The wins leaderboard is the visible side of whale activity on prediction markets. The losses leaderboard is the more instructive one. Whales lose, frequently, and the structure of the losses is often more revealing of how the markets actually work than the structure of the wins. This article walks through the largest paper-PnL losses on record across Polymarket and Kalshi, the structural patterns that recur across the worst outcomes, and what copy-traders can learn from studying the list rather than ignoring it. The full ranked list is available on the all-time losses page; the top entries are shown below.
- 1Roland Garros ATP: Felix Auger-Aliassime vs Flavio CobolliPolymarket·sports·stake $510.0K·@ 51¢-$510.0K
- 2Will IR Iran win on 2026-06-15?Polymarket·politics·stake $447.5K·@ 54¢-$447.5K
- 3Spread: Arsenal FC (-2.5)Polymarket·sports·stake $412.9K·@ 51¢-$412.9K
- 4Will Germany win on 2026-06-25?Polymarket·culture·stake $400.2K·@ 66¢-$400.2K
- 5Will Manchester City FC win on 2026-05-19?Polymarket·sports·stake $350.8K·@ 61¢-$350.8K
- 6Will Juventus FC win on 2026-05-24?Polymarket·culture·stake $330.1K·@ 68¢-$330.1K
- 7Will Türkiye win on 2026-06-14?Polymarket·culture·stake $303.1K·@ 57¢-$303.1K
- 8US x Iran permanent peace deal by June 15, 2026?Polymarket·politics·stake $302.5K·@ 86¢-$302.5K
- 9Spread: Morocco (-1.5)Polymarket·culture·stake $300.0K·@ 51¢-$300.0K
- 10Will Belgium win on 2026-06-15?Polymarket·culture·stake $296.7K·@ 62¢-$296.7K
The three patterns that produce the worst losses
The trades that appear at the top of the losses leaderboard share structural features almost as consistently as the trades that appear at the top of the wins leaderboard. Three patterns account for the majority of the entries.
The conviction trade that aged badly
The first pattern is a long-dated position taken at what looked like a defensible price, held through months of adverse news cycles, and resolved against the bettor at the final settlement. A whale opens a political contract at 0.55 in March because the analytical case looks strong. Over the following six months, the underlying narrative shifts, the price drifts to 0.30, and the eventual resolution is NO. The paper PnL on the position is the full size of the original stake.
The pattern is instructive because it shows the actual cost of long-dated conviction trading. Conviction is not free; the trader pays carry in the form of an extended period during which the position is exposed to information that the trader did not have at entry. The longer the holding period, the more opportunities exist for the market to absorb new information that contradicts the original thesis. For copy-traders, the lesson is that long-dated positions require both a strong analytical view and the willingness to size the position so the full-stake loss is survivable.
The late longshot with no catalyst
The second pattern is a large position taken on a low-probability contract with little time remaining before resolution. A whale buys a contract at 0.05 with three hours left on a market that has not moved for several days. There is no upcoming event on the schedule within the resolution window. The contract resolves NO and the position pays $0.
The late longshot pattern is the most identifiable and the most preventable of the three. The market is generally close to the correct probability when the underlying event is hours away, and a position taken at an extreme price at that point requires a specific informational advantage that retail copy-traders almost never have. The pattern is one of the clearest fade setups in the framework discussed in our copy or fade article.
The news chase
The third pattern is a position opened immediately after a sharp price move triggered by news. A headline prints and the contract gaps from 0.30 to 0.55 within seconds. A whale enters at the post-news price. The market subsequently mean-reverts portions of the initial reaction once the news is fully absorbed, and the whale is left holding a position at a price the market no longer supports.
The news-chase pattern is harder to fade in real time than the late longshot, because the news itself does contain information; the question is whether the price reaction is proportional to the long-term implication of the news. The Rivo dataset suggests that approximately half of major post-news price moves on prediction markets reverse materially over the following few hours, which makes the news chase a meaningfully negative expected value behavior even when the underlying news is genuine.
What the losses leaderboard does not show
The losses leaderboard, like the wins leaderboard, is a tail of the distribution. The bulk of whale activity produces moderate outcomes that do not appear on either list. A more complete view of whale performance, weighted by aggregate PnL across all trades rather than by single-trade extremes, is available in our profitability analysis. Copy-traders who read only the leaderboards form a view of whale outcomes that is too volatile in both directions; the population-level dispersion is less dramatic than the tail-event lists suggest.
The losses list also does not show the trades that a whale closed early to limit a loss. A position opened at $0.55 that was closed at $0.40 produces a smaller realized loss than the full-stake paper PnL shown on the leaderboard. The methodology used to rank trades assumes the position was held to resolution, which produces the largest possible paper-PnL figure for the entry. The decision to publish on this basis is described in more detail in our biggest wins article.
Why publishing the losses matters
The structural reason to publish a losses leaderboard alongside a wins leaderboard is to provide the corrective view to survivorship bias. Wins leaderboards exist on most platforms in some form; losses leaderboards are rarer because they are less flattering to the underlying activity. The asymmetry in publishing produces an overoptimistic public view of how prediction-market betting actually works.
For copy-traders, the losses leaderboard is the cheaper teacher. Reading the wins list produces confidence; reading the losses list produces calibration. The wallets that appear on both lists are the most informative individual cases, because the same trader was correct in one situation and wrong in another, and the comparison surfaces the specific conditions under which the trader's edge works and the conditions under which it does not. That kind of pattern recognition is the intermediate-level skill in copy-trading; it develops by studying losses as carefully as wins.
How to use the list in practice
The defensible weekly practice is to scan the losses leaderboard for new entries before scanning the wins leaderboard. The new losses surface the recent failure modes in the market: the political narratives that flipped against the consensus, the sports outcomes that defied the modeling, the crypto contracts that produced unexpected invalidations. Each of these is information that the next copy-trade is implicitly betting against.
Reading the losses list does not produce direct copy-trades, because the trades shown have already resolved. What it produces is a sharper sense of the current failure modes in the market, which informs the evaluation of fresh entries in the feed. A copy-trader who has just observed three political news chases reverse on the losses list will appropriately discount the next political news chase that appears as a fresh whale entry. The discipline is to let the recent losses adjust the bar for incoming trades, rather than to ignore them entirely.
Frequently asked questions
Are the losses on the leaderboard realized or paper?
Paper. The methodology assumes each position was held from entry to resolution at the entry price, gross of fees and execution costs. Realized losses for individual whales may be smaller if the position was closed early at a less unfavorable price.
Do whales who appear on the losses list also appear on the wins list?
Frequently. The same wallets appear on both lists with substantial overlap. The pattern is consistent with directional bettors operating with edge that is positive in expectation but variable in outcome. Studying the cases where the same wallet appears on both lists is one of the more informative ways to develop pattern recognition.
What is the most preventable loss pattern?
The late longshot with no catalyst is the most preventable. The pattern has identifiable structural features and can be filtered out by a copy-trader following the framework in our copy or fade decision guide. Avoiding this pattern alone removes a meaningful share of the largest losses on the leaderboard.
Why is the news chase hard to avoid?
The news is informationally legitimate; the question is whether the price reaction is proportional. The reaction often overshoots the long-term implication of the news, but the overshoot is not always identifiable in the moment. The defensible policy is to delay execution by several minutes after major news events, which allows the initial reaction to absorb before any copy-trade is taken.
Does Rivo flag entries that fit the loss patterns above?
The platform surfaces several of the relevant signals, including time-to-resolution, market depth, and entry-price extremity. The combination of signals supports the copy-trader in evaluating whether a fresh entry resembles one of the historical loss patterns, but the final judgment is the copy-trader's. Mechanical filtering on a single attribute will produce both false positives and false negatives.