It's well known that patent trolls forum-shop to get their cases in the Marshall courthouse of the Eastern District of Texas, to the point of setting up empty offices in the neighborhood to create venue. But, surprisingly, I've never seen a complete explanation written why. Let me do so.
1) Only two judges sit in the Marshall division. Correctly or incorrectly, they are perceived as judges who rarely grant summary judgment, and let juries decide issues. Moreover, judge-shopping is possible: one of the judges has a close relative in private practice, creating strategic disqualification opportunities. (Judge David Folsom's announced retirement may or may not change this dynamic, depending on who is appointed to replace him.)
2) The Eastern District of Texas is perceived to have plaintiff-friendly juries. (Only 20% of the population holds a bachelor's degree, meaning that technically complicated cases may well get decided on grounds other than the science.) The actual quantitative trial win-loss record of patent plaintiffs in the E.D. Tex. isn't so spectacular in recent years, which leads some to call the idea of plaintiff-friendly juries a myth. But that quantitative record needs to be discounted because, qualitatively, much poorer cases get to the jury in the Eastern District than elsewhere. (That said, adjustment of defense trial tactics in recent years seems to have had some effect.) And once plaintiffs do win, the juries tend to be on the high side on the damages calculations.
3) The Eastern District of Texas, like many other federal districts, has a "rocket docket": local civil rules of procedure for patent cases expedite the consideration of such cases. There's nothing inherently wrong with this, but the effect is a benefit for patent plaintiffs. A patent plaintiff who has its act together will have its ducks in a row pre-filing: all of its research, experts, and infringement theory of the case prepared. In most litigation, it's the plaintiffs who need discovery the most: lots of document discovery to find smoking guns, lots of depositions to induce witnesses into saying things that can be perceived as damaging admissions. Plaintiffs need discovery in patent cases, too, especially if they're hoping to prove "willfulness," but defendants need discovery more. Defendants will be working from scratch, and hoping to find evidence to invalidate a patent. An expedited discovery schedule makes that more difficult. An expedited discovery schedule also requires a larger team of litigators, because limited bandwidth makes lean staffing inadvisable. This has the potential to make a case more expensive than it would be with a longer schedule, both in terms of cash flow and because there's more wasted and duplicated effort conveying information when one is coordinating a larger team of attorneys, though there is the countervailing effect of the impossibility of no-stone-left-unturned/scorched-earth litigation that can take place with a conventional litigation schedule.
4) The remoteness of the Marshall location raises litigation costs considerably. Every 15-hour round-trip by a senior litigator having to change planes in Atlanta adds $10,000 or more to the bill. This, combined with the decreased likelihood of resolving the litigation through motion practice rather than expensive trials, and the increased costs of the rocket docket, means that there's an in terrorem effect from simply bringing the lawsuit: the patent troll can offer to settle for far less than the cost of litigation, and the defendant, who does not collect the full social value from invalidating a bad patent, often does not have the incentive to keep fighting.