Williams %R: A Fast Oscillator and Its Trade-Offs
Williams %R, created by Larry Williams, is a momentum oscillator that shows where the current close sits relative to the highest high over a lookback period, usually 14. It is the mirror image of the fast stochastic.
Values range from 0 to minus 100. Readings near 0 mean price closed near the top of its recent range, and near minus 100 means near the bottom. Common thresholds are minus 20 for overbought and minus 80 for oversold.
It is a fast indicator, reacting quickly to changes. That produces early signals but many false ones. In a trending market, it can sit in the overbought zone for long periods, so selling at minus 20 repeatedly loses money.
Better uses acknowledge this. In a ranging market, reversals from the extreme zones can offer entries near boundaries, with a stop beyond the range. In a trend, a dip into the oversold area in an uptrend, followed by a recovery above minus 80, offers a pullback entry in the trend's direction.
Because it is nearly identical in information to the stochastic, there is no need to use both. Choose one and learn its behavior.
As with all oscillators, the level of the indicator matters less than the context in which the signal appears, including trend, structure, and nearby levels.
Practical step: apply it to a trending and a ranging section of a chart, and count how many signals worked in each.
Williams %R: A Fast Oscillator and Its Trade-Offs
Williams %R: A Fast Oscillator and Its Trade-Offs
It’s Fairman 
Re: Williams %R: A Fast Oscillator and Its Trade-Offs
The pullback use in a trend is the one that makes sense to me. Dip into oversold, then back above minus 80, in the direction of the higher timeframe. Fewer signals, but they make sense. To make it concrete, if the daily chart on EURUSD is trending up, wait for the 1 hour Williams %R to drop below minus 80 during a pullback, then enter when it crosses back above, with the stop under the pullback low. Skip it if the pullback has broken the last higher low, because then the trend structure itself is in question. Test it on your pair before using it, since the 14 period default isn't always best.
It’s Fairman 
Re: Williams %R: A Fast Oscillator and Its Trade-Offs
Good call on not running it with stochastics. It's basically the same information flipped upside down. If you already like stochastics, there's no reason to add Williams %R. If you're starting fresh, either is fine, but pick one. A useful exercise is to put both on a chart for a week just to see how closely they match, then remove one. It's a good reminder that a lot of indicators are variations of the same calculation. The edge doesn't come from the indicator, it comes from the context you use it in and the discipline of following the same rules every time.
It’s Fairman 
-
FXS Prop Desk
- Posts: 1689
- Joined: Wed Oct 07, 2026 9:52 am
Re: Williams %R: A Fast Oscillator and Its Trade-Offs
Use it to time entries in the direction of the trend, for example buying when it reaches oversold during an uptrend, rather than selling when it is overbought.
Combining Williams %R with a trend filter, such as a moving average, can help reduce false signals.
Since Williams %R reacts quickly, many traders use a longer period on lower timeframes or confirm signals with a candle close back from the extreme.
For anyone comparing it with the stochastic, the two give nearly identical signals with the same period, so using both adds little.
Combining Williams %R with a trend filter, such as a moving average, can help reduce false signals.
Since Williams %R reacts quickly, many traders use a longer period on lower timeframes or confirm signals with a candle close back from the extreme.
For anyone comparing it with the stochastic, the two give nearly identical signals with the same period, so using both adds little.
FXS Official Team | The Idea Lab & Strategy
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Williams %R: A Fast Oscillator and Its Trade-Offs
The formula explains why it jumps around so much. With a 14 bar high of 1.1050, a low of 1.0990 and a close at 1.1038, the reading is the distance from the high, 12 pips, divided by the 60 pip range, times minus 100, which gives minus 20.
The problem appears when the lookback range is narrow. If the last 14 bars only span 12 pips, as often happens in a quiet Asian session, a 3 pip move shifts the reading by 25 points and it can swing from overbought to oversold on noise.
A simple fix is a minimum range filter. Ignore signals when the lookback range is below, say, half of the average range for that time of day, because the oscillator then measures spread and small ticks rather than momentum.
That filter also fits the trend pullback use described above. A real pullback in an uptrend usually happens inside a range of decent size, so the filter rarely removes the trades that matter there.
Has anyone tested whether the default of 14 bars works better or worse than a longer setting on the 5 minute timeframe? A longer lookback widens the range automatically, which might do part of the filtering on its own.
The problem appears when the lookback range is narrow. If the last 14 bars only span 12 pips, as often happens in a quiet Asian session, a 3 pip move shifts the reading by 25 points and it can swing from overbought to oversold on noise.
A simple fix is a minimum range filter. Ignore signals when the lookback range is below, say, half of the average range for that time of day, because the oscillator then measures spread and small ticks rather than momentum.
That filter also fits the trend pullback use described above. A real pullback in an uptrend usually happens inside a range of decent size, so the filter rarely removes the trades that matter there.
Has anyone tested whether the default of 14 bars works better or worse than a longer setting on the 5 minute timeframe? A longer lookback widens the range automatically, which might do part of the filtering on its own.
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Williams %R: A Fast Oscillator and Its Trade-Offs
The default 14 period has a side effect in the trend pullback use described above. After a strong leg the lookback holds the whole move, so reaching minus 80 needs price to give back about 80 percent of it.
Say EURUSD climbs from a low of 1.0900 to a high of 1.0960 over 14 hourly bars. Minus 80 on that range is a close at 1.0912, a 48 pip drop that would usually break the last higher low and cancel the trade anyway.
A 5 period setting only looks at the last five bars, whose low here is 1.0935, so a pullback close at 1.0938 already reads about minus 88 while the 14 period sits near minus 39, as the chart (illustration) shows. The shorter lookback flags the ordinary first dip that the default mostly misses.
The trade off is more signals, and some of them are the start of a real reversal rather than a pause. Requiring the dip to hold above the previous swing low keeps the extra ones honest.
So the period and the structure rule are really one decision. A long lookback asks for a deep pullback, the structure rule then rejects most deep pullbacks, and the combination leaves very few trades at all.
Say EURUSD climbs from a low of 1.0900 to a high of 1.0960 over 14 hourly bars. Minus 80 on that range is a close at 1.0912, a 48 pip drop that would usually break the last higher low and cancel the trade anyway.
A 5 period setting only looks at the last five bars, whose low here is 1.0935, so a pullback close at 1.0938 already reads about minus 88 while the 14 period sits near minus 39, as the chart (illustration) shows. The shorter lookback flags the ordinary first dip that the default mostly misses.
The trade off is more signals, and some of them are the start of a real reversal rather than a pause. Requiring the dip to hold above the previous swing low keeps the extra ones honest.
So the period and the structure rule are really one decision. A long lookback asks for a deep pullback, the structure rule then rejects most deep pullbacks, and the combination leaves very few trades at all.
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Williams %R: A Fast Oscillator and Its Trade-Offs
Last look is a practice where a liquidity provider gets a brief window, often a few milliseconds up to a couple hundred, to accept or reject a trade after the client has hit the quote. If price moved against the provider during that window, the trade can be rejected.
It exists to protect providers from latency arbitrage, where fast traders exploit stale quotes. The downside for clients is that orders in a fast-moving market can be rejected or re-quoted, usually when the client would have profited.
For scalpers the effect is direct. Rejections cluster in exactly the moments when momentum is strongest, so the best entries are the ones most likely to be declined.
Some brokers publish execution statistics including rejection rate and average hold time. Others do not mention it at all.
A practical question to ask any broker is whether its liquidity comes from last look providers, and what share of orders are rejected. The answer, or the refusal to answer, reveals a lot.
Has anyone dealt with rejection issues, and how did the broker explain them?
It exists to protect providers from latency arbitrage, where fast traders exploit stale quotes. The downside for clients is that orders in a fast-moving market can be rejected or re-quoted, usually when the client would have profited.
For scalpers the effect is direct. Rejections cluster in exactly the moments when momentum is strongest, so the best entries are the ones most likely to be declined.
Some brokers publish execution statistics including rejection rate and average hold time. Others do not mention it at all.
A practical question to ask any broker is whether its liquidity comes from last look providers, and what share of orders are rejected. The answer, or the refusal to answer, reveals a lot.
Has anyone dealt with rejection issues, and how did the broker explain them?