Understand what I mean when I say "like poker": the poker mindset is a very analytical, statistically oriented approach. The best players never make a bet/call/fold without some form of logical basis for their plays. They have an internal model that tells them "this play will succeed 50% of the time and I stand to win 3x my risked value and so it is profitable". Similarly in engineering, no decision should be made without an internal model where risk can be evaluated - "this structure will fail 5% of the time and kill somebody but it costs $10,000,000 to get more out of it and so it is not profitable".

Engineering - especially tunneling - is a gamble. You cannot design away every risk. All we can do is show up with our belts and suspenders on and hope our pants don't fall down when we're running out of the hole.

Everything else you described about risk is not unique to poker. In engineering, particularly major projects with real safety concerns, it's risk analysis. This component has a 4% chance of failing in this way, so add redundancy to mitigate it. Put in 10 sensors instead of 4, and require a consensus algorithm to reduce the risks from sensor failure or misreporting.

In engineering we do gamble, on a lot of things. But it's not a pure gamble, we aren't just throwing dice. We plan, prepare, rehearse, and execute.

And a bit of a nitpick, but we are just throwing dice. The entire idea of a safety factor is an abstraction to get away from statistics which is why it has been almost completely replaced in structural engineering with LRFD which is based on statistical analyses of component/system failures.

I'm a civil engineer, and I do structural design work for a living.

If you're not trolling and you do mean what you've been saying, you're being very disingenuous and in the process proving to be true and perfectly valid all comments regarding the virtues of the swiss system in contrast with the typical engineering work around the world, the one you so vocally identify yourself with.

Engineering is not poker. It never was. Only bad engineering is like poker, and it shows in their work. You're confusing basic aspects of reliability analysis (the "this structure will fail 5% of the time" comment) with being poker. European structural design standards dictate that the lower failure probability of an engineering work, for its entire design working life, is around 0,05%. This is not poker.

As you've mentioned, the design process is centered on risk analysis and balances the economic impact of safeguarding against risk scenarios, but this is entirely immaterial for the discussion. The discussion is about planning, and the lack thereof. All projects are planned to an extent, but the point is that planning is often insufficient in the sense that important variables are left unknown until breaking ground for any number of reasons, one of which is this mentality that its unacceptable to spend much on risk assessment ("why spend money on non-destructive tests to infer the geotechnical profile and properties of each stratum in the preliminary design stage if we're boring a hole in there and eventually we'll reach that point ourselves?").

Unknowingly or not, you've supported this observation throughout all this discussion by arguing in favor of the virtues of handling a civil engineer project like "poker", and tolerating and accepting the occurrence of surprises associated with unacceptably high risk probabilities that go way up to 5%. This is precisely the problem everyone is pointing out and you've tried to deny but ended up supporting. This is the sole responsible for the typical cost overruns in civil engineering projects, and you inadvertently demonstrated why sadly this still rings true up to this day.