Hacker News Reader

From the article: 'The lottery’s rules require “a close approximation of the odds of winning some prize for each game” and say those odds “must be displayed on a Video Lottery game terminal screen.”'

Curzi discovered that the odds displayed were not correct and the commission knew about it.

That is also very puzzling. It gives the above quote but it does not say whether these odds were actually displayed on the screen of this particular game, and it does not say which odds were displayed (the ones with using autohold or the ones without).

I used to work as a mathematician making OSL video poker games and I can tell you that there is no requirement that the autohold in the game use optimal strategy in Oregon. I don't remember if it was even required to have autohold enabled on the device. It was a configuration option in our software and could be turned on or off. It's possible that Oregon requires it is always on, but I can't remember.

I can't speak for everyone that manufactures these games, but the company I worked for always made autohold use a simplified strategy (similar to whats found here: http://wizardofodds.com/games/video-poker/strategy/jacks-or-...) for ease of implementation. 9-6 Jacks or Better had a slightly different optimal strategy than 8-5 Jacks or Better, and it was easier to code a simplified strategy that every paytable variation could use. Obviously a simplified strategy will deviate somewhat from optimal, but not by a huge margin. Autohold is generally as good as any unskilled player would expect to be on their own. Anyone considering themselves to be a skilled video poker player would never rely on autohold during their play.

In this specific instance it sounds like a software bug or a poorly implemented hold strategy. There's no reason the game should be offering such bad advice and I'd hope the game maker would correct it in an update. I'm inclined to assume stupidity rather than malice on the part of the game maker in this instance. It's absolutely ridiculous to imply some kind of conspiracy on the part of Oregon State Lottery however. As someone who made these games I would have happily blown the whistle on that kind of shenanigans had I encountered them. The truth is that nobody needs to rig these games to make money. We already have the laws of mathematics on our side and cheating just sows distrust in the player base.

If anyone wants more insight on how these games are made please ask. There are lots of misconceptions about gambling devices and how they really work.

Do you recall which 3rd party lab validated the games for OSL?

I've spoken with a few colleagues that are ex-WMS and the consensus has pretty much been "yeah, that sure sounds like a bug but why didn't GLI catch it?"

I worked with GLI when I was doing it.

Edit: Now that I've (finally) finished the article I see it was a WMS machine. I worked for WMS and on this exact product before I left. I'm not sure why they didn't name the manufacturer in the first paragraph of the article.

It's been years since I've worked on that poker product, but I remember the autohold on Jacks or Better used a very short, simple ruleset that didn't do anything obviously boneheaded. I'm having a hard time imagining it doing what the article described, but I don't have access to the system to check it out anymore. The autohold ruleset was built long before I started there and was ported from system to system.

The nuances of autohold are something that's really hard to explain in terms of the PAR sheet, in my opinion.

Is the goal of autohold to provide the player with any win, or give the proper chance at the optimal win given the upcards? Those two goals aren't always in alignment given certain hands.

(ex-WMS engineer here too. Glad to see you survived.)

Another purpose of autohold is to provide a floor for unskilled players in a game of skill. They don't use it in Las Vegas specifically to allow players to make as many mistakes as possible. Optimal strategy on 9-6 JorB is 99.54% RTP. Reverse-optimal strategy (making the worst selection every time) is something like 3% RTP. Use of autohold puts an absolute worst case on the downside to video poker.

But do states set a min RTP for something like autohold? I always understood it as something that, technically, didn't have to be held to a PAR sheet or validated at all, really.

Slots with bonus games could auto-percentage themselves, right? Like how does Jackpot Party hold to a PAR sheet percentage when the player is pretty much making random pokes? Or are the pokes gaffed to reveal the five virtual reels that were spun earlier in the play?

Each of the state jurisdictions have different rules regarding minimum RTP percentages in games. With regards to poker autohold there aren't any rules that I'm aware of other than it has to be above the regulated minimum for the state. For example, Nevada had a minimum RTP requirement to 75%, so autohold would have to be at least that good. (Note to tourists: never play the slot machines in the Las Vegas airport. They're probably barely over the minimum required by law)

I agree that the article wasn't very clear on this. It does, however, say that Oregon law requires the lottery to tell players the win chance.

The state lottery knew that the win chance they told players (0.9), was actually higher than the win chance when using 'auto-hold' (~0.85).

That doesn't mean that Curzi will win, of course.

One of the reasons I hate traditional media.

Articles generally get built around a story such as what we see here, leading with a human-directed hook, and then proceeding to build up the human backstory - even when it's only vaguely related to the actual topic. In this case I skimmed past maybe half the article as useless fluff.

Willamette Week isn't exactly traditional media - it's an alt-weekly known for investigative reporting. I think it's the only alt weekly ever to have won a Pulitzer, actually.

>For example - does oregon law require that the machines have certain winning percentage? And secondly did the state of oregon make a representation to gamblers that this machine had certain winning percentage?

What relevance to the claim being made does this have at all? The question is whether a feature on the machines intended to accelerate play by automatically choosing the ideal card draw was actually doing that. It has been determined that it was not (unlike every other supplier of machines that Oregon uses), and those machines have now been withdrawn.

>And finding documents marked "confidential" that say the machine has a certain winning percentage means nothing. You cannot say that a confidential document that was unknown to gamblers before Cruzi obtained it from a public records request mislead anyone. If people do not know about a document there is no way it can mislead them.

This objection is incomprehensible. It is not being claimed that players were being deceived by the listed winning percentage of machines within documents marked "confidential." The documents are being used to prove that Oregon was aware of an error in their favor within the autoplay algorithm on those particular machines, therefore consciously benefiting from that error by possibly over $100M.

Simple example: suppose healthcare.gov (US healthcare exchange under ACA) had a feature that calculated the amount of insurance plan subsidy that you were eligible for based on the information that it had about you. You would then be optionally allowed to review the calculation before selecting the plan, or click "NEXT." Further suppose that this feature was broken, and unless the calculation was reviewed, it would consistently ignore certain subsidies whether or not you qualified for them. Now, add that this error was discovered, but since it was in the government's favor, was ignored, resulting in a loss of over $100M to people who pressed "NEXT" (or lets say $100M * [US population/Oregon population] = $8.1 billion.

Would that be a problem?