The most repeated piece of fall walleye advice is that once a lake turns over, the fish drop into the basin and you should follow them down. It is wrong often enough to cost you the best month of the year. The guide consensus, laid out well at Wired2Fish, runs the other direction: fall walleyes frequently push shallow, onto wind-blown rock, rip-rap and whatever green weeds are left.
Be careful how strong a claim that is. This is accumulated guide experience, not a controlled study, and it's a pattern rather than a law. There are lakes and days where the deep fish are the only fish. But the mechanism is straightforward physics, and once you have it, the shallow bite stops looking like an exception.
What turnover actually does
Through summer, most lakes deep enough to stratify are built in layers. Warm light water on top, cold dense water underneath, and between them the thermocline, a narrow band where temperature drops off a cliff. That break does real work for you in July: it concentrates fish at a predictable depth and holds them there. We mapped how that plays out on summer bass, and the physics is identical regardless of species.
In fall, the surface cools. Cooling water gets denser, so it sinks, and the water it displaces comes up. Wind stirs the rest. The layers collapse and the column goes roughly uniform top to bottom in temperature and dissolved oxygen. That is turnover.
What matters is what turnover removes: the thermocline, and with it the reason depth was a refuge. In lakes where the deep water had gone oxygen-poor over summer, turnover reoxygenates it, but it also erases the gradient that had been stacking fish at a specific depth. Once depth stops sorting the fish, food does.
So the fall question is not "how deep" but "where is the bait." In a cooling lake, that's usually structure still holding warmth, cover and current.
Where they actually are
Three pieces of shallow water are worth your time, and they rank in this order for me.
- Wind-blown rock. Main-lake points, reefs and gravel bars taking wind on the face. Wave action pins plankton against the structure, baitfish follow the plankton, walleyes follow the baitfish. The chop also kills light penetration, which matters because a walleye's eye carries a tapetum lucidum, a reflective layer behind the retina that makes it deadly in low light and unhappy in bright.
- Rip-rap. Causeways, dam faces, roadbeds. Rock absorbs sun and holds a few degrees of it, plus crayfish and small forage in the cracks. Rip-rap fishes best taking wind, worst slick calm.
- Remaining green weeds. Most vegetation is dying and rotting by late fall, and rotting weeds pull oxygen out of the water. The patches still genuinely green are still producing oxygen and still holding bait while everything around them goes brown. Find green, fish green.
If you only remember one variable, make it wind. Wind direction is the most useful piece of information you have in fall, more useful than water temperature and much more useful than the moon. It tells you which side of the lake the food chain got pushed onto this morning. A spot dead yesterday on a north wind can be the whole day on a west wind. Not superstition. Just where the plankton went.
Presentation, including what doesn't work
Fall fish are eating hard, which means you can move faster than in midsummer. I fish a jig and a soft plastic more than anything else, because it lets me hop rock without hanging up constantly, and I'd rather cover a point in ten casts than twenty. Blade baits are the other honest answer: a lift-and-drop in cold water triggers fish that follow a steady presentation without eating it. Crankbaits ripped over the tops of rock work when the wind is really up.
Casting a jig into wind all day is also the clearest case there is for picking the right reel type rather than the one you already own — the spinning-versus-casting tradeoff gets decided by lure weight and wind more than by preference.
Now the part that gets left out. Slow-trolled live bait rigs dragged through the deep basin, the default October presentation, are frequently a waste of an afternoon on a lake that has turned over. Not because live bait stopped working, but because you're presenting it where the fish left.
And during active turnover, when the lake is churning and the water looks like weak coffee, the fishing is often just bad. I've never seen a lure that fixes bad water. Fish shallow anyway, fish fast, and accept that you may be scouting for next week rather than catching today.
Minnesota wants to cut the limit from six to four
The tactics have a regulatory shadow this year. The Minnesota Department of Natural Resources has proposed cutting the statewide walleye limit from six fish to four, with only one over 20 inches, effective March 1, 2027 if it clears the 2026 rulemaking process.
The DNR's stated rationale is three-part: climate, invasive species, and a sharp increase in fish-finding technology use. The six-fish limit dates to 1956, which is the number worth sitting with. It was set for anglers with paper maps and a flasher at best.
The invasive-species leg of that is not hypothetical in this basin. Whether carp can establish in the Great Lakes system is a genuinely contested question, and a management agency writing rules for 2027 has to price in a risk nobody has resolved.
Support looks real. An online survey found 61% in favor; on-water interviews from 2021 through 2023 found 67%. Separately, the DNR announced on March 5, 2026 that Mille Lacs would keep its three-walleye open-water bag for the season.
The technology clause is the interesting part
Naming fish-finding technology as a rationale puts this proposal into the forward-facing sonar argument. Forward-facing sonar, or FFS, is live imaging that lets an angler watch fish and lure in real time out ahead of the boat. We covered the 2026 tournament restrictions on FFS earlier this year.
Both sides have a real case. The efficiency gain is not imaginary, and if an agency sets a bag limit as a proxy for how many fish a typical angler can realistically catch, a large jump in catch efficiency is a legitimate reason to revisit the number. Regulation follows capability.
The counterargument is equally legitimate: bag limits should be set on population data, on recruitment and mortality and year-class strength, not on how anglers find fish. If the population supports six, six is the right number regardless of anyone's electronics, and folding a technology complaint into a harvest rule mixes a social question with a biological one.
Minnesota is not the only state watching. A statewide cut justified partly on technology is a template.
None of which changes what you do in October. Check the wind the night before, find the rock it's blowing into, and fish the top eight feet with a jig or a blade bait while everyone else idles over the basin looking for marks. If you are running live imaging anyway, a purpose-built spinning rod like the Jig Head Minnow is designed around exactly that presentation — light enough to feel a jig tick rock, with enough backbone to drive a hook at distance. The rest of the jigs and blade baits matter less than where you put them. Bring four home instead of six if you like. You probably weren't going to catch six anyway.