Stokastic Contest Generator: What It Does & When To Re-Run
By Jake Hari
July 13, 2026
Stokastic Contest Generator: What It Does & When To Re-Run
The question comes in almost every week, usually some version of this: "Do we need to refresh the Contest Generator when you guys get projection and ownership updates?" A fair question, because nobody wants to redo an hour of work every time a number moves.
The honest answer has two halves, and both of them fall out of one idea most people skip past: what the Contest Generator is actually building when you hit run. Once you see it as a field, not a button, the re-run question answers itself, and so does the order you should do everything in. There is also one lever people reach for right inside the generator that does the exact opposite of what they think it does. I'll get to that one, because it can quietly work against what you intended.
Watch: How The Contest Generator Builds Your Field
Before we get into the theory, here is the tool in motion. In our NFL DFS Contest Generator walkthrough we run the build, sort the output by projection and ownership rank, and let the sim grade every lineup by simulated ROI against that field, which is the same general flow every sport uses.
What The Stokastic Contest Generator Actually Does
Here is the whole thing in one line: the Contest Generator builds a realistic field of opponents, then the Stokastic Sims grade your lineups against it.
When you run it, the tool constructs thousands of opponent lineups shaped like the contest you are entering, matched to its size, its buy-in, and how chalky that kind of contest tends to be. You give it two inputs, pool size and contest archetype, and it fills out the field. Then the sim plays the slate out over and over and scores every one of your builds by simulated ROI against that specific field. This is the step that turns a page of raw projections into a contest-aware decision. Projections tell you who is good. The generated field tells you who is good relative to what everyone else is going to do, which is what actually wins tournaments.
That framing also unlocks a use most people miss. For cash games and head-to-heads, staff run the Contest Generator, sort the output by raw projection, and take the highest-projected lineups off the top. Same tool, different read: in cash you let raw projection lead, and ownership games matter far less than they do in a GPP.
The one idea to hold onto: the generator builds the field, and the Sims grade your lineups against it. Everything else on the page is either describing that field or reacting to it.
The Field Is The Whole Point
Worth sitting on that word, field, because it does all the work. Without the generator, projections grade in a vacuum. A lineup looks strong because its players score points. Fine, but everyone can see those same projections, so the "obvious" build is also the most owned build, and being right along with 40% of the field wins you very little in a top-heavy GPP.
The generated field is what prices that in. When the sim knows what the room is likely to roster, it can tell you which of your builds actually separate from the crowd when they hit. A leverage play only reads as leverage against a field that is fading it. Build the room first, and every ROI number after that describes real competition instead of a spreadsheet. That single shift, from grading players to grading lineups-inside-a-field, is why the re-run rules below work the way they do.
Do You Need To Re-Run When Projections Update?
Now the question we started with. The answer is: it depends entirely on what changed, and the dividing line is clean once you know where it sits.
If the input that feeds the field changed, you re-run. If you only nudged your own preferences after the field was built, you don't. Our point projections start updating early in the day and keep moving through lock as news breaks, so this comes up constantly. Staff coach the split this way:
| What Changed | Re-run? | Why |
|---|---|---|
| Point Or Ownership Projections Updated (Injury, Lineup, Weather, Odds) | Yes | New numbers change the field's makeup, and your lineups get re-graded against it |
| You Changed The Pool Size Or Contest Archetype | Yes | You changed the room itself, so the field regenerates from scratch |
| You Added ROI Boosts To Lean Into Or Off A Player | No | A preference applied on top of the built field, not an input to it |
| You Set Exposure Caps | No | A hard ceiling on your own pool; the field is unchanged |
| You Filtered To 5-Man Stacks Or Trimmed The Pool | No | You are slicing your existing lineups, not rebuilding the field |
The row that matters most is the boosts-and-caps row at the bottom, because it is the one that saves your evening. Those are the adjustments you make at the end, once you are close to final, so you are not redoing the same tweaks after every projection refresh. Do them early and every required re-run has you reapplying the same adjustments. That is the trap the original question was really about: "I need to go in and make the same adjustments all over again, correct?" You don't, as long as you do them in the right order.
The one-line rule: point or ownership projection changes, plus any pool-size or contest-archetype change, force a full re-run; ROI boosts, exposure caps, and stack filters do not. If you want to run it on tonight's board, the Stokastic Contest Generator lives inside Stokastic+, and code SIMFIELD10 takes 10% off your first Data + Sims or Sims MAX payment: build your first field.
The Order That Saves You An Hour: Heavy Inputs First, Light Dials Last
That table is really a workflow in disguise. Sequence your build so the expensive, re-run-triggering decisions happen first and the free ones happen last:
- Lock the room first. Pick your pool size and contest archetype to match the contest you are actually entering. Set pool size to mimic the real field, because realism beats raw size. For a small high-stakes contest of a few hundred entries, staff run a pool around 500 to get enough opponents to grade against; for the huge cheap fields, use the bigger presets. Archetype has two levers behind it, buy-in and field size, so a low-stakes contest like a $0.25 mini-max builds a softer, chalkier field and a $55 high-stakes contest, smaller and higher buy-in, builds a sharper one. Get this right before anything else, because changing it later regenerates everything.
- Generate the field and let projections settle. Run it, and if a meaningful projection update lands, re-run. Re-running is cheap here because you haven't layered any personal tweaks yet.
- Grade for your payout. Set percent-to-first to the contest's real top-prize share. A practical read: if a GPP pays $200,000 to first out of an $800,000 prize pool, that is a 25% percent-to-first. This changes how lineups are ranked, not what field gets generated, so a top-heavy GPP ranks contrarian ceiling builds higher and a flatter payout favors safer ones.
- Now apply your light dials. ROI boosts, exposure caps, and stack filters. These are last on purpose, because none of them force a re-run, and saving them for the end means an early-slate projection change won't send you back to redo them. Set uniques last of all, reading the min and average ROI at each count off the Portfolio Overview at the bottom of the board.
This is the exact "do it last" discipline we walk through in our PGA Sims exposure Short: the exposure work is the finishing move, not the opening one.
The Lever That Fools People: The Up-and-Down Arrows
Now the one I promised. Inside the Contest Generator, next to each player, are up and down arrows. Naturally, you assume clicking up gets you more of a player and clicking down gets you less. It does not.
As our PGA Sims exposure Short breaks down, those arrows change the player's projected ownership, which shapes the field, not your own exposure to that player. You are telling the sim, "the room will roster this guy more (or less)," which is useful when you genuinely disagree with where his ownership is heading. But if your goal is simply to end up with more or less of a player in your lineups, the arrows are the wrong tool, and a lot of sharp players quietly fade or force the opposite of what they intended by using them.
Callback to the field: the arrows edit the room, and your ROI boosts edit you. To actually move your own exposure, use ROI boosts (up to lean in, down to fade), or an exposure cap when you just need a hard ceiling on one name. Reserve a real projection change for the cases where you truly disagree with our median on a player, since that one does force a re-run. Here is what each control actually moves:
| Control | What it actually moves |
|---|---|
| Up/Down Arrows | The player's projected ownership, which shapes the field |
| ROI Boost | Your own exposure to a player (lean in or fade) |
| Exposure Cap | A hard ceiling on your own pool |
| Change A Projection | Your disagreement with our median (a field input) |
The row people trip over is the top one. Clicking those arrows feels like it should get you more of a guy, but it is really editing the field's projected ownership, not your slice of it. Knowing which control moves the field versus which moves your pool is the difference between a build that does what you meant and one that fights you all night.
A Worked Run: Modeling A $20 MLB GPP
Let me put the whole order on one contest. Say I am entering a mid-size $20 MLB GPP, roughly 15,000 entries, and I want 20 lineups.
- I set the room first. Pool size to the closest preset that behaves like a 15,000-entry field, archetype to the closest preset for that buy-in and field size. Because ownership matters more in MLB than in any other sport, something staff stress constantly, getting the field's sharpness right matters more here than it would in NFL.
- I generate and watch the clock. If the early projections are stable I move on. If a starting pitcher gets scratched an hour later, I re-run, because that is a projection change and the whole field has to regrade around it.
- I set percent-to-first to the contest's real top-prize share. Because this one is a top-heavy GPP, a higher percent-to-first pushes ceiling builds up the rankings where they belong.
- Only now do I tune. I ROI-boost the two hitters whose range I think is wider than our median, cap one chalky bat I don't want 60% exposure to, and set uniques off the Portfolio Overview at the bottom, which lays out min and average ROI at each uniques count for tonight's slate. That read might look like roughly 14.8% minimum ROI and 41% average at 3 uniques, with the minimum falling toward negative at 4, in which case I take 3.
Total re-runs in this example: one, driven by the pitcher scratch, and none caused by my own tweaks. All of that is the payoff of doing the heavy inputs first. The simulated ROI I read at the end describes a real $20 MLB field, and every dial I touched after the field was built stayed exactly where I put it.
In Summary
The Contest Generator is not a button that spits out lineups. It is the step that builds a realistic opponent field so your lineups get graded against real competition instead of against no one. Once you hold that, the re-run question stops being confusing: change what feeds the field (projections, pool size, archetype) and you re-run; change your own preferences on top of it (boosts, caps, filters) and you don't. Set the heavy inputs first, tune the light dials last, and remember that the up-and-down arrows move the room, not you. Do it in that order and you stop redoing work every time a number moves, and the ROI you are reading actually means something.
Run It On Tonight's Slate
The Stokastic Contest Generator builds your field, simulates your lineups against it over and over, and surfaces the builds with the best simulated ROI, with the pool-size, archetype, and percent-to-first controls covered above. It lives inside Stokastic+, and code SIMFIELD10 takes 10% off your first Data + Sims or Sims MAX payment: build your first field with 10% off. New to the Sims? Start free with the DFS Sims and pull a live slate from the MLB DataHub. If the underlying build order is still new, our MLB Sims walk-through and NBA Sims guide cover it end to end, and the pool-size guide goes deeper on sizing the field.
Frequently Asked Questions
What does the Stokastic Contest Generator actually do? It builds a realistic field of opponent lineups shaped like the contest you are entering (its size, buy-in, and chalkiness), so the Sims can grade your lineups by simulated ROI against real competition rather than in isolation. You set pool size and archetype, it generates the field, and the sim scores every build against it.
Do I need to re-run the sim when projections or ownership update? Yes, when the input that feeds the field changes. Point and ownership projection updates, and any change to pool size or contest archetype, regenerate the field, so you re-run. Your own adjustments, ROI boosts, exposure caps, and stack filters, do not require a re-run.
When should I make my ROI boosts and exposure caps, then? Last, once you are close to final. They sit on top of the built field rather than feeding it, so doing them early just means an early-slate projection change erases them and you redo the work. Build the field first, then tune these at the end.
Why don't the up and down arrows change my exposure? Because those arrows change the player's projected ownership, which shapes the simulated field, not your own exposure to that player. To move your own exposure, use ROI boosts (up to lean in, down to fade) or an exposure cap for a hard ceiling.
Can I use the Contest Generator for cash games? Yes. Run it, sort the output by raw projection, and take the highest-projected lineups off the top. In cash you let projection lead, and ownership games matter far less than they do in a GPP.
Stokastic+ (Data + Sims / Sims MAX): build your field and grade lineups inside the Stokastic Contest Generator
Use code SIMFIELD10
Get Started