Accelerator vs Awesome Oscillator: AC Turned First on 53% to 56% of Swings, but Price Had Already Turned 2 Times in 3

In Bill Williams’ system the two oscillators sit one under the other. The Awesome Oscillator is meant to show momentum, the Accelerator Oscillator its change, so the AC should turn first and give the earlier signal. That is the usual reason to put both on a chart. This page checks whether the AC is really earlier, how much noise it pays for that, and which of the two traded better on the same bars.

How are the AC and the AO different on paper?

An AC zero cross is the moment the AO crosses its own 5-bar average, the same trick as the MACD signal line. Taking away the average leaves only the short wiggle of the AO, which is why the two lines look alike in shape but hardly move together. The full test of the AC alone is in Accelerator Oscillator tested.

How was it compared?

• Data: EURUSD M5 from June 3 to August 26, 2026, 16,926 bars; gold M5 from June 16 to August 26, 13,736 bars. M15 built from the same bars.
• Timing: for every AC colour change I looked for an AO colour change in the same direction within 6 bars and noted which came first. A price turn is a close that is the lowest or highest of 6 bars on each side.
• Trading: AO zero cross, saucer and colour change against the five AC rules, with the exit from the forex indicators list, ranked: stop 1.5 ATR, target 3 ATR, 48 bars. Cost 1 pip on EURUSD and $0.30 an ounce on gold. A pass needs a profit after cost in both halves and a result above the best 5% of 200 random-entry runs.

Is the AC really faster than the AO?

Yes, on timing the textbook is right. When both oscillators turned around the same swing, the AC was first more than half the time and last only about one time in eight. The lead was short, 1 to 2 bars, which is 5 to 10 minutes on M5.

The 0.19 correlation is the surprise. Two lines made from the same AO, where one usually turns first, still went their own ways most of the time, because the AC throws away the AO’s level and keeps only its last few bars of change.

How much extra noise does the AC carry?

The AC changed colour about 30% more often and crossed zero almost three times as often. Its colour changes were also more often close to a real turn, 7 to 11 points more than the AO’s, so the extra changes were not all junk. But on gold, about two in three AC colour changes still came nowhere near a turn.

Do they turn before price?

Here the order Williams described, the AC first, then the AO, then price, did not hold. The order on our bars was price, then the AC, then the AO. Both are averages of past bars, and an average cannot turn before the bars it is made of. The AC is the less late of the two, by about one bar. The same finding in a different pair is in Awesome Oscillator vs MACD, where the “faster” AO beat MACD to the zero line only about half the time.

Which one traded better?

Trades in brackets. The best AC rule on each chart (including the two EMA 200 and AO-side variants from the main test) lost to the best AO rule on all 4 charts, by almost nothing on EURUSD M5 and by more than a pip on EURUSD M15. Colour change against colour change, the earlier AC won on 2 charts and lost on 2: being one bar earlier did not show up in the result.

Still, the AO’s two profitable runs did not pass either. The saucer on gold M15 made money in both halves, +$3.04 and +$0.56 an ounce a trade, but the best 5% of random runs with the same number of trades made $3.76 or more. Gold trended in this sample, and that lifted every long-biased entry.

Which one should you use?

• If you want one Bill Williams oscillator, keep the AO. It was calmer, made fewer trades, and its best rule did better than the best AC rule on every chart here.
• Use the AC only as a hint that the AO is about to turn. That part worked: the AC was first in more than half the cases. As a signal of its own it lost on 19 of 20 runs.
• Neither of them warns before price. Both turned after the close did, and the AO by 2 to 3 bars.
• Both work as part of a system, not alone. Williams used them with the Alligator, and our indicator combos test shows what stacking filters does and does not fix.

Frequently asked questions

Is the Accelerator Oscillator better than the Awesome Oscillator?

Not in this test. The best AO rule lost less, or made more, than the best AC rule on all 4 charts. The AC turned earlier, but that did not turn into better trades.

Does the AC give earlier signals than the AO?

Yes, by a little. It changed colour first on 53% to 56% of matched turns and the AO first on 11% to 14%, with a median lead of 1 to 2 bars.

Can you use the AC and AO together?

Williams did. Taking his AC rule only on the AO’s side of zero cut the trades by a quarter to a third, but it still lost on all 4 charts, from $0.58 an ounce on gold M5 to 2.29 pips on EURUSD M15.

Why does the AC cross zero so often?

Its zero line is the AO’s own 5-bar average, so every small wobble of the AO above or below that average is a cross. That gave 31 to 36 crosses a day on M5.

Are AC and AO the same indicator?

They share a formula, but on our charts their values correlated at only 0.19. The AO follows the trend of the averages; the AC follows the last few bars of change in the AO.