, and no trials played yet, you do not need to simulate anything to know the optimal daily income. You just read the EV of that starting state.

For the current card pool, that number is ~640,000 Stock Bills. That is the mathematical ceiling of the mode per day. It wobbles a little depending on the active deck, but not by much.

This is also why we can state the optimal figure as an exact expected value rather than a simulated average. The simple strategy has to be measured by brute-force simulation (run it a few million times, average the result) because it is a fixed rule rather than an optimal policy. The optimal strategy is the policy the table defines, so its value is the table.

Under the Hood: What the Optimal Table Reveals

Building the table for the initial card pool of [5, 5, 5, 8, 7] turned up some genuinely fun structure. Every number here is specific to that pool and will shift when the deck rotates, but the patterns are instructive.

A few things worth chewing on:

• The optimal table never keeps a 5, 6, 7, or 8 if it can reroll. When you have rerolls to spend, those mid-range hands are always worth throwing back. This is the same instinct the simple strategy encodes, just proven exhaustively.
• Doubling is aggressive when you are reroll-rich. With three rerolls still in the bank, the optimal play doubles every time, because you have so many chances left to land a strong hand that the doubled payoff is almost guaranteed to connect. The model only starts hesitating (declining to double a 9) once rerolls get scarce.
• 85% agreement is the whole story. The simple strategy and the optimal table make the identical decision roughly five times out of six. The 22K daily gap lives entirely in that other 15%, the edge cases where the table sees a subtle reroll-budget tradeoff the simple rule cannot.

The Card Pool Rotates Every 3 Days

The obvious worry: these numbers were computed on one specific deck, and the deck reshuffles every three days. Does any of this survive a rotation?

Yes, and here is the key distinction. The math does not change, only the inputs do. The dynamic-programming machinery that builds the EV table does not care what the card values are; it cares only that there is a deck, a set of decisions, and a reward structure. Feed it a new pool and it produces a new table for that pool. The method is rotation-proof even though any single table is not.

We sanity-checked this by generating several different random card pools and running the simple strategy against each. The daily averages stayed in roughly the same neighborhood. As long as future pools do not change drastically in character, there is no reason to expect the ~640K / ~618K story to move much. If a future rotation does something genuinely weird, the numbers will shift, but the approach to finding the new optimum stays exactly the same.

By Player Type: Who Should Bother With Perfect?

The 22K gap means different things to different players. Here is the honest segmentation.

The Daily-Grind Player

You log in, want to clear dailies fast, and move on. Use the simple strategy, full stop. You are capturing ~96.6% of the theoretical maximum for zero mental overhead. The 22K you leave on the table is noise against the time you save. Memorize the four-row table above and never think about it again.

The Optimizer / Theorycrafter

You enjoy the squeeze and do not mind tooling. If you want the literal maximum, you would need the precomputed EV table for the current pool on hand to consult each decision. That is the only way to claim the full ~640K. Whether the extra 3.4% justifies running a lookup every draw is your call, but at least now you know the exact size of the prize.

The Returning / Casual Player

You are not even sure you will run all three trials every day. Forget the gap entirely; the bigger lever for you is simply using both doubling charges and not letting Rewarded attempts go unused. Missing a doubled trial or skipping a day costs far more than the 22K optimization gap ever will.

Common Mistakes That Leak Stock Bills

The optimization gap is small. These mistakes are not, and they are what actually cost most players real income.

• Leaving a doubling charge unspent. This is the single most expensive error. Two charges, three trials, spend both every day, no exceptions.
• Hoarding rerolls you never use. Rerolls do not roll over. An unspent reroll at the end of the day was a free improvement you declined. The simple strategy spends them naturally; do not override it to “save” them.
• Keeping a 5, 6, 7, or 8 while you still have rerolls. The optimal table rerolls these 100% of the time. If you are holding them out of caution, you are leaving EV on the table.
• Refusing to lower your standard once rerolls are gone. With no safety net, an 8 is a keep. Holding out for a 9 you can no longer safely chase is how you bust into nothing.
• Treating each trial as a separate game. Rerolls and doubles are a shared daily budget. Spend them with the whole day in mind, not one hand at a time.
• Drawing past 10 on purpose for farming. Nothing in either strategy ever wants you above 10 in Rewarded Mode. The lone exception is the cosmetic medal trim, covered separately.

What Would Change This Math

The model is only as good as its assumptions. A few things could move the numbers, and they are worth watching as Version 1.3 matures.

• A drastically different card pool. We tested several random pools and the figures held, but a rotation that introduces wildly different card values or deck sizes would produce a new table and potentially new headline numbers. The method survives; the specific 640K/618K might not.
• The exact Stock Bills reward curve. The per-result payout table is still not fully published. Our EV figures rest on the reward structure as currently understood; if a patch retunes how payouts scale, every number here scales with it.
• A change to reroll or doubling counts. If a Trial Arena upgrade or a patch changes how many rerolls or doubling charges you get per day, the shared-budget math shifts and both strategies would want re-deriving.
• A graduated overflow penalty. If overflow above 10 ever becomes a partial cut rather than the current hard cliff, the value of edge-case high landings (the 20/21 nuance) would change, and aggressive draws near the edge would get more defensible.

Final Read

Trial of Swordmancy looks like a luck mode and is actually an arithmetic one, and the arithmetic has a clean answer. Played perfectly, it is worth about 640,000 Wuling Stock Bills a day. Played with a rule you can recite from memory, target 9 or 10, draw-then-reroll otherwise, accept 8 when the rerolls run out, and always spend both doubling charges, it is worth about 618,000. The 3.4% you give up by not carrying a calculator is, for the vast majority of players, the best trade in the mode.

If you take one thing away, make it the framing: you are optimizing a whole day, not a single hand. Rerolls and doubles are shared resources, spend them like it, and the rest takes care of itself. Watch the next deck rotation, re-check if a patch touches the reward curve, and otherwise enjoy one of the most reliable Stock Bill faucets in 1.3.

Frequently Asked Questions

How much is Trial of Swordmancy worth per day? Played optimally, about 640,000 Wuling Stock Bills per day across the 3 Rewarded Trials. With a simple by-hand strategy, about 618,000. Both figures depend on the active card pool but hold roughly steady across rotations we tested.

What is the simplest strategy that still gets most of the value? Keep a hand of 9 or 10. Otherwise draw if you can, or reroll if you have rerolls left. Once you are out of rerolls, lower your standard and also keep an 8. Always activate Rewards Doubling on two of your three trials. That averages ~618,000 Stock Bills, about 96.6% of perfect.

Why does the simple strategy accept an 8 only when out of rerolls? While you still have rerolls, you can afford to be picky and chase a 9 or 10, because a reroll lets you escape a mediocre hand for free. Once the rerolls are gone you have no safety net, so a guaranteed 8 beats gambling a draw that might bust.

Does it matter which trials I double? Not for expected value. Every trial draws from the same visible deck, so doubling the first two and doubling the last two have identical math. “Double the first two” is just the easiest rule to remember. The only real mistake is leaving a charge unspent.

Are rerolls and doubling charges per trial or per day? Per day, shared across all 3 Rewarded Trials. This is the whole reason the mode is a puzzle: a reroll spent on Trial 1 is unavailable for Trial 3, so you have to budget across the entire day.

How is the “perfect” number actually calculated? By modeling every reachable game state, assigning each an expected value, and using dynamic programming to fill a table of those values. Optimal play is then just “always pick the highest-EV action.” The daily maximum is read directly off the EV of the fresh start-of-day state.

How many game states are there? For the initial card pool of [5, 5, 5, 8, 7], there are 10,980 reachable game states. A different pool produces a different count.

How often does the simple strategy agree with the optimal one? About 85% of the time on the current pool. The remaining 15% (subtle reroll-budget tradeoffs the simple rule cannot see) is where the entire ~22,000-per-day gap comes from.

Will these numbers still be right after the deck rotates? The method is rotation-proof; any single table is not. The card pool reshuffles every 3 days, which changes the inputs but not the math. Testing several random pools kept the daily averages in the same range, so unless a rotation is drastic, expect the figures to hold.

Is it worth running a tool to play perfectly? Only if you genuinely enjoy the optimization. The gap is ~22,000 Stock Bills a day, roughly 3.4%. For almost everyone the simple rule is the better trade: nearly all the reward, none of the homework.