While federal regulation of online poker in the US may still come to pass, several states in the country are already moving forward individually to regulate online poker.
The common wisdom is that states will pool their players in interstate networks. Establishing such networks, however, will be a long and potentially complicated process.
What online poker players in the US are therefore likely to be left with - at least for some period of time - is a collection of individual online poker rooms run by a variety of states. So, which of the 50 US states could actually support such a room?
My opinion: About 11 states for certain, with another 5 that could make it happen if they (ahem) play their cards right. Read on for the longer answer.
Caveat: This article is a casual analysis and not a scholarly paper. A lot of my math is therefore quite rough. My goal with this article is just to spark conversation and definitely not to be the final word on the subject.
What constitutes a "viable poker room?" Let's consider two basic thresholds using data provided by PokerScout.com gathered on July 31st, 2012.
1000 peak / ~500 average players - With 1000 active cash game players in the lobby of a poker room, you can expect a solid micro-low stakes selection focused primarily on NLHE. You will also likely find only minimal mid-stakes and above games and few, if any, exotic games. As a point of reference, Everest Poker houses a shade over 1000 players at typical peak hours. The 7 day average traffic level for Everest is 550 players.
2000 peak / ~1000 average players - With 2000 active cash players, you'll get a broad selection of micro-low stakes games, a reasonable selection of mid-stakes and a smattering of higher stakes. This level of traffic is also easily substantial enough to support multiple formats (i.e. NLHE and PLO). To provide a real-world example, the Microgaming Network generally hovers around this number at peak times. The 7 day average traffic level for Microgaming (as of July 31st, 2012) is 1100 players.
So, anything around 1000 players at peak times / 500 players on average should be enough to guarantee you a modest, yet active and viable, online poker room if you're a US state.
The next step: trying to calculate what percentage of a typical US state's population will play online poker for real money.
There is no "standard" ratio. A number of factors, some of which we'll discuss in the next section, impact the participation rates of individual markets. It's still useful to explore some current / recent examples of closed online poker markets to at least get a rough baseline for participation rates.
The following sketches draw on data from Wikipedia, PokerScout.com, research by Ingo Fiedler and Ann-Christin Wilcke (The Market for Online Poker) and research by Kahlil Philander and Ingo Fiedler (Online Poker in North America: Empirical Evidence on its Complementary Effect on the Ofﬂine Gambling Market).
Sweden is an interesting example of a quasi-closed online poker market. The market itself isn't closed - Swedish players can (and do) play on a variety of sites like PokerStars and iPoker. However, Sweden is home to a ring-fenced poker room - Svenska Spel - that only accepts Swedish players.
Let's look at the numbers. Svenska Spel is one of the top 15 online poker rooms by real-money volume, with a peak of about 1900 players and a 7 day average of 730. Sweden has a population of about 9.4 million. Without doing any additional math, you could broadly conclude that - all things being equal - any state with a population greater than or equal to Sweden's could support a viable online poker room. Digging a bit deeper, you could chop the number down even further:
Some rough math taking the above into account puts Sweden's effective participation rate at around 85 average players per million. Given that rate, a demographically similar population of 6 million would meet or exceed the required player pools for a viable online poker room.
France is a pretty tightly ring-fenced market. A handful of major operators such as PokerStars operate a closed version of their poker room exclusively serving the French population. The total size of the market (all traffic at all operators combined) is 8805 players at peak and 4315 players for the rolling 7 day average. France has a population of about 64 million.
That presents a more pessimistic picture of participation rates than Sweden - roughly 66 average players per million. If France is our guideline, then a state would need a population closer to 8 million to meet our minimum thresholds for a viable online poker room. One possibly relevant note: The rake at French poker rooms is notoriously terrible (one of many complaints French players have about state-regulated online poker) and may artificially depress the country's participation rate.
Italy is more or less analogous to France, although a few unregulated sites still serve Italian players. Those operators aside, the country is home to a handful of Italy-only poker rooms, including Party, OnGame, iPoker and a few others. Total peak total players on these closed sites is 7660, with a 7 day average of 3860.
The population of Italy is about 60.6 million. Not surprisingly, Italy ends up being closer to France than Sweden, with roughly 64 average players per million - a participation rate that requires a minimum population closer to 8 million in order to sustain an online poker room.
This is a bit more speculative of an exercise, but still interesting nonetheless. Full Tilt Poker stopped serving all players from Washington State on November 12th, 2010. PokerScout concluded at the time that the decision cost FTP about 3% of their cash game traffic. Our best estimate of Full Tilt's average traffic at the time immediately prior to pullout from WA is right about 18,000.
If PokerScout is correct, that suggests Washington State - with a population of only 6.8 million - was responsible for about 540 of the average players on Full Tilt. That results in a participation rate of just under 80 average players per million. The actualy number is likely quite higher as other rooms were still serving players from WA at the time.
What's also interesting is that those numbers are from an era when the UIGEA was already in force and when Washington State had already passed a much-publicized law criminalizing the act of playing poker online. Even with those strikes against a high participation rate, players from Washington State apparently logged on in greater numbers than their French and Italian counterparts currently do in their tightly regulated markets.
By looking at traffic numbers before and after Black Friday, we can get a pretty good sense of the size of the US market. While not a perfect method, it's good enough for my purposes in this article. In May of 2011, PokerScout posted an analysis of post-Black Friday traffic. Given the relatively small gains realized by EU operators like 888 and Party, it seems safe to assume that the majority of traffic losses for Full Tilt Poker and PokerStars resulted from the absence of US players as opposed to a mass migration on the part of players from other countries. Based on those losses, and rough estimates of the smaller networks that still served US players, the pre-Black Friday market in the US consisted of something close to 16,000 average players - a participation rate of about 53 players per million - meaning we'd need a population just south of 10 million to launch our viable intrastate online poker room.
That rate, of course, is almost certainly depressed by a number of factors - the UIGEA, actions on the part of a few individual states to explicitly outlaw online gambling and a general perception among mainstream Americans that playing poker online was illegal. With such a large population, you would probably also expect significant variations in participation rates from state to state. You'd be right, and we'll cover that issue in a bit. For now, one final case study.
After considering a large population, let's take a look at a fairly small one: Canada's Northwest Territory. With a population of only 43,000, it seems like an unlikely haven for online poker players - yet an analysis of the Canadian market found that the Northwest Territory had the highest participation rate of any region in the country at 2.45%. How does that translate into average players? Messily. Using PokerScout traffic data we can fix the total size of the pre-Black Friday market at about 63,000.
The analysis linked above concludes that Canada is responsible for about 5.75% of the global player pool, or 4,246 of the average players online during that period. The Northwest Territories, in turn, account for about .12% of the Canadian market, or about 5 of those players. If you could somehow replicate that region across the globe, you'd be looking at an online poker participation rate of 116 average players per million - meaning you'd only need a population of 4 to 5 million to hit our magic number for a sustainable online poker room.
So, to recap - Country, average players per million residents (ap / m) and the population threshold each country's rate suggests is necessary to generate and support a poker room of > 500 average players:
There's always a but. A few factors - such as GDP, Internet access rates and cultural attitudes - can significantly impact the participation rate of a given country or region. To give you an idea of just how significant an impact: According to 2010 research from Philander and Fiedler, the top US state by participation rate in online poker relative to population (Nevada) had a rate more than three times higher than the state with the lowest rate (South Carolina). The same research concludes that every 10k increase in GDP per capita increases participation rates in online poker by .1%.
It's way beyond my scope here to articulate the specific impacts of each factor, but it's safe to say that:
There are also a few reasons to believe that at least some states in the US could outperform their international (or historical) counterparts when it comes to online poker participation rates:
Of course, it's not all sunshine. States could also drop the ball in a number of unique ways, resulting in worse-than-expected participation rates from residents:
Right, that was the original question. So, what we have so far is something like this: A state with a population of at least 7 million with at least a tolerant attitude toward gambling (or lacking stringent opposition to gambling) should be able to generate and support our hypothetical online poker room, a room that would look a good bit like Everest Poker or PKR Poker . As the per capita GDP of the state moves higher, the required population moves lower; as GDP drops, the required population increases.
Which US states fit the bill? Let's start with population. While we're looking for 7 million, I'll widen the range because a smaller state could probably overcome their population deficit with high participation rates, GDP per capita and internet access.
22 US states have a population above 5 million. Of those, we can quickly strike Missouri and Tennessee (borderline anyhow due to population size) due to historical attitudes toward gambling. Arizona and Colorado are also at the low end of the population range and can probably be removed from the list. That leaves us with 18 states. Of those 18, I think everyone would agree that California and Texas could easily support standalone poker rooms.
I would add New York, Florida, Illinois and Pennsylvania to that list - all have populations above 10 million and most have higher-than average GDP per capita (sorry Florida). Additionally, we've already concluded that Washington can support a room. So, we now have 7 "sure things" and 11 states that are best described as borderline in terms of their potential to support a room. Those 11, by population:
Here the research from Philander and Fiedler is again useful; it provides estimates of participation rates by US states in the pre-Black Friday market. The average rate for states is .46%. Here, I'm not worried about what the number means in an absolute sense - only a relative one. Washington's rate, if you're curious, is slightly below average at .41%.
How do our 11 states stack up?
With lower rates than average, Georgia and North Carolina likely drop off the list. Massachusetts nabs a spot with New York and company thanks to a much higher than average rate, as does Ohio thanks to a large population and a slightly above-average rate. New Jersey, Virginia are similar to Washington in terms of population and GDP, earning them the green light. That leaves Michigan, Indiana, Maryland, Wisconsin and Minnesota in the too-close-to-call category.
These states can almost certainly support a standalone room: California, Florida, Illinois, New York, Pennsylvania, Texas and Washington.
States that are less certain, but still highly likely to be able to support a room: Massachusetts, New Jersey, Ohio and Virginia.
States that would need a bit of luck: Indiana, Maryland, Michigan, Minnesota and Wisconsin.
All other states would face very significant obstacles that would almost certainly prevent them from running a viable intrastate online poker room.
That's really a topic for a separate article. Probable summary of that article: Delaware has no real shot (and doesn't really care about poker anyhow). Nevada might be able to pull some magic out of their gambling hat (especially if casinos are willing to operate at a deep loss to gain share and a leg up on the eventual national market) and somehow stitch together a viable room from some combination of their native population and the revolving door of gamble-happy tourists but the odds for the Silver State still seem a bit thin.