Accept dead spins, keep return out of the tail, narrow the volatility band: a different answer that holds
Per-game samples of dozens of JILI slots: low, scattered hit rates, a product line split between no-free-game and high-frequency titles, max odds capped in the hundreds and a narrow volatility band. The real cost of dropping free games, and which players this design is built for.
Look at JILI after PG and the first impression is usually "conservative": low caps, free games that may or may not exist, and none of that hair-trigger tension on screen. Yet JILI's product line has held up, and expanded fast. This article does not argue who is better. It unpacks one thing: the three math choices JILI makes that let it retain players without leaning on jackpots.
The jackpot is the most expensive component. For the tail to really carry return, the base game has to be thin, the free game thick, volatility high, and every part of the game has to serve it. JILI chooses not to pay that bill: return stays in the middle, what a player collects over any stretch is steady and predictable, and "never too bad" replaces "maybe very good".
This article is based on per-game samples of tens of thousands of spins across dozens of JILI titles, with hit rate, free-game trigger interval, max odds, odds distribution and volatility tabulated for each. As before, no absolute RTP values, only the shape of the distribution. Each section covers where a choice sits and what it costs, and the end covers what kind of player the three choices are betting on together.
PG pins its hit rate at one in three spins. JILI does not. Close to half of JILI's titles need four or more spins per hit, and the spread between titles is wide: some hit every three spins, others only every five or six. That is not poor tuning. It is the first choice: accept dead spins.
What a low hit rate saves is return. Every losing spin is a stake that can be handed to a later winning spin. Where JILI puts that saved return comes later; first, the cost: the first ten spins feel drier, and players need longer to sense a rhythm.
Sub-stake wins take nearly a fifth of spins in PG; in JILI they are markedly rarer. So a JILI hit usually returns more than the stake. That is the other side of the low hit rate: hit less, but make it count. The definition of "a win" differs as a result: in PG it is "something happened", in JILI it is "I got something".
In PG the free game is a core component every title has, with a locked cadence. In JILI it is optional. Spread the line out and it splits three ways: titles with no free game at all, titles that trigger every few dozen spins, and titles close to the mainstream cadence.
Roughly a quarter of JILI titles have no free game. These are pure base-game products: no trigger, no waiting, no "once you're in" segment. The player experiences the same distribution from the first spin to the last. They are the fastest to build, the simplest to verify, and the flattest to play.
A further tenth or so go the other way: a free game or similar small bonus segment triggers every few dozen spins. At that frequency each one cannot pay much, and its role is closer to PG's small wins: a "chapter break" outside the base game every short while.
The remaining titles follow what most people think of as slot cadence: a trigger every hundred-odd spins, with the free game carrying part of the return. But even on this line the share of return JILI puts inside the free game is more conservative than PG's, and the base game is not deliberately thinned.
This is the most visible choice and the one most often misread as "conservative". The vast majority of JILI titles cap max odds within the hundreds; titles reaching the thousands are rare. But a low cap does not mean less return. It only decides which band the return sits in.
In nine out of ten titles the max odds land within a few hundred times the stake. That means the apex of the odds distribution is close: from one times to tens to hundreds, every band carries real probability, and a player can see most of the distribution in a single evening rather than only its floor.
The key fact again is allocation: in JILI, wins above a hundred times the stake take only a small slice of total return, under a tenth. Where PG parks nearly a quarter in the tail, JILI keeps almost all of it in the middle. The player's curve is far flatter as a result: no night where the stake multiplies by thousands, but no night of hundreds of spins with nothing either.
PG deliberately spreads volatility so each title differs; JILI gathers it in. Seven in ten JILI titles sit in the same volatility range, which means whichever game a player picks, the evening's ups and downs feel about the same.
A narrow band follows naturally from the low cap: with no weight in the tail, the distribution cannot stretch. But it is also a product-line decision. Players moving between JILI titles do not need to re-adapt, so switching costs almost nothing. That matters for a strategy that retains players through volume and frequent releases.
As with PG, split the JILI sample into the higher-hit-rate half and the lower half and the median volatility barely differs. Hit rate decides how often you win; volatility decides how heavy the tail is; in JILI they are independent dials as well. JILI's scattered hit rate and clustered volatility are the proof that they are not the same dial.
The pure base-game line is JILI's most distinctive, and the one most teams want to imitate: fast to build, fast to verify, fast to release. But dropping the free game is not free, and the math shows a clear cost.
Separate titles with a free game from titles without, and the pure base-game group has a markedly lower hit rate. The reason is that the return has nowhere to go: no free game to absorb it, no appetite to raise the cap, so the only room left is fewer winning spins with each win thicker. The pure base-game feel is therefore "drier, but a hit means more".
Many assume that without a free game there are no big swings, so volatility must be lower. In the sample the two groups' volatility is nearly identical. Volatility comes from the shape of the odds distribution; the free game only moves part of that distribution into a separate segment, and moving it back does not change the shape. To make a game calmer, tune the tail, not the free game.
Put the three choices side by side: accept dead spins, keep return out of the tail, narrow the volatility band. They point at one kind of player: not someone who sits down for that one jackpot, but someone who wants roughly the same amount over every stretch, can switch titles without re-adapting, and can stop or come back at any moment.
This is neither worse nor better than PG; it is another complete answer. The only question is which player your game wants, and whether your three choices point the same way.
If your line follows the JILI route, or you want to drop the free game and ship a pure base-game title, run your sample first and answer three questions:
Answer those three and you know whether your products are really on the same road. That is also the first thing we do for clients: run the distribution, look at the shape, then decide whether to tune.
Further reading: for a completely different answer, see PG's four math components. For how hit rate and odds distributions are designed mathematically, see natural probability design; for what to verify once the distribution is measured, see QA math verification; and for why hit rate and volatility are two independent dials and how each is tuned, see the control system.