Fixed or progressive, and the three things that change
The two kinds of jackpot are not two sizes of the same prize: they are paid out of different money, and the choice between them moves the return, the variance and the dependence on strangers in other countries.
Two prizes paid out of different money
Put a "fixed" jackpot and a "progressive" jackpot side by side and the trade will tell you that one is small and the other large (and will usually leave it there), which is true and is also the least interesting thing about either of them.
The difference that matters is where the money comes from. A fixed prize is paid out of the "payout model" of the game in front of the player (the phrase is a database's, and a useful one), like any other combination on the reels. A progressive prize is paid out of a deduction levied on every stake placed across a whole "network" (the set of games sharing one "pool"). Those stakes include the money of people in other countries who will never meet him (and the pool is denominated in more than one currency).
From that single distinction everything else follows, and it follows in three directions at once: the return a player is paid, the shape of the distribution he is paid it from, and whether his prize depends on strangers (it does, on one of the two, and the dependence is not small).
None of the three appears on the comparison tables this trade publishes, which we find odd, because all three can be read off figures the suppliers themselves print.
Our table below sets out the comparison as we would draw it, and the last of its seven rows is the one to read first. The rest of the page is those three consequences, in the order in which they cost a player money (the order is deliberate).
The comparison, line by line
| Fixed jackpot | Progressive jackpot | Must-drop | |
|---|---|---|---|
| Where the prize comes from | the game's own payout maths | a slice of every stake on the network | a slice of the stakes, with a threshold |
| Does it grow? | never | yes, until it falls | yes, but it must fall by a threshold or a time |
| Does it reset? | not applicable | yes, to the seed | yes, to the seed |
| Does it feed a shared pot? | no | yes | yes |
| Does it depend on other players? | no | yes | yes |
| Largest win | the multiplier declared | unlimited in theory | bounded by the threshold |
| What it costs you in RTP | nothing | 0.99 to 7.60 points, depending on the network | the network's share |
The first consequence: the return stays in the game
A fixed jackpot takes nothing for a shared "pot", so the whole of the declared return remains inside the mathematics of the game, and that is not a theoretical nicety but the largest single effect we have been able to measure.
The cleanest demonstration is a game issued both ways, and such a game exists exactly once in this trade. Nobody has repeated the experiment. The Dream Drop version of Money Train Origins is published at 94.00 per cent and the version without the jackpot levels at 96.11, so attaching a pot to that set of reels costs 2.11 points of "return to player" (the figures are the supplier's own, for both configurations).
The same principle, applied to an older network, produces a far uglier number. Mega Moolah pays 88.12 per cent from its reels on the original title and 81.63 on the "Megaways" version, against roughly 96 per cent for a slot of comparable vintage with no pot attached (Wolf Gold, from 2017, is published at exactly 96.00). Between 5.30 and fourteen points of return have therefore left the game before a single spin is made.
Fixed jackpot games sit where you would expect them to sit once that is understood. Fountain of Youth is published at 96.81 per cent with a fixed maximum of 800 times the stake (Playtech, December 2015); Jokerizer stands at 98 per cent with a fixed "mystery win" of 6,000 coins (Yggdrasil, November 2014). Neither of them gives anything up to a pool, and neither has a "seed" to restart from.
Where both kinds sit side by side
Alderney AGCC Category 1, Baytree (Alderney) Ltd · Kahnawake, Baytree Interactive Ltd
18+ · Sponsored link · Terms on the operator's site
The third column, where the prize has a ceiling
The table carries a third column, because a "must-drop" jackpot behaves like neither of the other two and is usually filed with the progressives by mistake.
It grows, as a progressive does, out of a slice of the stakes. But it has to fall by a stated time or before a stated value, and that ceiling makes the prize "bounded" in a way a true progressive is not. Jackpot King carries two "capped pots", seeded at about £500 and £5,000 and ceilinged at about £3,500 and £35,000. The licensed operators carrying that network are the source for all four of those figures (two of them, and they disagree about which pot carries which name).
The Daily Drop network of Red Tiger works the same way with a clock instead of a ceiling: one level falls every hour and the other every day (the hourly prizes are described as typically three or four figures), while the daily one is put by a licensed operator at typically between £40,000 and £50,000. Even the largest of the modern pools has an upper limit, since the Dream Drop Mega is capped at €10 million (it climbs by €500,000 at every second win until it reaches that "ceiling").
The conclusion you should draw is that "unlimited in theory" belongs in only one of the three columns, and that, in our reading, the bounded prize is the one which actually arrives (every hour, in one case). We say more about who uses the term, and who refuses to, on what must-drop really means.
The second consequence: the median and the mean
This is the consequence the trade never states, and it is the one a reader should carry away from the page, because it explains why the published returns of progressive games feel wrong in practice.
A fixed jackpot of 1,000 times the stake is won with a probability which can be measured, and it pays an amount known in advance (the stake multiplied by the figure on the paytable, and nothing else). A top "progressive" is the opposite on both counts, being an event of extremely low probability with an enormous payout, and the two together do strange things to an average.
Put it in terms of distributions and the point is plain. The "median" result of a player (the outcome in the middle, with half of all sessions above it and half below) is identical whether the jackpot is attached to the game or not; what the jackpot changes is the "mean", and it changes the mean by the whole weight of a prize which will not arrive.
The arithmetic of how rarely it arrives is on how often these jackpots fall, and the headline is that the Mega Moolah top level has averaged about five and a half payments a year (9 weeks 4 days between wins, on the trackers) across every casino carrying it at once. A reader can work out his own share of five and a half!
So for the ordinary player, and the ordinary player is practically everybody, the return paid into the pot is return lost. It is not withheld, it is not stolen, and it is not a scandal; it is simply spent on a "lottery ticket" priced into every spin, whether or not the man paying for it wanted one.
The third consequence: nobody else is involved
The value of a fixed prize does not depend on how many people are playing, where they are playing, or which casino they happened to choose, and that independence turns out to matter a good deal more than it sounds.
The trouble it avoids is visible on the oldest pooled games, which are the subject of the first generation. King Cashalot has a pot and no players; the counter stands above $2.2 million while the last win was 9 years 8 weeks ago (the "average interval" for that level is 2 years 21 weeks), so the prize exists on a screen rather than in any practical sense.
There is a subtler version of the same dependence on the NetEnt pools, and it is verified on the supplier's own pages. On Mega Fortune two of the three levels are "local" to the operator and only the top one is "global" (the Rapid and the Major are the local pair), which is why the trackers list the Mega Fortune "Major" in nine different versions, one per casino, ranging from €12,138 to €47,633 (the "Mega" above them is a single global figure).
The consequence for a reader is worth stating flatly: on those games the pot he sees depends on the casino he happens to be in rather than on the game he has chosen, and no page on the game itself will tell him which of the nine versions he is looking at.
A fixed jackpot has one further property that is easy to overlook: because it is a multiple of the stake, the prize scales with the bet and with nothing else, so the figure you are chasing is one you can work out before you start. On a progressive the equivalent calculation needs the probability of the level, which no supplier publishes, as when a pot is worth chasing works out at length.
Which to choose, and on what ground
We are not going to tell a reader which kind of jackpot to play, because the choice depends on what he wants from the session rather than on any figure we can produce.
What we will say is that the two products answer different questions, and that the usual framing confuses them. If the point of the session is to play a slot at the best available price, the fixed jackpot wins on every figure in our document, and it wins by between two and fourteen points of return. If the point is to buy a small chance of a "life-changing" sum, then the price of that chance is those same points, and our tables put it at two of them on the newest network and fourteen on the oldest, and it is worth knowing what the chance costs rather than being told that it is free.
The one position we would reject is the one most often published, which is that a progressive jackpot game is simply a slot with an extra prize bolted on at no charge. That bolt has a price, and the price is printed on the supplier's own page; we have yet to find a comparison table anywhere which puts the two figures in the same row!
Three casinos with both families
VideoslotsMGA/CRP/258/2014, Videoslots LimitedVisit
Dove SlotsGambling Commission account 39175, Jumpman Gaming Ltd, domain Active in the public registerVisit
Express WinsGambling Commission account 39175, Jumpman Gaming Ltd, white label domain in the public registerVisit
18+ · Sponsored link · Terms on the operator's site
Questions
Is a fixed or a progressive jackpot better?
For the typical session the fixed one is better, because nothing is deducted from the game to feed a pool and the whole of the return stays inside the payout mathematics. The progressive is better only for the player who actually collects the top level, and the measured intervals between those wins run from weeks to years.
Which kind has the higher return to player?
The fixed one, all else being equal. The cleanest comparison in the trade is a single game issued both ways: Money Train Origins returns 96.11 per cent plain and 94.00 with the Dream Drop levels attached, so the pot costs 2.11 points. The same principle explains why Mega Moolah sits at 88.12 per cent at the base.
Why does the average win look so good on a progressive?
Because the average is lifted by an event almost nobody will experience. The median result of a player is identical with or without the jackpot attached, while the mean carries the weight of a prize that falls a handful of times a year across a whole network. For the ordinary player the return paid into the pot is return lost.
Does a fixed jackpot depend on other players?
No, and that is its structural advantage. The value of a fixed prize does not depend on the liquidity of a network, which is exactly the trouble with the oldest pooled games: King Cashalot has a pot but no players, and a counter nobody is playing for is a prize on a screen rather than in practice.
