Choosing a training method for your space, time and plant count

A single plant in a 0.8 × 0.8 m (2.6 × 2.6 ft) tent with six weeks of veg left, and eight small plants in a 1.2 × 1.5 m (4 × 5 ft) tent with three weeks left, are not choosing between the same eight techniques listed on every training round-up. They're answering two different questions, and the method that suits one would waste the other's time or blow its schedule. Four things decide which training method fits a given grow, roughly in this order: how much floor space is available, how many plants are in it, how many weeks of vegetative growth remain before the flip, and how much you can afford to get wrong if a cut or a tie goes badly. Work through them in that order and the technique picks itself more often than not.

Space sets the ceiling on what's worth attempting

A plant left untrained grows as one dominant cola with a scatter of smaller secondary sites below it, because the growing tip keeps suppressing the branches under it. Left alone, that imbalance shows up in the flower itself: in one large-plant trial, the lowest inflorescences ran up to 40% lower in cannabinoid concentration than the top of the canopy, a gap that showed up in the untreated control almost as much as in most of the pruning treatments tested, since cutting the top doesn't by itself get more light to the bottom ^[1]. Every training technique on this page works on that same dominance from a different angle, cutting it, tying it sideways or simply spacing plants apart, and how much room there is to do so shapes which techniques are worth the effort.

Under roughly 1 × 1 m (3 × 3 ft) of floor space per plant, a single plant is the realistic unit of training, and LST or light topping earns its keep: tying a branch down or removing one apex redistributes light across a canopy that's already within reach of the light source. A full mainlining or manifolding frame, which typically spreads eight to sixteen colas across one plant, wants more width than that footprint gives it; the extra colas end up shading each other rather than filling the tent, which defeats the point of building the frame at all. Sizing a tent to plant count and power covers the floor-area maths in more detail if the tent itself isn't decided yet.

Once the space runs to something closer to 1.2 × 1.2 m (4 × 4 ft) or larger with several plants in it, SCROG and SOG both become realistic, but for different reasons. SCROG needs the net more than it needs raw floor area: a screen strung roughly 20–30 cm (8–12 in) above the pots gives the plant something to weave through as it grows, and a wider, squarer tent lets the net actually spread growth sideways instead of just holding up one column of stems. SOG barely cares about individual plant architecture at all, since the technique's whole premise is many small, largely untrained plants standing shoulder to shoulder, so what it actually needs is enough floor area divided into that many pots with room for airflow between them, not headroom or width for a net.

Plant count flips the labour maths

Every training technique costs time per plant, and that cost either concentrates or spreads depending on how many plants are carrying it. One or two plants reward a detailed technique like mainlining or heavy multi-round LST, because the hour or two it takes to tie down or top a plant properly is time spent on the only plant, or one of only two, that will ever benefit from it. Run that same technique across eight or nine plants and the hour becomes eight or nine hours, repeated at every training session through veg, for a canopy that a lower-labour approach could have evened out almost as well.

That's the logic behind SOG: instead of training a few plants into one wide canopy, it runs many small, largely untrained clones or seedlings close together and lets plant density do the work training would otherwise do, trading per-plant attention for pot count. Density itself changes the yield-versus-uniformity trade: doubling planting density from one to two plants per square metre raised yield per area by 28–44%, depending on the architecture treatment, in one controlled trial, though the lowest inflorescences on the more crowded plants ran up to 90% below the top of the canopy in cannabinoid concentration, and both plant-to-plant and within-plant variation rose with it ^[2]. SOG accepts that unevenness deliberately: at high density, each plant contributes one small, reasonably uniform top cola rather than a tall structure with a strong top and a weak bottom, so density itself does some of the evening-out that SCROG or mainlining would otherwise need training to achieve.

None of this makes plant count the only lever on yield. In one factorial trial that varied light intensity, planting density and cultivar together, density and light intensity affected yield in a simple, additive way, but which cultivar was grown changed how much either one paid off, interacting with both rather than acting as a fixed bonus on top of them ^[3]. Training and spacing decisions interact with genetics and light; they don't replace them.

A single valuable mother plant or a limited run of a hard-to-replace cultivar sits at the opposite end from SOG's logic regardless of plant count: see the risk section below before deciding how much structure to put it through.

Remaining veg time is the hard constraint most guides skip

Every technique needs time to do two things after the intervention itself: recover, and then regrow into the shape the technique is meant to produce. Skip ahead to the flip before either has happened and the plant goes into flower still healing, or still short, which is usually worse than not having trained it at all.

LST needs almost none of that runway, because tying a branch down doesn't wound the plant or interrupt its growth; it can start the same week seedlings show their third or fourth node and keep going right up to the flip, or even a little into early flower, with no forced pause. Topping is the next smallest ask: the plant needs roughly one to two weeks after a cut heals before the two new tops it produces catch up to where the single top would have been, so topping wants at least two to three weeks of veg left to be worth doing at all. SCROG needs that same recovery repeated across several branches, plus the extra weeks it takes the canopy to actually fill the net edge to edge, typically three to five weeks all told. Mainlining or manifolding asks for the most: it's built from two or three sequential topping rounds, each needing its own one-to-two-week recovery before the next cut, so a frame that produces eight or sixteen even colas typically needs five to eight weeks of veg from seed or clone to be fully built before the flip. Recovery times after training breaks each of those numbers down technique by technique.

A grower with two to three weeks before a forced flip, whether that's a housemate's fixed schedule, a landlord's inspection date, or simply a plant that's already stretching, should default to LST or nothing rather than starting a technique it can't finish; a topping cut made a week before the flip leaves the plant recovering through the stretch phase instead of building flower sites, which usually costs more yield than skipping the cut would have.

Outdoor and greenhouse plants run on the same logic with a different clock: the calendar sets the flip instead of a timer switch, so a technique needs to be finished well before the days shorten enough to trigger flowering on their own, with less room to push training later if the season runs behind. The trade-off is worth it in one respect: outdoor and greenhouse plants that run the full season are usually larger to begin with, which is exactly the situation heavier techniques and real physical support (stakes, a trellis frame built to outlast wind rather than a tent's still air) pay off in, provided the structure goes in from the start rather than being added once the plant is already leaning.

Risk tolerance and plant value decide how much you can afford to get wrong

Not every cut carries the same risk, and the gap between a safe mistake and a costly one is bigger than most first-time trainers expect. In the large-plant trial cited above, topping a plant twice (a "double prune") raised CBDA yield to roughly 22.4 g per plant against 14.1 g on an untouched control, a real gain from a moderate, well-timed intervention. Removing a plant's primary branches entirely, a considerably more severe cut than topping, collapsed inflorescence yield by up to 99% in the same trial ^[1]. The lesson isn't that every cut is dangerous; it's that risk scales sharply with how much of the plant's existing structure a technique removes, and a technique that looks only a little more aggressive on paper can behave completely differently on the plant.

LST and SOG carry the least of that risk, in different ways: LST cuts nothing, so a badly placed tie can simply be repositioned or removed, and SOG barely trains the plant at all, so there's rarely a cut to get wrong in the first place. Topping and SCROG sit in the middle: a topping cut removes tissue permanently, but a missed or mistimed cut on one plant among several is a partial loss, not a total one, and a SCROG net doesn't wound the plant the way a cut does, though a screen mounted too low or filled unevenly can force an early flip or waste a season's light on bare squares. Mainlining carries the most risk of the five techniques compared here, not because any single cut is more severe than topping, but because the whole structure depends on several cuts landing correctly in sequence; a mistimed cut two-thirds of the way through the frame can leave an asymmetric plant that never recovers into the even structure the technique was built to produce.

Weigh that against what's actually on the tray. A cheap seed or clone of a common, easy-to-replace cultivar is a reasonable subject for a first attempt at any technique on this list, because losing it costs little more than the seed or the cutting itself. A single irreplaceable mother plant, a limited run of seeds from a cultivar that's hard to source again, or the only plant standing between a grower and an empty tent for the season are all better matched to LST or a light topping than to a full mainlining frame, regardless of how much veg time is available; the frame's payoff in yield uniformity isn't worth the plant if the sequence goes wrong partway through.

Five techniques side by side, and a tool to narrow it further

Read the table by whichever row matches the constraint that's actually binding: a forced flip date makes the first row decisive regardless of what the others say, while a single valuable plant makes the last row the one to weigh most heavily. The decision tool below asks the same questions in order and points to a starting technique.

Two worked paths, from tent to flip

For a fuller week-by-week reference across a whole veg period rather than just these two cases, see training schedules for photoperiod plants.

Path A: one plant, first grow, six-to-eight-week veg window

A single seedling in a 0.8 × 0.8 m (2.6 × 2.6 ft) tent, with no fixed flip date beyond "whenever it's ready," is the situation LST and topping were made for. A reasonable sequence:

• Weeks 1–3: grow the plant to four to six true nodes, then start LST, tying the main stem down at a shallow angle to open the centre of the plant to light.
• Around weeks 2–3: top once, just above a node with two healthy growth points below it, once the plant has recovered from the transplant and is visibly growing again.
• Weeks 4–6: keep tying new growth outward as it reaches for the light, so the canopy fills flat rather than tall.
• One to two weeks before the flip: a light lollipopping pass, clearing growth low enough on the plant that it won't receive useful light once the canopy closes in flower.

Every step in that sequence has at least one to two weeks of slack either side of it, which is exactly the margin a first grow needs, because a first attempt at tying or topping rarely goes perfectly on the first try.

Path B: several small plants, SOG-style, flip forced in three weeks

Eight small plants in a 1.2 × 1.5 m (4 × 5 ft) tent, with the flip date fixed three weeks out by a housemate's schedule, sit close to the opposite end of every decision above: many plants, short remaining time, and a layout where per-plant labour would eat the whole three weeks before it paid off. The right amount of technique here is close to none. Spend the three weeks on spacing and light instead: rotate pots so no plant shades another for long, keep canopy height even by moving faster-growing plants outward from the light's hottest spot, and skip topping or mainlining entirely, since neither has time to recover before the flip arrives. A light SCROG net, strung loosely across all eight pots rather than trained per plant, is the one addition worth considering in three weeks: it costs almost no per-plant labour and still pulls the canopy toward one even height as the plants stretch into flower, which un-netted SOG relies on density alone to achieve.

Layering two or three techniques is normal, not advanced-only

Most experienced home growers don't pick one technique from the list above and stop; they layer two or three, because each one solves a different part of the same problem. Topping breaks the plant's single dominant top and creates multiple main colas; LST redistributes those colas sideways so they all reach even light instead of shading each other; lollipopping removes the lower growth that would never catch useful light anyway. Run together on the same plant across a normal six-to-eight-week veg window, that combination is the default for most single-plant and few-plant home grows, not a step reserved for a second or third attempt.

The one real caution is timing, not technique count: don't stack two high-stress interventions in the same session. Topping and a heavy defoliation pass both send the plant into repair mode, and doing both at once asks it to recover from two wounds simultaneously instead of one, which usually shows up as a longer stall in growth than either would cause alone. Spacing a topping cut and a defoliation pass a week or more apart, and doing LST continuously in between rather than in separate sessions, gets the benefit of all three without compounding the recovery cost.

When the right answer is no training at all

Three situations turn every question above moot, because the plant isn't in a position to benefit from any technique yet. A plant recovering from a pest hit, a nutrient burn, or transplant shock needs to resume normal growth first; training adds a second stressor on top of one it hasn't dealt with yet, and the two compound rather than average out. A plant with less than the LST-level minimum veg time left, typically under a week, gets more value from being left alone than from a rushed tie-down that won't have time to hold its shape before the flip. Autoflowers running close to their own internal flip, rather than one triggered by a light switch, carry the least margin of all: their vegetative window is fixed by genetics and age, not by the grower's schedule, so a technique timed for a photoperiod plant's flexible flip date can just as easily land after an autoflower has already begun flowering on its own. When not to train covers each of these in more depth, including how to tell a stressed plant from a slow one before deciding either way.

If it's not obviously growing well this week, that's the answer: wait, don't train, and revisit the decision once new growth shows it's ready to carry the extra work.