MEMORANDUM, ORDER and JUDGMENT
Table of Contents
I. Introduction..269
II. The Problems Created by the Theory of Review...270
III. Facts...;.271
IV. Standard of Review .273
A. Law . 273 I
B. Analysis of Law . 03
Y. Merits of the Claim .. t — 1 OO 03
A. Issue of Reversal rH CO 03
B. Issue of Remand rH CO 03
VI. Conclusion. to 00 to
I. Introduction
Plaintiff Rita J. Mood sues defendants Yasuda Fire and Marine Insurance Company of America Long Term Disability Plan (the “Plan”) and Prudential Insurance Company of America (“Prudential”), the Plan administrator. She contends that she was improperly denied long-term disability benefits under the Plan, an employee welfare benefit arrangement governed by the Employee Retirement Income Security Act of 1974 (“ERISA”), as amended, 29 U.S.C. § 1001-1461. The district court reviews the denial to determine whether Prudential’s decision should be approved, or remanded for further proceedings, with *270a direction to provide benefits, or for other action. The leading opinion of the Court of Appeals for the Second Circuit leaves room for clarification. See Nichols v. Prudential Ins. Co. of Am., 406 F.3d 98 (2d Cir.2005). For the reasons stated below, a remand is ordered.
II. The Problems Created by the Theory of Review
Among the questions now posed are the following: Did Prudential reasonably construe the medical and other evidence before it? Does the record cover medical determinations as to which reasonable doctors can differ, such as the claims of “multiple chemical sensitivity” and of carpal tunnel syndrome and the intensity of the claimed disabling symptoms? Even if there is a “multiple chemical sensitivity” defect causing some difficulty to claimant, is it reasonable for a person in her condition under current socioeconomic conditions not to seek employment? Would a reasonable potential employer hire this claimant even if she were eager to work, though she was perhaps somewhat disabled? Triers such as plan administrators and courts bring their own biases and life experiences into an assessment of what is reasonable in answering each of these questions.
The problem is quite different from that of determining whether a specific event took place in the real world — e.g., did defendant fire a gun into the deceased’s heart, or did an employee fall off a ladder while obtaining a box of Wheaties for a customer? The terms “disability,” “permanent disability,” “employability,” “nature and degree of disease,” and the like, all are somewhat subjective and somewhat vague, with varying factual and legal content. They are analogous to concepts such as “negligence” in tort actions, “disability” in social security actions, and “reasonableness” in applying Supreme Court rulings in habeas corpus proceedings.
Solution is further complicated because the federal courts (appellate as well as trial) are expected to decide not how they would have decided the issue were they the administrator, but whether the administrator’s decision should be validated by the courts. In this respect, the problem is somewhat similar to that described in United States v. Copeland, 369 F.Supp.2d 275, 276 (E.D.N.Y.2005), where the trial court was directed to decide how a hypothetical administrative immigration body would have decided a fact-law issue had it ruled on the issue a decade ago.
A judge may say, “I believe a mistake was made by the administrator,” but in the back of his or her mind will be the question, “What would other federal judges conclude?” Were there an attempt to think of the problem in quantitative terms it might proceed somewhat like this: How would a group of randomly chosen federal appellate judges (or trial judges) think a group of randomly chosen administrators would evaluate the diagnosis of claimant made by a group of randomly chosen specialist doctors (or general practitioners, as the case may be); how would a group of randomly selected employees with claimant’s alleged medical problems react to their difficulties in looking for a job; and how would a group of randomly selected potential employers react to a job-seeker such as the claimant with her alleged medical difficulties? And, in this chain of probabilities, what numbers should be required in each link?
As indicated below, infra Part IV, the courts in the Second Circuit, as elsewhere, approach the problem by first deciding whether federal trial and appellate courts should apply a “de novo ” or an “arbitrary and capricious” standard in reviewing the administrator’s decision. Conceptually, *271the phrases “de novo ” and “abuse of discretion” are not well-defined cubbyholes, but areas on a spectrum of doubt about the reasonableness of the administrator’s decision. “Arbitrary and capricious” is not completely arbitrary and capricious, like tossing a coin might be, but unreasonably arbitrary and capricious under the circumstances. And “de novo,” while reflecting the usual Latin obscurantism of the law, does not really mean the judge will decide as on a clean slate without considering the decision of the administrator (or on appeal the decision also of the trial court). Even the most egocentric federal judge applying de novo revieiv will give some weight to what the presumably more expert or experienced plan administrator actually did, even if that weight is applied sub rosa or subconsciously. Both de novo and arbitrary and capricious standards pack a good deal of reviewer subjectivity.
Another way of describing the court’s review process is to paraphrase Holmes: It is a prediction of what a large group of randomly selected federal judges would by consensus conclude on the same record as the administrator had before it. See Oliver Wendell Holmes, Jr., The Path of the Law, 10 HaRV. L.Rev. 457 (1897). If most would disagree with the administrator, there was an abuse of discretion; if a considerable number would disagree, de novo review requires remand. Cf, e.g., Williams v. Taylor, 529 U.S. 362, 409, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000); Henry v. Poole, 409 F.3d 48, 68, 2005 WL 1220468, at *18 (2d Cir.2005) (“Williams also made clear that a federal habeas court may permissibly conclude that federal law has been unreasonably applied by the state court even though not all reasonable jurists would agree that the state courts’ application was unreasonable.”).
Unlike an affirmance or a reversal, a remand involves different considerations. The reviewing court may not have a clear idea of how the administrator dealt with the problem procedurally — e.g., were all aspects of the claim and relevant available data considered; were the individual diseases evaluated; and was the whole effect of all diseases, together with such factors as age, education, obesity, and the job market considered? Here the reviewing judge may need to have some assurance that the claim has been “fairly” (itself a flexible term) considered by the administrator in a procedural sense.
It does not help in the judicial evaluation of the fairness of the medical and job-related decisions of the administrator that judges have no expertise in either field. Their scientific knowledge and understanding of the record may well be based on faulty assumptions. A useful approach to deciding how federal judges should view the record is suggested in Anderson v. City of Bessemer City, 470 U.S. 564, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985) (involving a review of findings of the district court in a sex discrimination case). The Supreme Court held that a ruling on the record is “clearly erroneous” when “ ‘although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.’ ” Id. at 573, 105 S.Ct. 1504 (citation omitted). While soft around the edges, this standard suggests the desired modesty of a reviewing court. Humility before reversing is called for under any standard of review.
In short, much of what is done by federal courts in reviewing a plan administrator’s decision, as in the instant case, or a hypothetical administrative decision, as in Copeland, is a stab in the dark. Nevertheless, the case must be decided.
IIL Facts
Mood brings this action under section 1132(a)(1)(B) of title 29 of the United *272States Code, alleging that she was improperly denied her claim for long-term disability benefits. She was employed by Yasuda as an Insurance Business Analyst from February 7, 2000 to October 13, 2000, and was a “participant” in its long-term disability insurance plan administered by Prudential.
Based on a claimed disability, she stopped work on October 13, 2000. On October 27, 2000, she applied for disability benefits because of a diagnosis of “sensory disturbance,” with the symptoms of “transient spells and involuntary muscle activity triggered by office environment.” Administrative Record (“AR”) at PRU 565. As noted below, the claim was later supplemented with a diagnosis of “multiple chemical sensitivity” and allegedly related symptoms such as carpal tunnel syndrome. She was given short-term disability benefits for the maximum period of 26 weeks. Pl.’s Rule 56.1 Statement at 2.
The policy reads in part:
How Does Prudential Define Disability?
You are disabled when Prudential determines that:
• you are unable to perform the material and substantial duties of your regular occupation due to your sickness or injury; and
• you have a 20% or more loss in your indexed monthly earnings due to that sickness or injury.
After 36 months of payments, you are disabled when Prudential determines that due to the same sickness or injury, you are unable to perform the duties of any gainful occupation for which you are reasonably fitted by education, training or experience.
AR at PRU 154 (italics added; bold in original). The policy defines the bolded terms. See id. The definitions for “sickness” and “injury” each specify that “[disability must begin while you are covered under the plan.” Id. (emphasis added). “Regular occupation” is defined as “the occupation you are routinely performing when your disability begins. Prudential will look at your occupation as it is normally performed instead of how the work tasks are performed for a specific employer or at a specific location.” Id. (emphasis added). “Material and substantial duties” means duties that:
• are normally required for the performance of your regular occupation; and
• cannot be reasonably omitted or modified, except that if you are required to work on average in excess of 40 hours per week, Prudential will consider you able to perform that requirement if you are working or have the capacity to work 40 hours per week.
Id.
The policy specifies that it does not cover a disability which “begins within 12 months of the date your coverage under the plan becomes effective; and is due to a pre-existing condition.” Id. at PRU 160. The 12-month period will be reduced if certain conditions are met. See id. at PRU 160-61. A 90-day exclusion period for the purposes of the pre-existing condition determination was in Mood’s case from November 9, 1999 to February 6, 2000. Defs.’ Statement in Resp. to Pl.’s Rule 56.1 Statement at 2.
On February 21, 2001, Prudential denied Mood’s claim for long-term disability benefits because (1) the results of the diagnostic tests were within normal ranges and thus she did not meet the definition of disability, and (2) her symptoms were excluded from coverage as a pre-existing condition. AR at PRU 238-40.
By a letter dated May 22, 2001, Mood informed Prudential that she was appeal*273ing the denial and would provide further evidence from her treating physicians. Id. at PRU 559. On February 16, 2002, Mood submitted additional materials, such as various notes from her treating physician, Dr. Jung Youn. Id. at PRU 505-06. Dr. Youn met the pre-existing claim objection of Prudential by explaining that his treatment note for “January 5, 2000” should have been dated “January 5, 2001,” id. at PRU 532, and that “[t]he chart notes on patient Rita Mood do not document multiple chemical sensitivity between 11/9/99 and 2/6/00,” id. at PRU 531. Also submitted were materials from a Dr. Roger G. Mazlen, who diagnosed Mood as having “multiple chemical sensitivity.” Id. at PRU 507-08.
On June 18, 2002, Prudential denied Mood’s first request for reconsideration based in part on a report from its own consulting physician, Dr. Howard Kipen, and his review of the medical documentation in her file. Id. at PRU 219-21. Dr. Kipen concluded that the medical records failed to provide evidence of any impairment that would prevent her from working. Id. at PRU 220; see also id. at PRU 441-44. Prudential also noted that Mood’s claim was barred by the pre-existing condition exclusion since she had been seen by Dr. Youn on January 5, 2000 for the same symptoms which caused her to stop working in October of 2000. Id. at PRU 220.
On April 22, 2003, Mood again appealed. Id. at PRU 382-83. She submitted additional documentation, including Dr. Ma-zlen’s rebuttal to Dr. Kipen, id. at PRU 384-85, a neurological evaluation from an examining physician, Dr. Sana Bloch, id. at PRU 404-08, and notes from another neurologist, Dr. Slavina Gardella, id. at PRU 409-10.
On June 25, 2003, Prudential denied Mood’s second request for reconsideration. Id. at PRU 204-06. The denial was based on the medical file review performed by another of Prudential’s consulting physicians, Dr. Ahmed Khalifa; he concluded that Mood’s condition did not preclude her from performing the duties of her occupation as a business analyst. Id. at PRU 205.
Mood appealed a third time with submissions on January 8, 2004 and March 18, 2004. Pl.’s Rule 56.1 Statement at 9-10. Presented were additional materials from her doctors, Dr. Mazlen and Dr. Gardella. A functional assessment (of mobility, strength, sensation, etc.) by Ellen Rader Smith was also submitted, together with Dr. Mazlen’s critique of Dr. Khalifa’s report. Id.
On May 11, 2004, Prudential’s three-person Appeals Committee denied Mood’s third and final request for reconsideration. It based its decision largely on the review of the file by Prudential’s medical director. AR at PRU 197-99. The medical director concurred with the assessments of Dr. Ki-pen and Dr. Khalifa, that the records in the file supported the conclusion that Mood was able to perform her regular occupation. Id.
Mood commenced this action on April 9, 2004. She now moves for judgment on the administrative record, arguing that the applicable standard of review for Prudential’s denial of benefits is de novo. Defendants cross-move for summary judgment, contending that the proper standard of review is “arbitrary and capricious” and that denial was “reasonable, supported by substantial evidence, and not arbitrary or capricious.”
IV. Standard of Review
A. Law
A de novo standard is used to review a denial of benefits under ERISA, “unless the benefit plan gives the administrator or fiduciary discretionary authority *274to determine eligibility for benefits or to construe the terms of the plan.” Firestone Tire & Rubber Co. v. Bruch, 489 U.S. 101, 115, 109 S.Ct. 948, 103 L.Ed.2d 80 (1989); see also Nichols v. Prudential Ins. Co. of Am., 406 F.3d 98, 108-09 (2d Cir.2005) (citing Firestone, 489 U.S. at 115, 109 S.Ct. 948).
“Where the plan reserves such discretionary authority, denials are subject to the more deferential arbitrary and capricious standard, and may be overturned only if the decision is ‘without reason, unsupported by substantial evidence or erroneous as a matter of law.’ ” Kinstler v. First Reliance Standard Life Ins. Co., 181 F.3d 243, 249 (2d Cir.1999) (quoting Pagan v. NYNEX Pension Plan, 52 F.3d 438, 442 (2d Cir.1995)). “Substantial evidence” is that which “a reasonable mind might accept as adequate to support the conclusion reached by the [administrator and] ... requires more than a scintilla but less than a preponderance.” Celardo v. GNY Auto. Dealers Health & Welfare Trust, 318 F.3d 142, 146 (2d Cir.2003) (citation and internal quotation marks omitted). “[W]here both the trustees of [an ERISA plan] and a rejected applicant offer rational, though conflicting, interpretations of plan provisions, the trustees’ interpretation must be allowed to control.” Id. (citation and international quotation marks omitted).
Scope of judicial review is narrower under the “arbitrary and capricious” standard. Id. The burden of proving that the arbitrary and capricious standard applies is placed on the plan administrator. Kinstler, 181 F.3d at 249. “Magic words such as discretion and deference may not be absolutely necessary to avoid a [de novo ] standard of review.” Id. at 251 (citation and internal quotation marks omitted).
Even if a plan administrator is given discretion, the Court of Appeals for the Second Circuit has applied the de novo standard of judicial review if the administrator was “in fact” influenced by a conflict of interest. Fay v. Oxford Health Plan, 287 F.3d 96, 108-09 (2d Cir.2002) (citation and internal quotations marks omitted). A conflict “may be ‘inherent’ to some extent when a plan is both administered and insured by a single entity,” but the de novo standard is only applicable if there is an actual conflict of interest. Id. at 109.
B. Analysis of Law
The parties disagree about whether the Plan gives Prudential discretionary authority to determine eligibility for benefits. Mood argues that Nichols v. Prudential Ins. Co. of Am., 406 F.3d 98, 108-09 (2d Cir.2005), “fully supports” its argument that the phrase “when Prudential determines” does not confer discretionary authority and thus does not result in the “arbitrary and capricious” standard of judicial review. Prudential relies upon Judge Cote’s opinion in Simone v. Prudential Ins. Co. of Am., No. 04 Civ.2076, 2005 WL 475406, at *2 (S.D.N.Y. Feb.28, 2005), which addressed the same language at issue in the instant case. In Simone, the court concluded that the language (i.e., that a claimant is disabled only “when Prudential determines” that the claimant meets the specified criteria) conferred discretionary authority on Prudential to make a decision regarding eligibility for disability payments. Id. at *6; see also Wenzel v. Prudential Ins. Co. of Am., No. 03-CV-5751, slip op. at 5-7 (E.D.NY. Mar.25, 2005) (reaching the same conclusion as Simone, albeit with respect to what appears to be somewhat different policy language).
The following chart compares the relevant language of the long-term disability policies at issue in the leading cases in this circuit, Nichols, Wenzel, Simone, and O’Sullivan v. Prudential Ins. Co. of Am., No. 00 Civ. 7915, 2001 WL 727033 (S.D.N.Y. June 28, 2001), with that in the instant case.
*280In Nichols, Prudential had paid the claimant short-term and long-term disability benefits in accordance with the plan. 406 F.3d at 101-02. The plan limited long-term disability benefits to 24 months for disabilities based in part on mental disorders. Id. After two years’ of those payments, Prudential notified claimant that her long-term disability benefits would be discontinued. Id. Claimant appealed, but apparently did not submit any medical evidence and refused to submit to an independent medical examination. Id.
Nichols does not quote in full the language of the plan at issue. Its description of the plan is as follows:
The plan states that a disability “exists when Prudential determines that all of these conditions are met” and then goes on to list specific conditions. The plan also states that Prudential will pay benefits “when Prudential receives written proof of the loss.” The latter language is clearly objective. The phrase “when Prudential determines” is more troubling, but ultimately lacks sufficient indicia of subjectivity as required by Kinstler. See O’Sullivan v. Prudential Ins. Co. of Am., No. 00 Civ. 7915, 2001 WL 727033, at *2 (S.D.N.Y. June 28, 2001) (finding that identical language vested no discretion in plan administrator). The language gives Prudential the power to make the determination, but the list of specific conditions requires that such power be exercised only in accordance with objective standards. To find discretion, we would have to read in language, effectively amending the provision to find disability “when Prudential determines to its satisfaction that all these conditions are met.” We therefore approve of O’Sullivan and hold that the plan vests no discretion in Prudential.
Id. at 108-09 (emphasis added). Since the Court of Appeals for the Second Circuit held that the language of the plan did not vest discretion in the plan administrator, the de novo standard of review was applicable. Id. In the alternative, the court held that even if the plan vested discretion in the plan administrator, the administrator’s inaction was not a decision to which deference could be given. Id. at 108-10.
It is unclear whether Nichols controls here. It can be distinguished. First, although the plan language in the instant case includes the phrase “when Prudential determines,” it is used in a different context from that used in Nichols. Second, the enumerated conditions in the current case differ from those in Nichols. Two of the three conditions in Nichols involved “objective” language — i.e., the claimant (1) is not working for wage or profit and (2) is under the regular care of a doctor. Greater subjectivity is involved for the remaining condition in Nichols, which is similar to the conditions in the instant case, with respect to the terms “material and substantial duties” and “occupation” or “regular occupation.” Third, the language addressing payment differs: In the instant case, a claim has to be approved by Prudential and payment is made “for any period for which Prudential is liable,” whereas in Nichols payment was to be made “when Prudential receives written proof of the loss.”
As discussed below, the court need not now decide which standard of review is appropriate. The arbitrary and capricious standard appears, however, to be appropriate given the language in the plan.
Mood seems to argue that there was a conflict of interest — i.e., Prudential did not act as a disinterested fiduciary because it did not credit Mood’s physicians. See Pl.’s Mem. of Law, Mar. 16, 2005, at 24-26. Prudential is not required *281to defer to Mood’s physicians. In any event, in the instant case there is no evidence of a conflict of interest that would itself authorize application of the de novo standard.
V. Merits of the Claim
A. Issue of Reversal
Here, as is often true in such cases, the conclusions of claimant’s doctors differs from those of Prudential’s doctors. Even assuming arguendo that the de novo standard of review were applicable, the evidence in the administrative record suggests that judgment could not be granted in Mood’s favor unless the court were willing to credit her less-than-compelling submissions and reject those of Prudential. Prudential had two independent bases to deny benefits to Mood: (1) her medical record did not support the conclusion that she could not perform her regular occupation, and (2) she was ineligible for benefits due to the pre-existing condition exclusion.
Prudential’s evaluation was based on the reports of various doctors, including Mood’s additional submissions from her doctors and “independent” doctors consulted by Prudential, Dr. Kipen and Dr. Khali-fa. Prudential’s experts, including its medical director, agreed in their conclusion that Mood could perform her regular occupation. Prudential’s experts suggested that Mood’s doctors largely provided con-clusory statements and diagnoses that were unsupported by evidence. E.g., AR at PRU 44S (Dr. Kipen’s Report) (“Dr. Mazlen’s findings of pericarditis and pleurisy, which are both inadequately documented as there is no accompanying relevant history or chest film to complete these evaluations, do not have any clear implications for the ability to work. The findings are undermined by absence of supportive data in numerous examinations.”); id. at PRU 533 (Dr. Youn’s note of May 1, 2001) (“Rita is disabled and has been unable to work since October 13, 2000.”).
Although Mood’s doctors expressed disagreement with Prudential’s experts, unlike social security disability cases, an administrator need not accord “special deference” to treating physicians. See Black & Decker Disability Plan v. Nord, 538 U.S. 822, 831, 123 S.Ct. 1965, 155 L.Ed.2d 1034 (2003) (“Nothing in [ERISA] itself ... suggests that plan administrators must accord special deference to the opinions of treating physicians.”).
As for the pre-existing condition exclusion, application of the provision is unclear. It was conceded at argument that Dr. Youn’s date “correction” should be credited so that his treatment note is not for January 5, 2000, but for January 5, 2001. Yet it appears that Mood sought treatment from Dr. Fay Yang for symptoms similar to those that she now presents, such as lightheadedness, dizziness, and work-related anxiety, on September 1, 1999; November 8, 1999; November 15, 1999; and January 7, 2000—dates well before the claimed onset of her diseases. AR at PRU 239, 471-73. Some of these days fall within the period that could bar Mood from receiving the benefits sought.
B. Issue of Remand
The absence of Appeals Committee documentation suggests that it accepted, without its own analysis, the conclusions of its medical staff. This situation is quite unlike social security cases in which the courts are accustomed to reviewing detailed reports and an evaluation by an administrative judge, not a physician, of the evidence from both sides. Whether an acceptance by the Committee of its medical staffs opinion without its own analysis of the underlying evidence should be ac*282ceptable in these ERISA matters need not be decided at this time.
The record now presented is unclear as to whether the earlier symptoms necessarily led to the conclusion that the claimed disability was pre-existing. Also unanswered is whether the question of disability was considered in the context of the claimant as a whole person, including her obesity, age, education and available jobs. Dr. Khalifa did note factors such as Mood’s age and obesity and rejected obesity as impairing her from working, see AR at PRU 351-52, but it is not clear what the Appeals Committee did in evaluating claimant’s condition in light of the job market. Until the record is clarified, an appeal cannot be fairly decided.
In this case, the court would reach the same conclusion requiring a remand regardless of the standard of review applied. Serious issues of fact were not resolved by Prudential. It should be given an opportunity to make relevant determinations in the first instance.
Mood’s claim should be remanded to Prudential to make determinations such as the following: Is carpal tunnel syndrome a symptom of “multiple chemical sensitivity”? What jobs would be available to Mood? Would sedentary jobs available to Mood require use of computers given the prevalent use of modern technology in the workplace? Would her disabilities, if any, preclude her use of this and other office equipment necessary for employment?
VI. Conclusion
The matter is remanded for further consideration and for clarification of the record by Prudential. The case is not dismissed. Cf Nichols, 406 F.3d at 103-05 (distinguishing remand from dismissal for purposes of appeal). It shall remain closed administratively in this court unless restored to the active calendar on motion of any party.
SO ORDERED.