Dishing hurts wheel strength and durability. Besides reducing dish or using heavier components, another approach which is sometimes used is asymmetrical lacing: either more spokes on one side, or heavier/thicker spokes on one side.

The following discussion uses "cone side" and "flat side" rather than "left" and "right", because rear wheels are dished one way and front disk brake wheels are dished the other. Here, "flat" means the side where spokes are most nearly in a plane, the right or sprocket side of a rear wheel, or the left or disk side of a front disk-braked wheel.

For starter, two simplifying assumptions:

• From a spoke stiffness perspective, more spokes has the same effect as thicker spokes: either one means more total metal, and thus a stiffer spring. In practice, the two approaches are not quite the same due to slightly different load distribution, but in many respects the two close enough to treat the same.
• Unless noted otherwise, the following discussion treats more/thicker spokes on one side the same as to fewer/thinner spokes on the other. In practice, more/thicker spokes builds a heavier wheel and also has more stiffness relative to the rim, so there are some detail differences.

Several qualitative (or hand-wavy quantitiatve) observations:

• When spokes start going slack, the load is close to the load that causes wheel collapse.
• Radial loads cause stiffer spokes to go slack at less displacement, and thus spread less load along the rim.
• For typical rims and spoke counts, moving material between the flat and cone sides (either via changing spoke diameter or moving spokes from one side to the other) does not change the net radial stiffness of the spokes, and therefore does not change radial load sharing. This argues that moving material from the cone side to the flat side is okay, but adding material hurts load sharing and thus radial strength.
• Intuitively, we would like the same lateral stiffness on both sides of the wheel. If spokes are half as steep on the "flat" side, lateral wheel motions have rougly twice the leverage against "flat" spokes as against "cone" spokes. Thus, we intuitively want spokes twice as stiff on the "flat" side, so lateral loads are equally supported by both sides. This argues for twice the spoke material on the flat side.
• Cone side spokes are at lower aggregate tension, so it takes a smaller change in absolute tension to make cone-side spokes go slack than it does flat-side spokes. Thus, we want more flexible spokes (fewer or thinner) on the "cone" side so more loads towards the cone side are handled by the flat side rather than by detensioning the cone side. This also argues for more material on the flat side.
• If spoke material is moved from cone-side to flat-side, then overall leverage is worse. That is, if the bracing angle on the flat side is half the angle of the cone side, then a gram of spoke material on the cone side has twice the lateral stiffness of that same gram moved to the flat side. Lower stiffness tends to be good for radial strength, but higher stiffness tends to be good for lateral strength This argues that moving material from cone side to flat side hurts lateral support.
• Flat-side spokes are under higher aggregate tension due to their lower bracing angle. With an equal number of cone-side and flat-side spokes, the tension per spoke is thus higher for the flat-side spokes. One limit to total spoke tension on a wheel — and thus one limit to it's total radial strength — is very high spoke tension pulls the spoke out of the rim bed or hub flange. Using more spokes on the flat side divides the total flat-side tension among more spokes and thus circumvents hub and flange strength limits, allowing greater total spoke tension and thus greater radial strength. This argues that increasing the number of spokes on the flat side can increase radial strength.
• On the other hand, if the spoke bed is already strong enough, higher spoke tension simply causes rim collapse, and is thus of no use. In that case, moving material from cone side to flat side reduces lateral support and thus reduces the total rim compression before the rim buckles: a rim with good lateral support can take more rim compression than the same rim with poor lateral support. Lower rim compression means lower spoke tension, which in turn means a lower radial load before spokes start going slack. This argues that moving spoke material from cone-side to flat-side can reduce total spoke tension and thus radial strength.

To summarize so far:

• If spoke bed or hub flange strength is the limit on total spoke tension, more total spokes is preferred (use thinner spokes to achieve the same weight). If more spokes is not an option, putting more spokes on the flat side may increase total spoke tension and thus build a stronger wheel.
• If spoke bed and hub flange strength are not limits, moving some material from cone side to flat side can increase the radial displacement before cone-side spokes go slack and thus increase the radial load at which cone-side spokes go slack.
• However, this must be balanced against the worse bracing angle of material on the flat side, which reduces the lateral load at which cone-side spokes go slack.

A next question is: what is the magnitude of asymmetrical lacing and spoking? I do not have complete numbers, but do have some qualitative observations:

• Asymmetrical lacing patterns lead to gross differences. As example, a 36-spoke wheel is normally laced 18/18; it can be laced asymmetrically to 12/24. A 32-spoke wheel can be laced 8/24 or 1:3. A 28-spoke wheel can (with certain rims) be laced radial/crow's-foot to 7/21 or 1:3 (provided the cone side is laced radially). A 24-spoke wheel can be laced 8/16 or 1:2. And so on. This shows that asymmetrical left/right patterns lead to large changes — which may be more than is desired.
• Changes due to spoke diameter are smaller, but perhaps not as small as they first appear, because spoke cross-section is squared in radius/diameter.

  	2.0	1.8	1.6	1.55	1.5
  2.0	100%	81%	64%	60%	56%
  1.8		100%	79%	74%	69%
  1.6			100%	94%	88%
  1.55				100%	94%
  1.5					100%

  So, for example, a 1.6mm spoke has 79% of the stiffness of a 1.8mm spoke.

Asymmetrically-laced wheels have been used for decades. What is the anecdotal or "field" evidence?

• Some users report asymmetrical wheels are stiffer, stronger, and/or more durable. However, the advantage, if any, seems relatively small for common wheels: historically, light wheels built asymmetrically are not vastly stronger or mor durable than symmetrical wheels. This is probably due to the 'race" between factors above.
• More recently, it has become common to build wheels using a small number of spokes to reduce air drag. (A heavier rim is needed for equivalent structure, due to the larger distance between spokes. Thus, total wheel weight is not reduced, despite lower spoke weight.) With fewer spokes, the tension on each spoke must be higher to get the same total spoke tension, rim compression, and load-carrying capacity. For these wheels, spoke bed strength is a limit to spoke tension, and so asymmetrical lacing is advantageous to spread the flat-side load among more spokes.

To conclude:

• As a rule of thumb, the strongest high-dish wheel of a given weight will be weaker than the strongest low-dish wheel of the same weight.
• In specific cases, asymmetrical lacing can be used to reduce the loss of strength compared to a low-dish wheel and improve strength over an otherwise similar high-dish wheel.
• Specifically, asymmetrical lacing patterns help wheels with few spokes, stiff rims, and significant dishing — by allowing higher total spoke tension without suffering spoke pull out from the rim or hub.
• For a wheel with more spokes and more flexible rim, asymmetrical spoke counts or spoke diameters may be of little benefit or even a net loss because the lateral support provided by a gram's worth of spoke material on the cone side is greater than if the gram is moved to the flat side.
• Even with many spokes and and a flexible rim, a very highly dished wheel may still benefit from asymmetrical spoking because cone-side spoke tension is so low that common loads cause complete loss of cone-side tension and because the flexible rim precludes higher total rim tension. Asymmetrical lacing may thus increase the radial load at which spokes go slack.
• However, even with asymmetrical lacing or spokes, such a wheel is still weak.
• Therefore, the main benefit of asymmetrical patterns is for low-spoke-count wheels with significant dish.
• And for the kinds of wheels covered in this class (many spokes, offset rear spoke bed, medium or better rim stiffness, not-to-severe dishing) asymmetrical spoking is not typically of value.

To summarize the summary: for a low-spoke-count wheel with a stiff rim, using more of the spokes on the flat side can win because it allows higher total spoke tension without suffering spoke pull out from spoke bed or hub flange. For high-spoke-count wheels with a relatively flexy rim, using asymmetrical spoke counts or asymmetrical spoke diameters may help keep cone-side spokes from going slack, but that must be balanced against lateral strength reduction that happens because lateral bracing of a gram's worth of spoke on the flat side is worse than the same gram of spoke material on the cone side.