A price is an opinion. A comp record is evidence. The difference between guessing what a Shepard Fairey print is worth and knowing what it is worth is not access to secret data — it is a disciplined method for reading the data that is already public. This post gives you that method: the exact framework Gauntlet Gallery uses to value any Fairey print, walked end to end on two real records so you can run it yourself.
Across this series we have looked at the Fairey market from the top down — by the numbers, by edition size, by tier, by liquidity. This installment inverts the lens. Instead of asking what the market does in aggregate, we ask a narrower and more useful question: given one specific print, with one specific pile of past sales attached to it, what is a defensible number to put on it today? That is the question every buyer, seller, and insurer actually needs answered, and it is the one that separates a collector who overpays from one who does not.
The framework rests on our master comps database of 1,004 catalogued Fairey works, of which 595 carry enough pricing signal to price and 79 have a clean, deep enough sales record to serve as reference comps. Those 79 records aggregate 3,509 individually tracked clean sales. Every number below is drawn from that database, and every methodological choice — the floor, the window, the averaging method, the sources — exists to solve a specific way that raw comps lie to you. Learn the choices and you learn the trade.
Why raw comps mislead — and what the framework fixes
The naive way to value a print is to search a marketplace, find the last few prices, and average them. That method fails in four predictable ways, and each rule in our framework is a direct patch for one of those failures.
First, marketplaces mix apples with oranges. A single title like "Peace Goddess" spawns a signed-and-numbered edition, an artist-proof, an unsigned open version, colorway variants, and the occasional damaged copy sold cheap. Average them together and you get a number that describes nothing. The fix is a hard floor: for signed and numbered Fairey prints we do not count any sale below $90. Anything under that floor is almost certainly an unsigned poster, a print of a print, or a misdescribed listing — noise that drags the average down and misrepresents the asset you actually hold.
Second, old prices describe an old market. Fairey's secondary market of 2013 is not the market of 2026; demand, supply, and collector attention have all moved. The fix is a rolling ten-year window. We weight and trend against sales inside a 2016–2026 span and treat anything older as historical context, not as a live comp. A 2011 auction result is a data point about history, not about what a buyer will pay next week.
Third, outliers dominate small samples. One euphoric auction night, one desperate fire-sale, and a mean built on a dozen sales lurches wildly. The fix is a trimmed mean: we strip the extreme high and low tails before averaging so a single $2,000 anomaly cannot masquerade as the going rate. Where the sample is thin or skewed, we lean on the median — the middle sale — which by construction ignores the tails entirely.
Fourth, a single source is a single point of view. eBay skews retail and current; auction records skew high and occasional; archival databases skew historical. The fix is multi-source aggregation. We pull from WorthPoint, eBay, LiveAuctioneers, and Heritage, then reconcile them, because triangulating across four vantage points is the only way to see a print's true trading range rather than one channel's slice of it.
Those four rules — floor, window, trimmed mean, multi-source — are the entire philosophy. Everything that follows is mechanics: how those rules turn a raw sales list into the fields you see on a Gauntlet Gallery comp record, and how you read those fields to reach a number.
Anatomy of a comp record
Every reference print in our database resolves to a compact record. Learn to read one record and you can read all seventy-nine. Here is the full anatomy, field by field, using the Lenin Record (a 2005 print, edition of 300) as the specimen on the table.
| Field | Lenin Record value | What it tells you |
|---|---|---|
| Sale count (n) | 16 | How much evidence exists. Sixteen clean sales is a usable but modest sample — enough to trust the median, thin enough that the mean is fragile. |
| Min | $124.99 | The floor of observed reality. The cheapest a clean copy has changed hands. Useful as a "you should never pay less to buy, never expect less to sell" anchor. |
| Max | $2,005 | The ceiling of observed reality — a single euphoric or exceptional sale. Aspirational, not typical. Sixteen times the minimum. |
| Average (trimmed mean) | $899.94 | The tail-trimmed arithmetic center. Still pulled upward here by high sales — note it sits well above the median, a red flag we read below. |
| Median | $525 | The middle sale. The single most robust "typical price" figure, immune to the tails. When mean and median diverge, trust this one. |
| Recent average | $409.80 | The average of the most recent cluster of sales. The trend signal — where the market is now, not where it has been on average. |
| Recent count | 5 | How many sales feed the recent average. Five is enough to notice a direction, too few to bet the house on it. |
| eBay count / eBay average | 39 / $306.59 | The retail-channel view. Deeper sample (39) but a lower center — eBay skews toward accessible, current, buyer-friendly pricing. |
| Price history span | 2018-12-12 → 2026-01-06 | The date range the record covers. Confirms the comps sit inside our live ten-year window. |
| Price history | 9 dated, sourced sales | The raw evidence itself — each with a date, a price, and a source. This is what everything else is computed from. |
| Suggested price | $356.25 | Our derived, conservative fair-ask — a blend that leans on recent and retail signal, floored and discounted for liquidity. |
| Price range | $250 – $650 | The defensible band. Below the low you are leaving money on the table; above the high you are asking the market to make an exception. |
| Recommendation | HOLD | The action tag. Given a soft recent trend against a higher historical average, the record says wait rather than sell into weakness. |
Notice that the record contains four different "prices" — min, max, mean, median — plus two trend prices (recent average, eBay average) and two synthesized prices (suggested price, price range). A beginner sees eight numbers and freezes. The framework tells you which number to trust, when, and why. That is the whole skill.
Why median beats mean here
Return to the Lenin Record's two center-of-gravity figures: a trimmed mean of $899.94 and a median of $525. Those describe the same sixteen sales, yet the mean is 71.4% higher than the median (arithmetic on the brief: $899.94 versus $525). That gap is not a rounding artifact — it is the signature of a right-skewed distribution, where a handful of exceptional sales (up to the $2,005 max, sixteen times the $124.99 min) drag the average upward while most sales cluster far lower.
When mean and median diverge this violently, the median wins, for a simple reason: the mean answers "what would the average sale be if you could repeat all sixteen," which nobody is doing, while the median answers "what does the typical single copy sell for," which is exactly the transaction you are about to make. You are selling one print, once. You want the price the middle of the market pays, not a figure inflated by nights that will not repeat. A trimmed mean helps — it has already removed the most extreme tails — but with a sample this skewed even the trimmed mean overstates the center. The median is the honest number.
The rule generalizes: the wider the mean-median gap, the more you should trust the median and discount the mean. On a tightly clustered print — where nearly every sale lands within a narrow band — mean and median nearly coincide and either serves. On a wild one like the Lenin Record, the median is your anchor and the mean becomes a diagnostic: its distance above the median measures how top-heavy and unreliable the raw average is.
The football field: min, median, max, recent
Numbers in a row are hard to feel. The single most useful thing you can do with a comp record is lay its key figures out as a horizontal band — what analysts call a "football field" — and read the shape. Four marks matter: the min (left goal line), the max (right goal line), the median (the honest center), and the recent average (where the ball is now). Here are the two worked examples laid out that way.
| n | Min | Median | Recent avg | Max | Field width (max÷min) | |
|---|---|---|---|---|---|---|
| Lenin Record (2005, ed. 300) | 16 | $124.99 | $525 | $409.80 | $2,005 | 16.0× |
| Peace Goddess (2006, ed. 300) | 31 | $174.12 | $700 | $849.42 | $2,006 | 11.5× |
The width of the field is your first read. Both prints have enormous spreads — Lenin's ceiling is 16.0 times its floor, Peace Goddess's is 11.5 times (arithmetic on the brief). A wide field means high dispersion: condition, colorway, signature, and timing swing the outcome dramatically, so which copy you have and when you sell matter as much as the title. A narrow field would tell you the opposite — a commodity-like print where any clean copy fetches roughly the same. Neither print here is a commodity.
The second read is where the recent average sits relative to the median. This is the trend, and the two prints tell opposite stories from nearly identical geometry.
On the Lenin Record, the recent average of $409.80 sits 21.9% below the $525 median (arithmetic on the brief). The ball is on the wrong side of the field — recent buyers are paying less than the historical middle. Worse, that recent figure is 54.5% below the $899.94 trimmed mean. The market is cooling off a print whose long-run average was inflated by past highs. That is precisely why the record carries a HOLD tag: selling today means selling into a soft patch, below the median, far below the average. The disciplined move is to wait or price at the conservative $356.25 suggestion if liquidity is the priority.
On Peace Goddess, the geometry inverts. The recent average of $849.42 sits 21.3% above the $700 median and even edges 2.6% above the $828.05 trimmed mean (arithmetic on the brief). Here the ball is deep in scoring territory — recent sales are outrunning both the historical center and the historical average. Momentum is positive. A seller has the wind at their back; a buyer should expect to pay a premium to the median, not a discount.
Same title family, same era, same edition of 300, near-identical fields — and opposite trend signals. That contrast is the entire argument for the framework: you cannot see it from a headline average. You only see it when you read the recent average against the median inside the field.
Reading recent-average versus all-time average
The relationship between the recent average and the all-time trimmed average is the record's trend gauge, and it reads cleanly in three states:
- Recent well below all-time (cooling). Lenin Record: recent $409.80 versus all-time $899.94, a 54.5% gap. The print traded richer in the past than it does now. Read: patience, or a discount if you must sell.
- Recent roughly equal to all-time (stable). When the two converge, the print is trading at its own long-run center — a mature, predictable comp. Value with confidence near the median.
- Recent above all-time (heating). Peace Goddess: recent $849.42 versus all-time $828.05, +2.6%. The market is bidding the print above its own history. Read: momentum; a seller can reach toward the upper band.
One caution the sample sizes force on you: Lenin's recent figure rests on just five sales and Peace Goddess's on six. A trend built on five or six data points is a whisper, not a shout. It is enough to tilt a decision at the margin — enough to justify a HOLD over a SELL — but not enough to override the median as your anchor. The median is built on the full sample; the recent average is a small, current overlay. Weight them accordingly.
Reading the raw price history, sale by sale
Every summary figure on the record — the min, the median, the recent average — is a compression of the raw price history, and it pays to open that history and read the individual sales, because the sequence often tells you something the summary cannot. The Lenin Record's history contains nine dated, sourced sales spanning December 2018 to January 2026: $249.99 (WorthPoint, 2018), $400 (2021), $299 (2021), $650 (2021), $650 (2024), $199 (2024), $375 (2025), $124.99 (eBay, 2025), and $700 (eBay, 2026). Read straight down that column and the print's personality appears.
| Date | Price | Source | Read |
|---|---|---|---|
| 2018-12-12 | $249.99 | WorthPoint | The historical anchor — near the bottom of the range, early in the window. |
| 2021-10-31 | $650 | WorthPoint | A 2021 strength cluster; the print reached toward its upper band. |
| 2024-06-09 | $199 | WorthPoint | A soft print — the kind of low sale that pulls the recent average down. |
| 2025-07-03 | $124.99 | eBay | The record low, and the eBay channel doing what it does: retail, buyer-friendly, at the floor of reality. |
| 2026-01-06 | $700 | eBay | The most recent print — a hard bounce above the median, but a single event. |
Two lessons fall out of the sequence. First, the volatility is genuine, not an artifact: within eighteen months the print traded at $199, then $124.99, then $700 — the $124.99-to-$700 swing is a 460% move between two adjacent sales (arithmetic on the brief: $124.99 to $700). That is the football field's 16.0× width made concrete, and it is why no single recent sale can be trusted as "the price." Second, the source tags matter: seven of the nine sales are WorthPoint (the archival, historical view) and only two are eBay — yet those two eBay sales bracket the extremes, the $124.99 low and the $700 high. A thin channel sample can swing wide, which is exactly why the record also carries a separate, deeper eBay average of $306.59 across 39 listings. When you read a raw history, always separate the signal (the clustered middle) from the theatre (the isolated extremes), and always note which source produced which.
The comparison across the full window is instructive too. From the earliest tracked sale ($249.99 in 2018) to the latest ($700 in 2026), the print is nominally up 180% (arithmetic on the brief) — but that headline is meaningless in isolation, because it connects two single points at opposite ends of a violently noisy series. This is the trap the framework exists to prevent: a first-and-last-point calculation on an illiquid asset produces a number that looks authoritative and means nothing. The median and the recent average, built on the whole distribution, are the honest summaries. The endpoints are just the endpoints.
Sample size and how much to trust the record
Not all seventy-nine reference records deserve equal confidence, and the framework builds that humility in through the sale count. The rule of thumb: the smaller the n, the wider your error bars and the more you lean on the range rather than a point estimate.
Peace Goddess, at n=31, is a deep, trustworthy record — its $700 median is built on enough sales that you can quote it with confidence. The Lenin Record, at n=16, is usable but thinner; the median holds, but the trend (five recent sales) is a soft overlay. Contrast both with the cautionary case buried in the same data: a "Peace Goddess Gold" variant carries an n of just 1 — a single $999 WorthPoint sale from 2018. One sale is not a market; it is an anecdote. On a record like that, the framework refuses to pretend precision, defaulting to a conservative synthesized figure (its suggested price lands far below that lone $999 print) and a wide range, because a single old sale tells you almost nothing about what a copy clears for today. When you encounter a print whose entire "record" is one or two sales, treat the number as a starting hypothesis to be tested against fresh listings, never as a valuation.
This is why the database separates the 595 priced prints from the 79 clean reference records. Having a price is not the same as having a trustworthy price. The 79 earned their status by depth of evidence. When you value a print that sits outside that set, you are doing genuine primary research — and you should price it with the corresponding caution, quoting a range, flagging the thinness, and refusing to let a lonely data point masquerade as a market.
How suggested price and price range are derived
The suggested price and price range are not raw observations — they are the record's two synthesized outputs, the point where the framework stops describing and starts recommending. Understanding how they are built keeps you from misusing them.
The price range is the defensible band — the interval inside which a clean copy can be bought or sold without either party needing the market to make an exception. It is anchored below the median-and-recent center and reaches up toward, but rarely to, the historical highs. For the Lenin Record it is $250–$650; for Peace Goddess, $150–$575. Read the low end as "below this you are leaving money on the table as a seller" and the high end as "above this you are asking a buyer to pay a premium the recent record does not yet support."
The suggested price is deliberately more conservative than a naive read of the median would produce. Lenin's suggestion is $356.25 — below its $525 median and below its $409.80 recent average — and Peace Goddess's is $304.23, well below both its $700 median and its $849.42 recent average. Why so cautious? Because the suggested price answers a specific question — "what should we ask to move this print with confidence" — and it bakes in three haircuts the headline median ignores: a discount for liquidity (thinner prints must price to sell), a lean toward the retail eBay channel (Lenin's eBay average is $306.59, 41.6% below its blended median, and that channel is where most actual transactions clear), and respect for the floor and the recent softness where present.
The practical takeaway: use the median when you want the honest typical value (for insurance, for a fair-market estimate, for a private sale between informed parties), use the price range when you want the negotiating envelope, and use the suggested price when your goal is a confident, liquidity-aware ask that will actually clear. They answer different questions. Reaching for the wrong one is the most common valuation error we see.
Multi-source aggregation in practice
The single most underrated field on the record is the source split between the blended figures and the eBay-only figures, because it quietly reveals channel bias. On the Lenin Record, the blended median is $525 but the eBay average is $306.59 — the retail channel runs 41.6% below the blended center (arithmetic on the brief). On Peace Goddess, the blended median is $700 while the eBay average is $452.62, a 35.3% gap the same direction.
That consistent direction is the lesson: eBay reliably prints lower than the auction-and-archive blend. It is a retail, buyer-friendly, high-liquidity venue; auction houses and archival records like LiveAuctioneers and Heritage skew toward higher, occasional, headline results, and WorthPoint aggregates the long historical tail. If you value a print off eBay alone you will systematically underprice it; if you value it off auction records alone you will systematically overprice it. The blend — four sources reconciled — is the only view that sees the whole animal. When you read a comp record, always check whether the number you are leaning on is a blended figure or a single-channel figure, and adjust for the known bias of that channel.
The repeatable checklist: valuing any Fairey print
Here is the framework compressed into an eight-step checklist you can run on any Fairey print with a sales record. Run it in order; each step feeds the next.
- Confirm the asset. Is it signed and numbered? If so, the $90 floor applies — discard any "comp" below it as almost-certainly-not-your-asset. Identify the exact colorway and edition; a variant is a different comp.
- Gather multi-source sales inside the window. Pull sales from WorthPoint, eBay, LiveAuctioneers, and Heritage, and keep only those inside the rolling 2016–2026 window. Older sales become context, not comps.
- Count the evidence (n). A double-digit n supports a trustworthy median; a low single-digit n means you value with humility and wide error bars.
- Trim and find the center. Strip the extreme high and low tails, then read both the trimmed mean and the median. Note the gap between them.
- Trust the median when mean and median diverge. A wide gap (like Lenin's 71.4%) means a skewed, top-heavy distribution — anchor on the median, treat the mean as a skew diagnostic.
- Lay out the football field. Plot min, median, max, and recent average. Read the field width for dispersion and read the recent average's position for trend.
- Read recent versus all-time for direction. Recent below all-time = cooling (lean HOLD/discount). Recent above all-time = heating (lean toward the upper band). Weight the trend by its recent count.
- Synthesize the output. Pick the median for a fair estimate, the price range for a negotiating envelope, or a conservative liquidity-aware suggested price for a confident ask — and adjust for channel bias (eBay low, auction high).
Eight steps, no guesswork, every number sourced. That is the difference between an opinion and a valuation.
Worked example, start to finish: the Lenin Record
Let us run the full checklist on the Lenin Record in plain sequence, so you can see the machine turn.
Step 1 — asset. Signed and numbered, edition of 300, a 2005 work. The $90 floor applies; we discard sub-floor noise.
Steps 2–3 — evidence. Sixteen clean sales across a 2018–2026 span, blended from WorthPoint and eBay, all inside the live window. Sixteen is a usable sample — enough to trust the median, thin enough to keep the mean at arm's length.
Steps 4–5 — center. Trimmed mean $899.94, median $525, a 71.4% gap. That gap flags a skewed distribution, so we anchor on the $525 median and treat the mean as evidence of top-heaviness, not as the price.
Step 6 — field. Min $124.99, max $2,005 — a 16.0× field, very wide, telling us condition and timing swing this print hard. Recent average $409.80 sits on the low side of the field, below the median.
Step 7 — trend. Recent $409.80 is 21.9% below the median and 54.5% below the all-time mean. The print is cooling. With only five recent sales, that is a soft signal — but it points one direction, and it is down. Hence the HOLD tag: do not sell into this softness unless liquidity forces your hand.
Step 8 — output. For insurance or a fair private estimate, use the $525 median. To negotiate, work the $250–$650 range. To sell with confidence and speed, ask near the conservative $356.25 suggested price, which respects the eBay-channel reality ($306.59) and the recent cooling. Same print, three legitimate numbers — each correct for its own purpose.
Worked example, start to finish: Peace Goddess
Now the same machine on Peace Goddess, where the trend flips.
Step 1 — asset. Signed and numbered, edition of 300, a 2006 work; floor applies. (Note the record also carries colorway siblings — Gold, Red, Burgundy — each a distinct comp; do not blend a Burgundy sale into a base Peace Goddess valuation.)
Steps 2–3 — evidence. Thirty-one clean sales across 2018–2026, blended from WorthPoint and eBay. Thirty-one is a deep sample — the median here is robust.
Steps 4–5 — center. Trimmed mean $828.05, median $700, an 18.3% gap. Narrower than Lenin's, but still skewed enough that the median leads. Anchor: $700.
Step 6 — field. Min $174.12, max $2,006 — an 11.5× field, wide but tighter than Lenin's. Recent average $849.42 sits high in the field, above the median.
Step 7 — trend. Recent $849.42 is 21.3% above the median and 2.6% above the all-time mean. The print is heating. Six recent sales is still a modest base, but the direction is clearly up — the market is bidding Peace Goddess above its own history.
Step 8 — output. A fair estimate anchors on the $700 median, but the positive trend justifies reaching toward the upper band; a seller with a clean copy and time can test above $700 given recent prints near $850. The published $150–$575 range and $304.23 suggested price are the conservative, liquidity-first envelope — deliberately cautious to guarantee a clear, and a reminder that the "ask to move fast" number and the "true typical value" number are not the same thing. Read against Lenin, the contrast is the whole lesson: identical structure, opposite momentum, opposite recommended posture.
The same record, three different readers
A comp record does not hand every reader the same number, because a buyer, a seller, and an insurer are asking different questions of the same evidence. Watch how the Lenin Record answers each.
The buyer wants to know the most they should pay without overpaying. They anchor on the median ($525) as fair value, but they push toward the low end of the range ($250) and lean on the recent softness — a $409.80 recent average and a cooling trend — as negotiating leverage. A disciplined buyer of the Lenin Record today would open below the median, cite the recent sub-median prints, and refuse to chase the $700 January 2026 outlier. The HOLD tag works in their favor: the seller who understands the record is not desperate, but the recent data is on the buyer's side.
The seller wants the most they can realize without the print sitting unsold. If they need liquidity, they price near the suggested $356.25 — a number engineered to clear — and accept that the cooling trend caps their upside. If they can wait, the HOLD tag is permission to do exactly that: hold through the soft patch rather than realize a below-median price. The seller's key discipline is honesty about their own time horizon, because that, more than the print, determines which number they should use.
The insurer or estate wants a defensible fair-market value, neither the fire-sale floor nor the aspirational ceiling. They take the median ($525) as the honest typical value, note the range for context, and document the four-source methodology so the figure survives scrutiny. They deliberately ignore the $124.99 low and the $2,005 high — both are real, both are irrelevant to a fair-market estimate, which by definition describes the typical transaction between a willing, unhurried buyer and seller.
One record, three correct-but-different numbers, each derived by pulling the appropriate field for the appropriate question. That flexibility is a feature, not a contradiction: the framework does not pretend a single price exists, because a single price never does. It gives you a structured way to pick the right one for what you are actually trying to do.
Common ways this framework gets misread
Three mistakes recur often enough to name explicitly.
Chasing the max. The $2,005 Lenin and $2,006 Peace Goddess ceilings are seductive and useless as list prices. They are single events at the far tail. Anchor to the median; treat the max as the outer edge of the possible, reached only with an exceptional copy and an exceptional buyer on an exceptional day.
Trusting the mean on a skewed print. A collector who reads Lenin's $899.94 average and lists at $899 will watch it sit. The mean is 71.4% above the median because a few high sales inflate it. On any print where mean and median diverge sharply, listing at the mean is listing above the market.
Ignoring channel and window. Valuing off a single eBay sweep underprices; valuing off a lone auction record overprices; valuing off a pre-2016 result values a market that no longer exists. The floor, the window, the trim, and the four-source blend are not bureaucracy — each one patches a specific, repeatable error. Skip a rule and you reintroduce its failure mode.
Where this fits in the collection
This framework is the engine under every other valuation in the series. When we published our entry-level guide to the best Fairey prints under $300, every price in it came out of exactly this process — floored, windowed, trimmed, blended. When we quantify what a signature and edition number add in our study of signed and numbered premiums, the premium is measured by comparing comp records built this same way. The method is the connective tissue.
You can practice it yourself. Browse the Shepard Fairey collection, pick a print, and pull its record from the Fairey index. Run the eight steps. Read the mean against the median, plot the field, check the recent average for direction, and note the channel split. The more records you read, the faster the shape jumps out — and the harder it becomes for anyone to sell you a number without evidence behind it.
Because in the end, that is the entire point of a comp framework. Not to hand you a single magic price, but to make you fluent enough in the evidence that you can defend your own number to a seller, a buyer, an insurer, or yourself. A price is an opinion. A comp record, read this way, is an argument. Learn to read it and you never have to guess again.
This is analysis, not financial advice. Art is an illiquid, idiosyncratic asset; past sales do not guarantee future results, and every valuation carries uncertainty that widens as the sample thins. Comp figures are drawn from Gauntlet Gallery's aggregated database (WorthPoint, eBay, LiveAuctioneers, Heritage) and reflect tracked sales, not a live bid. Do your own diligence and consult a qualified professional before buying, selling, or insuring any work.
