This is not a good picture, but it does emphasize the rugged terrain along which new roads are being constructed to the site of the Galore Creek not-soon-to-be project.
More on the NI 43-101 report from that superb mining news provider, Mineweb:
In an analysis published Tuesday, Haywood Securities metals analyst Kerry Smith, a professional engineer, said the decision by Teck Cominco and NovaGold to suspend the Galore Creek project “calls into question the quality of other feasibility studies.”
“This recent development now calls into question the quality of other feasibility studies conducted by engineering firms for large projects, such as Pebble, Donlin Creek, Fruta del Norte, Cerro Casale, Fort Hills, Pascua Lama, Andacollo sulphides,” Smith suggested.
However, Smith blames neither Hatch Engineering, which prepared the original 2006 feasibility study for Galore Creek, nor AMEC, which reviewed the original study and found that capital costs would be significantly higher. The original US$2 billion estimate has now more than doubled to US$5 billion.
“Part of the problem is the difficulty is accurately estimating the capital cost for a project of the scope of Galore Creek in a tight mining construction market where costs are rising rapidly and competition for skills is high,” Smith noted.
I agree with Mr. Smith regarding the quality of other feasibility studies and advise potential investors to read carefully before relying thereon. But I disagree with him regarding the difficulty of accurately estimating capital costs. The simple fact is that it is not difficult; it is impossible.
But there is a way: probabilistic cost estimating. This was done in a half-hearted way for Galore Creek. It was not done, however, with the cold-blooded, hard-hearted ruthlessness and honesty the process demands if it is to yield probability distribution curves reflective of reality.
People do not like probability. Engineers do not take kindly to distribution curves that give ranges of costs particularly when there is a finite probability of huge cost increases and project failure. I know from experiences I had many years ago working with ARCO. We were dealing with contaminated site cleanup on one of the many sites ARCO was responsible for cleaning up. The site contained radio-active waste. The remediation cost estimate based on ordinary-old deterministic cost estimating procedures was about $10 million.
The ARCO project manager called in specialist consultants who made us sit down and honestly estimate the range of possible costs associated with handling radio-active waste in a densely populated part of the USA. We screamed and resisted the push to go beyond a mere few percent one way or the other. But over the course of a few days we had to admit the potential (albeit with low probabilities) of huge cost increases. Using a primitive version of current computer codes for Monte Carlo probability cost estimating, we were appalled to hear that there was a five percent probability of the remediation cost going to $30 million.
Impossible, we screamed. That is what comes out of your numbers, the code cruncher replied. Needless to say, this high number forced a hard rethink by ARCO and ourselves. I cannot tell the rest of the story, other than to say that this exercise in honest probabilistic cost estimating headed off a cost over-run disaster.
My point is the techniques exist for dealing with situations like Galore Creek. They are not nice or easy to use. They cause distress and disbelief. But they are scientific and ultimately factual. Until the mining industry embraces them fully, I fear we will see more Galore Creeks.
