There have been many articles published in recent months, and indeed over the last
several years, about slot hold, with 📈 many at least partially attributing the industry’s
woes to rising hold.
“Hold” is the expected amount of each wager that the 📈 slot machine
“holds” over time. A slot machine with 5 percent hold is expected to produceR$5 revenue
for the casino 📈 for everyR$100 in wagers. The same machine may be described as having 95
percent “RTP” or return-to-player.
Over the past two 📈 decades, we’ve seen average slot
holds rise considerably, largely due to the proliferation of higher-hold penny video
slots. See Nevada’s 📈 hold, for example, in the figure below.
To contextualize this
increase, bettingR$1 per spin at eight spins per minute, aR$100 budget 📈 would last on
average 249 minutes at the 5.02 percent hold we saw in 1993, but only 187 minutes at
📈 the 6.7 percent hold we saw in 2024. If we were instead to look at the effect of moving
from 📈 9 percent hold to 10.68 percent hold (the same 1.68-point increase, more
indicative of penny slot hold), time on device 📈 for thatR$100 budget would decrease on
average from 138.9 minutes to 117 minutes, a decline of more than 15 percent 📈 or nearly
22 minutes of play.
Increased hold is decreasing the average time of slot sessions.
This isn’t a controversial viewpoint; 📈 it’s just math—if the machine holds more per
spin, players with a fixed budget necessarily spend less time on machines.
The
📈 question, “Can players ‘feel’ the effect of hold changes?” has been studied by
academics, and they’ve concluded that players cannot. 📈 Industry experts have countered
this research by arguing that increased hold is nonetheless degrading the experience of
the slot player, 📈 for example by decreasing time on device. These critics argue that a
player-centric rather than a machine-centric review is necessary.
These 📈 views may seem
irreconcilable, and intuitively, how could players not feel a decrease of 20 minutes on
device? This article 📈 is intended to bring together these views. I’ve spent a lot of
time studying these questions with slot operators, finance 📈 teams, economists, and data
scientists, and as such, have a unique perspective on the problem.
Players Can’t Feel
Hold Changes
Anthony Lucas, 📈 a professor at the University of Nevada, Las Vegas, has
published several articles with various co-authors on how players can’t 📈 “feel” the
effects of a hold change. The evidence has been several-fold, with the most compelling
arguments being that:
A computer 📈 (let alone a human) can’t accurately distinguish
between two different return-to-player (RTP) settings in a number of observations that
mimic 📈 a human slot session on a reel machine; and,
Lucas and his co-authors have run a
number of live experiments in 📈 casinos with side-by-side machines of the same theme, one
with low hold and one with high hold, and observed empirically 📈 that the high-hold
machines perform better financially.
Both of these findings are sound. That is, in
nearly all cases, a player 📈 cannot accurately tell the difference if a machine’s hold is
changed, sometimes even dramatically. Though, I should note there are 📈 several ways of
going from hold A to hold B, and some of these may be easier for players to 📈 “feel” than
others. We’ll discuss why hold changes are not all created equal later on.
In the cases
studied, using variants 📈 of actual paytables from reel slot machines, Lucas and
co-author A.K. Singh argue (correctly) that players can’t possibly tell the 📈 difference
between hold settings, because even computers cannot tell the difference with any
degree of certainty across 500 spins (representing 📈 approximately an hour of continuous
slot play) or even across longer sessions. Certainly, players may guess, as can
computers, but 📈 they are wrong nearly as often as they are right.
Players can’t “feel”
the time decrease because slot machine outcomes are 📈 volatile. As a wise man once said,
“you never know what you’re gonna get.”
This effect is born out on slot 📈 floors, as
described in Lucas’ recent work with Kate Spilde, where they measure the performance of
high-hold and low-hold versions 📈 of the same slot machines placed next to each other on
the floor, finding that the high-hold machines outperform the 📈 low-hold machines
empirically.
My own experiences echo these results. In nearly every example I’ve ever
seen—including the dozens of tests I’ve 📈 run with slot teams on real, live casino
floors—the higher-hold machine of a pair of like machines generates higher win 📈 than the
lower-hold machine. That is, slot patrons don’t shift their play to the lower-hold
device. Players truly can’t feel 📈 hold changes.
How Can Player Behavior Be Impacted If a
Player Can’t Feel Hold Changes?
It would be easy to conclude, as 📈 Lucas does in several
of his articles, that casinos can perhaps increase their revenues by increasing slot
hold. But upon 📈 reflection this is far from clear.
First, the side-by-side machine
comparison fails to ask about the rest of the slot floor, 📈 about the rest of the
player’s wallet. Is the increased financial performance of the high-hold machine simply
displaced win from 📈 the rest of the floor? Or, asked differently, do the players that
lose less on the lower-hold machine exhibit increased 📈 play elsewhere on the floor? In
other words, do players generally lose the same amount on the visit, but those
📈 experiencing lower hold just lose slower and on more machines, getting more time in the
casino?
Second, even if we were 📈 able to measure the overall wallet impact of
experiencing lower or higher hold on a single visit, how does this 📈 experience impact
likelihood to return, or frequency of visitation? Is it possible that a lower-hold
experience today means that a 📈 player will return to the casino sooner, producing the
same amount of revenue or more over more visits?
As an extreme 📈 thought experiment,
consider that a machine that holds 100 percent—never returning a dime to a player—will
perform financially better in 📈 the short term, for some definition of “short term.” But
as a player, if you walked into a casino withR$100 📈 and lost on every spin of your
machine, would you consider yourself unlucky on that trip? Would you hesitate before
📈 returning? How would you feel if it happened again on your next trip?
This thought
experiment—even if 100 percent hold is 📈 extreme—provides a useful way of thinking about
how players can be impacted by hold changes even if they don’t know 📈 that the hold is
higher. Players don’t experience theoretical hold. Players experience the random
sequence of outcomes that the machine 📈 produces in the short amount of time that they
play on the machine. They experience “Did I have a good 📈 time while I was at the
casino?”
This question will have different criteria for different players: How long did
my budget 📈 last me? Did I get to experience fun bonus games on the machine? Did I have
positive staff interactions? Was 📈 my restaurant or valet experience good? And the answer
to “Did I have a good time while I was at 📈 the casino?” influences player behavior
related to return trips: Will the player return, and how soon?
A player who has a 📈 bad
session at low theoretical hold has the same negative experience as a player who has a
bad session at 📈 high theoretical hold. Tying this all together, increased hold leads to
a higher proportion of players experiencing losing sessions, short 📈 sessions, and
therefore, overall negative experiences.
We know that actual loss correlates to overall
experience, and you can validate this with 📈 your own guest survey results. Players who
have “winning experiences” as measured by the duration of play that their budget 📈 allows
or as measured by the experience of low actual hold (including those who win on the
trip) tend to 📈 report better satisfaction with staff interactions, beverage service, and
several other areas of guest experience. And we all believe that 📈 experience matters in
choosing whether entertainment budget should be spent at a casino, and furthermore when
choosing which casino to 📈 visit.
By increasing theoretical hold, even if any individual
player can’t tell that we’ve done so, we increase the number of 📈 players whose random
sequence of slot outcomes leads them to have poor overall experiences at the casino,
and this can 📈 have downstream effects in terms of visitation and spend.
So, What’s an
Operator to Do?
It’s important to stress that I don’t 📈 think there’s a one-size-fits-all
solution to hold changes. For large commercial properties on the Las Vegas Strip, where
revenue is 📈 shifting rapidly to non-gaming predominance, where casual visitors to Las
Vegas have small gambling budgets relative to their overall vacation 📈 budgets, where the
overall trip experience has many components beyond their experience on the casino
floor, and where the time 📈 between trips is lengthy, it may make sense to push hold high
and capture the tourist gambling dollar before the 📈 competitor down the street can
capture it.
Next year, when planning their annual Las Vegas trip, the thought of how
quickly 📈 theirR$100 budget was captured by the slots will be dwarfed by their pool,
dining, nightclub and hotel experiences, and by 📈 the “sin” in Sin City.
In regional
markets, by contrast, casinos may have large segments of patrons visiting upwards of 30
📈 or 40 days per year. Gambling is the main concern at these properties, and markets are
quite competitive, with many 📈 having four or more easily accessible casinos, not to
mention the regional or national destination markets—Las Vegas, Atlantic City, Biloxi,
📈 and so on—that are also competing heavily for these guests. Here, hold is a more subtle
concern.
Casinos in these markets 📈 must carefully assess the impact that hold has on
their businesses, but understanding the tradeoff between short-term financial gain
(take 📈 the money quickly) and long-term business stability (Do we alienate our guests
and cause them to reduce or cease visitation?) 📈 is not an easy task. In contrast to the
Las Vegas market, the gambling experience at regional casinos by-and-large is 📈 the
customer experience, so operators should approach the gambling experience with
caution.
Macro Considerations For a Slot Hold Strategy
In assessing the 📈 impact of a
changing slot hold strategy, we must understand the balance between guests who are
time-constrained, those who will 📈 leave the casino before they’ve exhausted their
gambling budget, and guests who are wallet-constrained, those who will exhaust their
monetary 📈 budget before they exhaust their allotment of time. If a guest is
time-constrained, a reduction in slot hold will reduce 📈 the rate at which they lose (on
average), and the casino will capture less of their gaming budget on their
📈 (fixed-length) trip. In order to make this decision profitable, the casino would need
to increase the visitation of those guests 📈 to compensate for the reduced revenue.
If a
guest is wallet-constrained, however, a reduction in slot hold will simply increase the
📈 time that the guest’s budget lasts, providing more time in the casino and more positive
experiences, i.e., more “bang for 📈 their buck.” Of course, if we increase the duration
of the guest’s wallet too much, the guest may become time-constrained, 📈 and we run the
risk of losing the guest’s available budget. It seems natural to argue that a perfect
balance 📈 would be struck if we could have each guest expend their monetary and time
budgets simultaneously.
Quantifying time-constrained versus wallet-constrained guests
📈 is difficult to do scientifically. But as an example of this thinking, in a market like
Biloxi—where many patrons are 📈 lodgers and as such are a relatively captive
audience—guests are likely more wallet-constrained than time-constrained, and a lower
slot hold 📈 environment may increase player satisfaction (and ultimately visitation,
etc.) while effecting a very limited impact on gaming spend.
And besides making 📈 the
casino experience more fun, which we would hope leads to increased visitation, the
limited revenue loss from the gaming 📈 floor on that visit may be recuperated by retail
and dining outlets, albeit at a different margin.
My own experience in 📈 the Louisiana
and Mississippi markets suggests that Biloxi casinos tend to provide
richer-than-typical free-play offers. Increased free play and reduced 📈 slot hold have a
similar effect, increasing time on device, which is low cost to the casino so long as
📈 the patrons are wallet-constrained and not time-constrained.
Of course, many of these
arguments can be applied to Las Vegas as well, 📈 but Las Vegas visitors are more
time-constrained than one might imagine. The allure of other amenities, or even other
casino 📈 properties, limits the amount of vacation time allocated to gambling in any one
location. And with the proliferation of regional 📈 gaming, this makes sense. Most
visitors come to Las Vegas for the party, for the pools, for the weather, for 📈 the food.
The gambling is a nice-to-have, as opposed to Biloxi, where the gambling may be a
primary focus of 📈 the visit.
Additional considerations when developing an overall hold
strategy for a property may include:
The floor’s utilization: Higher utilization
suggests a 📈 higher hold strategy, as reduction of time on device can alleviate any
periods of prohibitive utilization, which itself degrades the 📈 guest experience.
Higher
utilization suggests a higher hold strategy, as reduction of time on device can
alleviate any periods of prohibitive 📈 utilization, which itself degrades the guest
experience. The quality and diversity of a property’s amenity set: The more
opportunities a 📈 property has to provide great experiences to a guest suggests a higher
hold strategy, as the slot experience may contribute 📈 less to the overall guest
experience.
The more opportunities a property has to provide great experiences to a
guest suggests a 📈 higher hold strategy, as the slot experience may contribute less to
the overall guest experience. The frequency of visitation of 📈 the patron database:
Higher-frequency properties might consider a lower-hold approach, since there is a high
dependence on return visitation.
Higher-frequency properties 📈 might consider a
lower-hold approach, since there is a high dependence on return visitation. The
competitiveness of the local market: 📈 Properties in highly competitive markets might
consider a lower-hold approach as a way to improve guest experience.
Micro
Considerations For a 📈 Slot Hold Strategy
I mentioned before that moving from one hold to
another isn’t a universal concept. That is because there 📈 are many ways to change a pay
table. As an example, consider the following simple mock game:
In this game, we
📈 wagerR$1, and we either lose ourR$1, or we winR$1,R$2,R$10, orR$10,000. The bonus game
that produces a win ofR$10 is triggered 📈 on average every 40 spins, and theR$10,000
jackpot is triggered on average every 100,000 spins.
Now let’s say our aim is 📈 to
increase the hold to 13 percent. One way to do this is to decrease the frequency of the
bonus 📈 game to 1 in 50:
Could a player “feel” this difference? How quickly? At eight
spins per minute, this represents a 📈 loss of approximately 2.4 bonus games per hour.
This is a question we can answer with science, but keeping this 📈 concept in mind,
consider this alternative version of the simple game that also achieves 13 percent
hold:
In this variation, we’ve 📈 returned the bonus game to a 1-in-40 proposition but
reduced the frequency of the top award to 1-in-220,000. This pay 📈 table should provide
identical game play as the 7.5 percent pay table, to nearly everyone who plays the
game. No 📈 one will be able to detect with any certainty that the top award has become
less frequent, as no one 📈 expects to hit the top award anyway. Given the option, we
would certainly put the “fewer jackpots” version on the 📈 floor before the “fewer bonus
games” version.
In other words, there are ways to raise hold without impacting the
player experience, 📈 and there are ways to raise hold while lowering the occurrence of
relatively frequent events that the player celebrates. To 📈 the extent that we can
accomplish the former, we should do so enthusiastically. With the latter, we should
proceed cautiously.
Only 📈 a careful review of PAR sheets, which detail pay tables and
frequencies of game awards, can give a clear indication 📈 of how hold changes will affect
player experience, and these can be cumbersome (I’ll say, politely) to read and
interpret. 📈 A broad-based hold increase without regard to how hold is increased will
certainly affect player experience.
An ideal hold strategy would 📈 be designed at the
game level. Operators and manufacturers would work together on how to provide the best
player experience 📈 while achieving operator financial goals.
Another consideration
pointed out to me by savvy slot operators is the speed of the processors 📈 in newer
games. They keenly note that players don’t necessarily experience hold as a percentage
of slot handle, but rather 📈 as a loss-per-hour. We are seeing max bets and cost-to-cover
on penny games increase, processor speeds producing more spins per 📈 hour, and holds
rising, resulting in even more substantial increases in loss per hour.
Conclusion
I’ve
worked with several properties on their 📈 slot hold strategy as an operator and as a
consultant. While there is no one overarching method for measuring the 📈 impact of slot
changes, I’ve been fortunate in my roles to work with talented teams of slot operators,
economists, statisticians 📈 and data scientists to develop methodologies to evaluate the
performance of slot hold changes.
We’ve developed benchmarks and metrics to look 📈 at
player behavior, machine performance, and overall property performance, each providing
a different lens into the effects of these changes. 📈 With forward-thinking operations
teams, we’ve run tests as aggressive as altering the hold on more than 30 percent (!)
of 📈 the machines on a casino floor. As expected, higher hold approaches have produced
more revenue on average in the short 📈 term, though at a mildly diminishing rate.
Most of
the studies were run for only six to 12 months, so I 📈 don’t know if in two, three, five
or 10 years we’d conclude that a lower hold strategy would produce the 📈 loyalty and
guest experience effects needed to outweigh the short-term effects of raising hold. Or
whether we’d find in the 📈 end that cranking up the hold produces stronger financial
outcomes across the board.
Casinos continue to navigate the tradeoffs of immediate
📈 gains at the risk of degrading guest experience in many areas—resort fees in hotels,
outlet fees in bars, ATM fees 📈 approachingR$10, and even parking fees. Regardless of the
enterprise’s overall strategy, taking a tactical, property-specific and game-specific
approach can help 📈 achieve the desired short-term financial outcomes while managing the
guest experience impacts and mitigating some of the potential long-term effects 📈 on the
business.
