How do you actually choose a wick size without burning fifty test candles to find out?

The efficiency comes from testing in parallel rather than in series, which is the thing nobody tells beginners.

Do not pour one candle, test it, adjust, pour another. Pour three or four jars in the same batch with different wick sizes: same wax, same fragrance, same pour temperature, same day. Cure them together, burn them side by side.

One week and one afternoon, and you get a comparison rather than a data point. That is the difference between months and a fortnight.

How to run the burn:

Burn for four hours, which is what a customer does, not one hour.

Look at the melt pool at the two hour mark and at four. You want it reaching the glass by about two hours in an 8oz jar and staying a sensible depth.

Note the flame height and whether it flickers or mushrooms.

Weigh the jar before and after if you want a real number, consumption per hour is the most objective measure you can take and it separates two wicks that look similar.

Then burn again the next day, because a lot of problems only appear on the second and third burn as the pool deepens.

One more thing that saves a season: change one variable at a time. People test a new wick and a new fragrance together and learn nothing.

Wick size is set by how much wax the flame has to melt, and everything else is an adjustment to that. In order of how much each factor matters:

Jar diameter, by a long way. The flame has to produce a melt pool that reaches the glass. Diameter is the whole job: a taller jar of the same diameter needs the same wick, which surprises people.

The wax. Soy, paraffin and blends have different melt points and different viscosities, so the same wick behaves differently. A chart for one wax does not transfer to another.

Fragrance load. Fragrance oil thins the wax and adds fuel, and above about six percent it starts genuinely changing the burn. This is real and it is smaller than people think - usually one wick size, not three.

Colour. Dye is the least important, and pigment-heavy colours can clog a wick. Liquid dye at normal amounts is close to irrelevant.

Wick material. Cotton, cored and wood all behave differently and are not interchangeable by number.

So the practical starting point: pick the wick series for your wax, look up the size for your diameter, and treat that as the middle of your test range rather than as the answer. The chart is a starting point that manufacturers publish knowing you will adjust.

How to read what you already have, since your four failures are informative rather than wasted.

Tunnelling, pool does not reach the glass. Wick too small for the diameter. Go up. Note that tunnelling on the first burn is often the user rather than the wick, since a first burn must be long enough to reach the edge; but tunnelling repeatedly after a proper first burn is the wick.

Drowning - the wick sinks into a deep pool and goes out. Wick too small, or the pool is getting too deep because it is burning down faster than out. Usually up a size.

Smoking and sooting. Wick too large, or too long. Trim to about five millimetres before every burn and retest before changing size: a lot of soot complaints are trimming, not wicking.

Large flame, pool over about a centimetre deep, glass too hot to hold. Too big. Go down.

So two of your four are pointing up and two are pointing down, which suggests your starting point is roughly correct and the wax or fragrance is the variable.

On changing fragrance: you do not start again. Test the new fragrance at your established wick and one size up. Two candles, not four. Most fragrances land on the same wick and the ones that do not are usually the heavy vanilla-type oils, which are the known troublemakers.