Sherman Shatwell went to the hospital complaining of neck pain. Tests showed he probably had throat cancer. It was treatable but required immediate attention. Thanks to a variety of bureaucratic. blunders the news never made it to him. Instead, Mr. Shatwell was sent home with a prescription for antibiotics. By the time he learned the truth a year later, it was too late.
Eventually, his widowed wife pursued negligence claims against the doctors and hospital. Through twenty months of motions practice and discovery and all the way through their submissions for the final pretrial order the defendants maintained a unified front, denying any negligence by anyone. Then, two weeks before trial, some of the defendants settled. Dr. Kenneth Phillips wasn’t one of those. Left to stand trial and with just days before jury selection,' he sought permission to amend the pretrial order so he could revamp his trial strategy. Now he wanted to pursue a defense pinning the blame on the absent settling defendants, arguing that they were indeed negligent and that they- — not he — should be held responsible for any damages. Dr. Phillips’s motion to amend the final pretrial order sought permission to introduce new jury instructions, exhibits, and witnesses aimed at advancing this new defense. But the district court denied the motion and at the trial’s end the jury found him liable for damages of a little over $1 million. Dr. Phillips now asks us to overturn the judgment, contending that the district court’s refusal to amend the *851final pretrial order and allow his new defense amounts to reversible error.1
Final pretrial orders seek to “formulate a trial plan.” Fed.R.Civ.P. 16(e). In their complaints and answers lawyers and parties today often list every alternative and contradictory claim or defense known to the law; during discovery they sometimes depose every potential witness still breathing and collect every bit and byte of evidence technology, time, and money will allow. Final pretrial orders seek to tame such exuberant modern pretrial practices and focus the mind on the impending reality of trial. “The casual pleading [and discovery] indulged by the courts under the Federal Rules ... has quite naturally led to” — some might say required — “more and more emphasis on pre-trial hearings and statements to define the issues” for trial. Meadow Gold Prods. Co. v. Wright, 278 F.2d 867, 868-69 (D.C.Cir.1960). Leaving the reins so loose at the front end of the case requires some method of gathering them up as the end approaches. At trial you just can’t argue every contradictory and mutually exclusive claim or defense you were able to conjure in your pleadings: juries would lose faith in your credibility. Neither can you present the millions of documents and the scores of witnesses you were able to dig up in discovery: no sensible judge would tolerate it. Final pretrial orders encourage both sides to edit their scripts, peel away any pleading and discovery bluster, and disclose something approximating their real trial intentions to opposing counsel and the court. Toward those ends, the parties are often asked — as they were in this case — to specify the witnesses and exhibits, supply the proposed jury instructions, and identify the claims and defenses they actually intend to introduce at trial.
While pretrial orders entered earlier in the life of a case often deal with interstitial questions like discovery staging and motions practice and are relatively easy to amend as a result, a final pretrial order focused on formulating a plan for an impending trial may be amended “only to prevent manifest injustice.” Fed.R.Civ.P. 16(e). Even that standard isn’t meant to preclude any flexibility — trials are high human dramas; surprises always emerge; and no judge worth his salt can forget or fail to sympathize with the challenges the trial lawyer confronts. For all our extensive pretrial procedures, even the most meticulous trial plan today probably remains no more reliable a guide than the script in a high school play — provisional at best and with surprising deviations guaranteed. See, e.g., Sill Corp. v. United States, 343 F.2d 411, 420 (10th Cir.1965) (pretrial orders should not be treated as “hoops of steel”). At the same time, the standard for modifying a final pretrial order is as high as it is to ensure everyone involved has sufficient incentive to fulfill the order’s dual purposes of encouraging self-editing and providing reasonably fair disclosure to the court and opposing parties alike of their real trial intentions. See, e.g., Case v. Abrams, 352 F.2d 193, 195 (10th Cir.1965) (“A policy of too-easy modification not only encourages carelessness in the preparation and approval of the initial order, but unduly discounts it as the governing pattern of the trial.” (quoting A. Sherman Christenson, The Pre-Trial Order, 29 F.R.D. 362, 371 (1961))). This court will review a district court’s decision to amend or not to amend a pretrial order only for abuse of discretion. Koch v. Koch Indus., Inc., 203 F.3d 1202, 1222 (10th Cir.2000).
*852We see nothing like that here. Dr. Phillips says that he was surprised when his co-defendants left him to stand trial and that the court was insufficiently sympathetic to his desire to revamp' his trial strategy in light of ,the last-minute settlement. But can a partial settlement really come as a surprise in an age when virtually all cases settle in part or in whole, many on the eve of trial? Especially in multiparty litigation, where an incentive exists to break ranks, settle relatively cheaply, and leave others on the hook before the jury? The truth is, what happened in this case was hardly unforeseeable. Like many before him in multi-defendant cases, Dr. Phillips initially saw profit in presenting a united front with his co-defendants only to regret the decision later. United front defenses often present a tempting choice at the outset of multiparty cases and through discovery. Parties can pool their resources and efforts in joint defense arrangements. Besides, no one likes to throw overboard someone else in the same boat. But the complications associated with this strategic choice often come home to roost as trial nears. An attractive partial settlement may be dangled before one defendant and not others. The settling defendant may get a good deal, replenish an opponent’s litigation coffers depleted through exhausting pretrial litigation, and leave others exposed at trial for the bulk of the plaintiffs damages. Remaining defendants can be left wishing for a defense or evidence or witnesses forgone. If a remaining defendant’s attorney counted on a colleague working for a settling party to do the heavy lifting at trial he may feel flat-footed when it comes to examining witnesses and arguing motions. Even if all the defendants do go to trial, failing to obtain experts and gather evidence to show contributory negligence by co-defendants can exact its toll and lead to regret. Multiparty litigation presents a variety of collective action problems and other strategic pitfalls and those Dr. Phillips encountered here are well known, not the stuff of surprise. See, e.g., Manual for Complex Litigation (Fourth) § 13.21 (2004), 2004 WL 258728, at *1 (settling with one of many adverse parties on the eve of trial to weaken another party’s position is “a common and legitimate litigation strategy”); Jerome Wahlert, Contribution Between Joint Tortfeasors as Affected by Settlement with Injured Party by One or More Tortfeasors, 17 A.L.R.6th 1 (2006) (observing that “virtually all” multi-defendant tort cases “involve potential contribution issues that can arise 'when one or more tortfea-sors consider entering into settlement agreements”); Geoffrey C. Hazard, Jr., The Settlement Black Box, 75 B.U. L.Rev. 1257, 1269-70 (1995) (noting that “the interests of multiple defendants in a specific level of settlement normally diverge”).
It’s hard, as well, to ignore the prejudice the other side can experience in these circumstances. Dr. Phillips effectively sought to force the plaintiff to prepare for an entirely different trial on a few days’ notice. For the better part of two years the defendants presented a united front. Even in their final pretrial order submissions they didn’t designate experts to suggest one or another of the codefendants acted unprofessionally. They didn’t submit documents to prove such a claim. They didn’t propose jury instructions asking that someone else be held accountable. The closest they came to suggesting contributory negligence was to prepare boilerplate blaming unspecified others for Mr. Shatwell’s injuries. Beyond that, through the long months of discovery and into their final Rule 16(e) submissions, nothing. In these circumstances, the plaintiff and her lawyers had some reasonable expectations about what trial would look like and the sort of evidence they would — and would *853not — need. They knew they’d need to prove negligence by the defendants who chose to go .to trial but they wouldn’t have to worry about finger pointing between defendants; trial would present one set of challenges but not another.
It may be that the district court could have allowed Dr. Phillips to rejigger his defense at the last minute and afforded the plaintiff more time to prepare for it. But we do not see why that outcome was mandatory. A district court does not abuse its discretion in holding a party to a long-scheduled trial and to the strategy he articulated though pleading and discovery and in the face of such obvious risks, especially when indulging an eleventh-hour strategic shift would mean either imposing prejudice on the other side or inviting more delay. So beware: when a fellow litigant settles on the eve of trial you can’t bank on the right to claim surprise and rewrite your case from top to bottom. See, e.g., 6A Charles Alan Wright, Arthur R. Miller & Mary Kay Kane, Federal Practice and Procedure § 1527.1, at 408-12 (2010) (explaining that a'court may decline a motion to amend the pretrial order “if the evidence or issue was within the knowledge of the party seeking modification at the time of the pretrial conference”); R.L. Clark Drilling Contractors, Inc. v. Schramm, Inc., 835 F.2d 1306, 1308 (10th Cir.1987) (concluding that the district court did not abuse its discretion by holding the defendant to its prior “deliberate” choice not to pursue a particular legal theory).
Many of Dr. Phillips’s remaining arguments echo his Rule 16 complaint and fail with it. For example, he challenges the court’s evidentiary ruling at trial prohibiting him from asking one of the plaintiffs expert witnesses questions designed to elicit suggestions of negligence by the settling defendants. And he argues that the court erred when it refused to instruct the jury to apportion liability and damages between him and the settling defendants. But these evidentiary and instructional complaints are no more than collateral attacks on the district court’s Rule 16 ruling. When the court decided to hold Dr. Phillips to his self-declared trial plan presented only right before trial, that necessarily included the evidence he could submit and the issues he could put to the jury. See Trierweiler v. Croxton & Trench Holding Corp., 90 F.3d 1523, 1543 (10th Cir.1996) (“We interpret the assertion of an issue not listed in the pretrial order as the equivalent of a formal motion to amend the order....”). Dr. Phillips replies that, however that may be, the plaintiffs own questioning of her experts at trial “opened the door” to questions about the settling defendants’ liability. But the district court saw no creak in the door and, after our own review of the transcript, we see no abuse of discretion in that assessment either. See United States v. Burch, 153 F.3d 1140, 1144 (10th Cir.1998).
In places, Dr. Phillips seems to suggest that we should, on our own motion, reduce the jury verdict against him to account for the alleged misconduct of the settling parties. Even though the jury found him personally responsible for the particular damages it awarded, he suggests we should relieve him of some of that amount and (re)apportion it to those who settled before trial. But if that’s what he seeks, we can no more oblige this request than his others. Under Oklahoma law governing this diversity dispute, it’s not for courts to absolve defendants who elect to go to trial from damages a jury has lawfully found them personally liable for and attribute some of those sums to others not found liable: “a pretrial settlement cannot be used as a credit where the ghost tortfeasor’s liability was not submitted to *854the jury.” Pain v. Sims, 283 P.3d 343, 345 (Okla.Civ.App.2012) (discussing Okla. Stat. tit. 12, § 832(H), which permits set-offs only when a settling party is “liable” in tort for the same injury); Nichols v. Mid-Continent Pipe Line Co., 933 P.2d 272, 280 (Okla.1996). The proper way to handle the question of a settling defendant’s liability is through a claim of-contributory negligence put before the jury, something (again) Dr. Phillips sought only at the last moment when it was too late — revealing this argument to be no more than yet another, collateral attack on the court’s Rule 16 ruling.
What’s left of this appeal at this point doesn’t warrant extended discussion. Dr. Phillips complains about the district court’s exclusion of evidence about Mr. Shatwell’s use of tobacco and alcohol. But at trial the cause of Mr. Shatwell’s cancer wasn’t at issue, only the failure to warn him of its existence, so the evidence bore no relevance to the question of liability. Of course, if Mr. Shatwell’s tobacco and alcohol use affected his prognosis or life expectancy, it would have become relevant to the question of damages. But, the district court found, Dr. Phillips failed to present competent evidence suggesting a material link between the amount of tobacco and alcohol Mr. Shatwell used and outcomes for the sort of cancer he had. In these circumstances, the court held that most inquiries about Mr. Shatwell’s use of tobacco and alcohol would be more prejudicial than probative of any issue in dispute. See Fed.R.Evid. 403. Appellate courts review evidentiary complaints only for abuse of discretion, and the degree of a district court’s discretion in evidentiary rulings under Rule 403 is “particularly” wide. Sprint/United Mgmt. Co. v. Mendelsohn, 552 U.S. 379, 384, 128 S.Ct 1140, 170 L.Ed.2d 1 (2008). The district court here didn’t overstep those bounds for the very reason it gave: Dr. Phillips failed to present competent evidence indicating a connection between Mr. Shatwell’s conduct and his prognosis. That’s not to say a link doesn’t exist or couldn’t be suggested with enough plausibility to permit or even require its admission in another case. Only that none of this happened in ours.
Affirmed.
. Technically, the judgment lies against Dr. Phillips and his employer, Tahlequah Emergency Consultants, PLLC, but for simplicity’s sake we refer just to Dr. Phillips.