Suppose there is a yes/no question on a prediction market which currently has "yes" at 0.20 and "no" at 0.80. Some data is about to be published that unambiguously makes "yes" more likely, but it's less obvious how much more likely and a more accurate prediction requires more resources or skill.
Further suppose you were going to spend resources analyzing the data and after that your conclusion would be that "yes" is a buy at up to 0.75, i.e. you think there is a 75% chance of it happening. And it turns out you were on the right side because it does end up being yes.
If everyone gets the data at the same time, you can start buying "yes" when it's still at 0.20 and keep buying it until it gets up to 0.75. If someone gets the data before you, even if they're (wrongly) less confident than you and only buy at up to 0.55, they start buying a day earlier and by the time the data is published, "yes" is already at 0.55. The result is that you, who were a buyer at up to 0.75, get significantly fewer contracts (because the other sellers already sold to the insider), and worse yet you only get the ones that cost more than 0.55 instead of getting many of the ones that were 0.20. In other words, you make significantly less money when it ultimately turns out you were right.
If your average profit is now less than the amount you needed to justify doing the analysis, you stop doing it, and then you don't buy any contracts at all. Then the price ends up stuck at 0.65 "yes" instead of 0.75 "yes" when "yes" was the right answer, because there wasn't enough profit left to fund an analysis that could justify higher confidence in the correct result.
Further suppose you were going to spend resources analyzing the data and after that your conclusion would be that "yes" is a buy at up to 0.75, i.e. you think there is a 75% chance of it happening. And it turns out you were on the right side because it does end up being yes.
If everyone gets the data at the same time, you can start buying "yes" when it's still at 0.20 and keep buying it until it gets up to 0.75. If someone gets the data before you, even if they're (wrongly) less confident than you and only buy at up to 0.55, they start buying a day earlier and by the time the data is published, "yes" is already at 0.55. The result is that you, who were a buyer at up to 0.75, get significantly fewer contracts (because the other sellers already sold to the insider), and worse yet you only get the ones that cost more than 0.55 instead of getting many of the ones that were 0.20. In other words, you make significantly less money when it ultimately turns out you were right.
If your average profit is now less than the amount you needed to justify doing the analysis, you stop doing it, and then you don't buy any contracts at all. Then the price ends up stuck at 0.65 "yes" instead of 0.75 "yes" when "yes" was the right answer, because there wasn't enough profit left to fund an analysis that could justify higher confidence in the correct result.