Deputy Editor
Winning players who late-registered a simulated poker tournament vastly outperformed their early-registering counterparts, according to a game theoretician and his computer model.
“Tombos21,” who frequently posts about poker theory on social media, shared on Reddit a study he recently conducted. According to Tombos21, registering as late as possible for poker tournaments can nearly double a winning player’s statistical return on investment (ROI)… as long as the other pros aren’t all doing the same thing.
There are two sides to the late-registration debate. Proponents point to “ICM effects” (short for Independent Chip Model) that make a starting stack worth more in equity as the field gets smaller and the tournament gets closer to the money. However, contrarians point out that weaker players are quicker to bust out on average, so the field is softest early on and late-registering pros miss out on exploitation opportunities.
To test the relative strength of these arguments, Tombos21 broke down his theoretical player pool into three segments: early-registering pros, late-registering pros, and recreationals.
Here’s what each group’s ROI looked like across 1 million simulations, with certain starting assumptions.
- Early pros: +15.2%
- Late pros: +27.9%
- Recs: -18.4%
“Late reg cohort won by a landslide,” Tombos21 wrote. “It’s hard to beat the ICM benefits conferred from late registration (10-15% ROI boost right off the bat from pure ICM advantage).
Now obviously one simulation doesn’t prove anything. But I question the mantra that you should always play soft fields from start. Is it really worth giving up 10-15% of your ROI just to have a slightly softer field?
Late-Regging as an Edge is a Well-Established Concept
For late registration advocates, Tombos21’s study isn’t telling them anything they don’t already know. The connection between late-registering poker tournaments and a higher win rate has been well-established. Kenny Hallaert published a more in-depth study on the matter back in 2019. PokerStrategy’s Barry Carter produced a similar conclusion when writing for GTO Wizard.
In a nutshell, both demonstrated that every player gains equity from those who bust, and late registrants reap all that equity with no risk.
Carter, went further, arguing that late registration is even more advantageous to non-professionals, as playing with a shorter effective stack reduces the skill deficit.
“A small losing player might become a breakeven or even winning player if they max late-registered tournaments,” he wrote.
Of course, purely recreational players get less entertainment for their buck if they’re buying in late, so it’s worth considering what one’s purpose is in playing — to make money, have fun, or both.
In that sense, Tombos21’s conclusion is wholly unsurprising. However, as he acknowledges, the model isn’t perfect. In its abstraction, it misses some potentially important context.
Skill Advantage Isn’t a Constant
Tombos21 explained their methodology in the study. The virtual players had stacks of 1 and went all-in against each other at random. To simulate skill advantage, Tombos21 gave the pros 52-48 edge in each such matchup against a recreational.
Tombos21 admitted it’s a “caveman MTT simulation” that probably only captures “90% of the info we want.”
One key piece of context is that the advantage of playing the early levels increases with the skill edge. Repeating the simulation with a 60-40 edge (rather than 52-48) made it about equally advantageous to register early as late.
Furthermore, Tombos21 pointed out that late-registration can be self-defeating if everyone does it. The more of the winning players who register late, the softer the early tournament field is. Hypothetically, if every pro registers as late as possible, a single pro willing to buck the conventional wisdom would get all of the fish to themselves. Real life tournaments probably do feature a less extreme form of this dynamic, as the majority of players registering close to deadline are going to be pros. Recreational players tend to prefer early registration, getting more play for their dollar.
There is one additional consideration that’s harder to account for. PokerScout Managing Editor Alex Weldon commented on Tombos21’s Reddit post to point out this flaw, which the theorist readily agreed with. It is the flip side of Carter’s point about stack depth and skill edge: pros’ advantage over recreational players isn’t fixed, but is greatest when the stacks are deepest. Setting the edge at 52-48 (or 60-40 for that matter) isn’t representative of real poker dynamics. A more detailed simulation would feature a higher skill edge early on, a lower one close to the registration deadline, and then an increase as the bubble approaches and ICM considerations complicate the decision-making.
Late-Reg Question Still More Important to Operators than Players
At the end of the day, Tombos21’s simulation is unlikely to move the needle, but it is an interesting opportunity for discussion and education. Most clued-in players are already well-aware they can gain an edge by registering late, and the dissenters probably won’t be convinced by a rudimentary simulation.
That said, there are plenty of arguments for registering early as well. Since many professionals and winning players have the bankroll to support multiple entries in an event they’re playing, and most events these days allow multiple entries, it’s not really an either/or question.
To recap:
- Registering late is generally good, because of the ICM advantage.
- Registering early is good for ROI only if your skill edge is huge, or all strong players are registering very late.
- Registering early also makes sense for recreational players, provided they care more about duration of play than their expected winnings.
However, the thornier question is for tournament organizers: If allowing late registration for pros is bad for recreational players, how long should registration stay open for? It means more entries and is good for the operator’s immediate bottom line, but is arguably bad for the ecosystem.
Hallaert grappled with this as an occasional tournament organizer himself. Despite the known problems, many operators still allow players to register extremely late. The World Series of Poker, most prominently, allows players to register with as few as 10 big blinds in some events.
Others are bucking the trend. The World Poker Tour, for instance, cuts off registration much earlier than most of its competitors.
Both tours have had and continue to enjoy plenty of success, so there’s no clear-cut answer in navigating the ongoing talking point of late registration in poker tournaments.
