It’s the question every live poker player has asked themselves at least once. You played perfectly. You read your opponent. You made the right decision. Yet you lost. The wrong card came out on the river. And the tournament was over.
Your fault or bad luck?
This tension — between skill and chance — is at the heart of the debate on live poker regulation in Italy. And it is the starting point of the Skill Poker Index (SPI).
- Is Texas Hold’em gambling? What is still missing
- What a skill index actually measures
- The two factors: what you know and what you can do
- The matryoshka effect: why the phases are not equal
- Why 60, 70 and 80, and not the exact numbers
- What the indices are for
- Is poker a sport? What would have to change
Is Texas Hold’em gambling? What is still missing
The debate on skill in poker has never settled on a shared position, inside the community before anywhere else. Yet a definition exists, and it is precise. It concerns the quality with which a player understands and uses the strategies and principles they have acquired.
The problem is that having it is not enough to see it. Even the professional who wins a tournament is judged lucky, and not entirely without reason: to go all the way you have to start behind and win the pot at least a couple of times. That is where variance concentrates, and it is why at the table skill stays invisible.
Italian authorities, meanwhile, classify poker as gambling. To leave that category you would have to prove that skill decides the outcome: it is the proof CONI requires to recognize a sport. It does not exist, and not because skill is missing: because no one has ever built a tool that measures it phase by phase, the same way for every table. As long as it is missing, the police chiefs decide, province by province.
What a skill index actually measures
A skill index measures the minimum ideal status a player needs to bring the skills they have into play. By status the model means something precise: the optimal cognitive readiness in relation to the size of one’s stack.
It is a distinction worth fixing straight away, because it is not intuitive. The index does not measure how good the person at the table is: it measures how much skill that phase allows anyone to exercise, whoever is sitting there. It is a property of the game, not of the person.
And it serves one purpose only. Without an all-in the indices do not intervene at all: the hand runs to the river and the rules stay as they always were. With an all-in, they stop a pot from being decided entirely by two cards yet to come.
The two factors: what you know and what you can do
An index arises from the meeting of two quantities, and both are measurable.
The first is information. Every card in the deck is worth 1.92% of everything that can be known in a hand, and as the board is revealed those shares add up: pre-flop you know a minimal fraction, by the turn most of what counts.
The second is the stack. Below a certain threshold you cannot play a whole hand: it takes around sixty big blinds to reach the river still having choices, the condition Arnold Snyder calls competitive utility. A player with fewer chips is not giving up skill by choice — they lack the means to exercise it.
The two quantities rise together, and at the river both reach their maximum. Information, because at showdown it is explicit and complete: nothing is left to be revealed. The stack, because by then it has stopped mattering: either you are already all-in, or you bet and were called, or you called your opponent’s bet. The gap between stacks, which in every earlier phase decided how much you could do, at the river decides nothing.
Neither factor limits anything any more. And that is why the remaining pot is not split: it is assigned in full.
The matryoshka effect: why the phases are not equal
The phases of a hand do not merely follow one another: they contain one another. The flop carries what the pre-flop had put on the table, the turn contains the flop, and so on. It is a matryoshka process, where each phase adds the previous one’s information to its own: this is why the indices rise instead of staying still.
At the river there is no share to assign: both factors are at their maximum, and the remaining pot goes in full to whoever holds the best hand there.
And that is what makes the model less forgiving than it seems. Whoever collects their own phase’s share collects nothing else: had they been ahead on the flop but no longer at the river, all the rest would go to the opponent. The advantage of one phase does not carry over.
Why 60, 70 and 80, and not the exact numbers
The indices are not born round. The calculation returns 63.84% pre-flop, 71.52% on the flop and 75.36% on the turn: three values with two decimals, which at a real table do not translate into chips. A dealer forced to assign 63.84% of a pot would hold up the hand for minutes, every hand.
This is why the model rounds. The original study rounded down and obtained 60, 70 and 75. The model adopted today follows a different rule: it stops at the nearest ten, because at the table a pot is divided into tenths, not twentieths.
The two rules give the same result in two phases out of three. They diverge on the turn, and only there, because the turn is the only phase whose index passes the halfway mark of its own ten: 75.36 sits just above the line of 75.
That margin is not hidden. On the turn the rounding falls upward, deliberately, to give more weight to the skill component than to chance. It is the second to last phase: the information is almost all on the table, and the model chooses to reward skill a little more than the calculation alone would require.
What the indices are for
Put to the test, the model does three things at once.
It rewards skill, because it favors whoever shows the best hand at showdown: whoever played the spot better than their opponent, not whoever caught the right card afterwards.
It pushes play to the river. The hyper-aggressive strategies that shove everything to make the opponent pass stop paying off, because the pot can no longer be taken whole by giving up the cards still to come. This is where fold equity loses its value.
It penalizes the mistake without erasing it. A player who had been behind at showdown and improved their hand by the river would still take the rest of the pot, but only that: the share of the phase would already be lost, and that is what the mistake would have cost them.
Is poker a sport? What would have to change
A recurring objection concerns dealers: assigning partial shares would complicate counting. It is less solid than it seems. In Hi-Low variants the dealer already evaluates the best and the worst hand of every player, and often splits the pot between equivalent hands: an operation more involved than this one.
What matters remains. Until today there was no objective criterion able to prove that the winner is the more skilled player and not the luckier one. The indices provide it without touching the game: the cards stay the same; the blinds go up normally. All that changes is what happens when someone gives up playing them.
It is the first step toward recognizing live Texas Hold’em poker as a sporting discipline in Italy, and toward EMPOKERMENT, the sporting variant that applies the Skill Poker Index.
This article opens the series dedicated to the Skill Poker Index. Next: why mathematics alone is not enough to answer this question, and which disciplines the study had to bring together.
