Most fishing tools are built to help you go fishing.
That makes sense. You open an app, check a report, look at a map or scroll through recent catches because you're already halfway to a decision. You want to know where to launch, what the tide is doing, what people are catching and whether tomorrow looks promising.
The Fishability Index Engine has a slightly different job.
Sometimes it needs to tell me not to go.
That sounds simple until you think about what the tool is actually competing against.
A fishing trip already has momentum before the truck leaves the driveway. I've checked the weather. I've been thinking about the water all week. The batteries are charged. Gear is staged. Maybe I've bought bait, filled the tank or blocked out a day that could have been spent doing something else.
At that point, WAIT isn't a neutral answer.
It's expensive.
The Cost of a Bad GO
The obvious failure of a fishing recommendation is sending someone into bad conditions.
But "bad" isn't limited to dangerous weather or an unfishable river.
A day can be perfectly fishable and still be a poor use of time.
Maybe the target species isn't there in meaningful numbers yet. Maybe water temperature has moved outside the productive range. Maybe wind makes the technique I intend to use miserable. Maybe recent catch data has collapsed. Maybe the tide, flow, run timing and pressure trend are individually acceptable but collectively mediocre.
None of those necessarily mean you can't catch a fish.
That's an important distinction.
The engine isn't trying to predict whether a fish can be caught.
It's trying to answer a more practical question:
Given what we know right now, is this trip worth making?
Those aren't the same thing.
Someone will always catch a fish on a bad day. I've been that person. I've also been the person who drove several hours, burned fuel, launched a boat and spent most of a day discovering what the data could have told me before I left.
The second outcome is what interests me more.
One of these is on the record now
On the first of August the engine put a Red Flag warning at the top of a sheet for two anglers. They cancelled.
The part that mattered wasn't the warning. The National Weather Service found that hazard, not me — the engine just refused to bury it under a score.
What mattered was what sat underneath it. The fishing that day scored 64 to 66. They weren't giving up a good day. They were giving up a mediocre one, and knowing that turned an agonising decision into an easy one.
That's the version of WAIT I want the engine to produce. Not "don't go."
"Here is exactly what you're giving up, and here is what it would cost you."
Added after publishing, twice — and the second time corrects the first.
The first version of this note said the engine had got it wrong.
I'd found out they were launching at first light, and the advisory didn't start until eleven. That looked decisive. A whole-day veto had cancelled a trip that would have been off the water before the hazard arrived — the day-unit problem turning up in the one output where it's most expensive to get wrong. I wrote that up and published it.
Then I asked how long they'd meant to fish.
All day.
Which puts most of their trip inside the advisory, against gusts to fifty, and makes the WAIT the right call at the right size. I retracted a true claim on half the facts.
The part I'm least comfortable with is that I already knew which fact was missing. How long were they going to fish was written down as the open question in three separate places, including this essay. I published the correction between asking and hearing back.
There is still something real here, and it's smaller than what I claimed. An engine that could say "go, and be off the water by eleven" would have offered them a morning instead of nothing. Maybe they'd have taken it. Maybe five hours isn't worth the drive for a day that scored 64. But "don't go" and "go, with a deadline" are different answers, and right now the engine can only produce the first.
So the WAIT was right, the day is still the wrong unit, and the most expensive answer in this piece turned out to be one of mine.
WAIT Has to Mean Something
If every marginal day eventually gets rounded up to GO, the score stops being useful.
That's especially tempting with fishing because uncertainty is built into the activity. There is almost always a reason to believe things might work.
The weather might improve.
The fish might move.
The bite might turn on.
The report might be stale.
All true.
But if uncertainty always becomes an argument for going, then the engine isn't evaluating anything. It's just finding increasingly sophisticated ways to tell me what I already want to hear.
So WAIT has to survive that temptation.
If run timing is weak, conditions are deteriorating and recent evidence doesn't justify the trip, WAIT should be the answer even if I'd rather see GO.
That is probably the most valuable answer the engine can produce.
And it may be the one I like least.
The Other Half of the Deal
There is a problem with giving WAIT recommendations, though.
Sometimes they will be wrong.
The same is true of GO.
Fishing has too many variables for any scoring system to eliminate uncertainty. Conditions change. Data arrives late. Fish behave inconveniently. A variable I thought mattered heavily may turn out to matter less, while something I discounted may prove important.
That means the field notes can't exist only to document successful predictions.
When the engine gets one wrong, that result belongs in the record too.
Not as a confession. As data.
What did the engine predict?
What actually happened?
Which inputs were accurate?
Which assumptions weren't?
Was the weighting wrong, or was the outcome simply inside the uncertainty that fishing always carries?
Those questions are more useful than pretending the score was right because I can explain the result afterward.
A system that can only explain its successes isn't learning much.
Two that went wrong in August
In the middle of the month the engine gave a tester an 80 and a GO for a Monday, and told him to fish it between 5:45 and 9:00 in the morning.
He launched at four in the afternoon instead. It was the best session of his trip by a factor of four — eighteen fish across three species.
The day call was right. The day was worth fishing. But the three hours the sheet pointed at were hours he never fished, and the sheet had nothing to say about the ones he did.
I only know that because he mentioned when he'd actually gone out. Nothing in the score would have revealed it. An 80 that produces eighteen fish looks like a win from every angle except the one that matters — whether the advice is what produced them.
A week later it emitted a three-hour window for a jetty session — eight in the morning until eleven.
Nothing happened for the first hour and forty minutes. The rockfish started at 9:40 and were still coming at 11:15. The salmon hit at 11:47, forty-seven minutes after the window I'd given had already closed.
So the window was wrong at both ends. It opened too early and it shut while the fish were still biting. The engine had told me to stand on a jetty for the hour and forty minutes that produced nothing, and to go home during the part that worked.
Neither of those is a catastrophe. Both are the engine being less useful than its score implied.
And both are more instructive than the days it got right — because a day it got right could always have been luck.
The Test
I thought this would be the test of the Fishability Index Engine.
It turned out to be the test sooner than I expected.
Not whether it can identify a beautiful fishing day. Most of us can do that.
The harder question is whether it can identify the days that look tempting but don't justify the trip — and whether I'm willing to listen when it does.
Sometimes I've followed the recommendation. Once I passed it along and two people stayed home — and having now got that story wrong myself and corrected it in public, I still think it's the most useful thing the engine has done.
Twice it was wrong about when — once because a tester happened to mention what time he'd launched, and once because I was the one standing on the rock.
Either way, the result goes back into the engine.
Because the goal was never to build something that tells me to go fishing.
I already know how to do that.
The goal is to build something that helps me decide when it's worth going.
Want a sheet for your next trip? Any water in Washington State — tell me where and when.