This case raises the question: what could be worse than having most of your home burn down in a fire? The answer, of course, is having the rest of it burn down a couple of days later in a second fire. What would make the situation dramatically worse, however, is if the fire department determined that the second fire was intentionally set (possibly by you) and called in federal authorities to investigate, thus requiring you to invest substantial energy, time and money defending against such allegations. Such a scenario would be particularly outrageous if the fire department did not actually believe that the second fire was intentionally set but was merely trying to draw attention away from the possibility that it had been negligent in putting out the first fire. According to Charles M. McDonald of Winnetka, Illinois, this is exactly what happened to him. McDonald responded by bringing a constitutional equal protection "class of one" claim in the Northern District of Illinois against the Winnetka Fire Department, following our precedent in Olech v. Vill. of Willowbrook, 160 F.3d 386 (7th Cir.1998), aff'd, 528 U.S. 562, 120 S. Ct. 1073, 145 L. Ed. 2d 1060 (2000). After extensive discovery, and an unsuccessful motion to dismiss, the district court eventually granted defendants' motion for summary judgment. For the reasons discussed infra, we affirm the decision of the district court because McDonald, in invoking the constitution, has failed to identify someone similarly situated but treated differently.

* Given the limited ground upon which we affirm the district court, our discussion of the facts of this case is relatively abbreviated. For a more detailed factual account, we refer the reader to the district court's lengthy discussion. See McDonald v. Village of Winnetka, No. 00 C 3199, 2003 WL 168637, at *1-*15 (N.D.Ill. January 23, 2003). We note that McDonald contends that the district court's recitation of the facts is one-sided and fails to construe the facts in the light most favorable to the non-moving party. Upon independent review, however, we find the district court's discussion of the facts of this case to be fair and balanced. There simply is not enough space in the Federal Reporters or on Westlaw's hard drives to discuss explicitly every fact and counter-fact which supports or affects each party's claim. Merely because a fact or argument has not been explicitly laid out does not mean that it has not been given serious consideration by the court. That being said, we take you to McDonald's first fire.

Sometime during the afternoon of May 10, 1999, a fire erupted at McDonald's home at 894 Sunset Road, Winnetka, Illinois. The Winnetka Fire Department (WFD) arrived at McDonald's home at approximately 4:09 p.m. Among those firefighters present were Ronald Colpaert, the Fire Chief; Scott Smith, the Deputy Fire Chief; and other firefighters from Winnetka and from various surrounding communities. Colpaert and Smith did not know McDonald before the May 10 fire. After more than four hours of fighting the fire, it was eventually extinguished at approximately 8:27 p.m. that same day.

In a subsequent investigation, WFD was able to determine that the May 10 fire started in the sunroom at the east end of the first floor of McDonald's home and was caused by a spark from an electrical outlet that ignited some paint stripping product. Fire, smoke and heat damaged various rooms on the first and second floors of McDonald's home, including the breakfast nook.

At the time of the May 10 fire, firefighters with the WFD knew that McDonald's home contained cellulose insulation that burned and smoldered. Cellulose insulation was present throughout the walls of the attic-level bedroom, the other rooms in the attic and above the kitchen and break-fast nook ceiling, among other places. The WFD had dealt with other insulation fires. In 1995, the WFD trained its personnel with videos that specifically pertained to attic fires and cellulose insulation. If not fully extinguished, cellulose insulation can smolder undetected for periods exceeding 2-3 days and can then ignite nearby combustible materials, such as beams or studs. Therefore, in accordance with common practice, a WFD firefighter returned to McDonald's home on the evening of May 10, to inspect the house for "hot spots." One hot spot was found and extinguished.

Diane Curtis, McDonald's wife, was outside the home on May 11 and 12, but did not report that she smelled smoke or any burning or smoldering material. Tom Robertson of the WFD was also at McDonald's home for several minutes on May 11, around 2:30 p.m. He viewed the exterior of the building from the street and did not see any signs of smoke. Various other employees of the Village of Winnetka were at McDonald's home on May 11 and 12, and did not report seeing or smelling any smoke or any burning or smoldering material.

At approximately 11:00 a.m. on May 11, Steven Strus, the General Adjuster employed by McDonald's insurer, Atlantic Mutual Insurance Company, met with the insureds at the house. They surveyed both the interior and exterior of the home and did not report seeing any smoke or smelling anything burning or smoldering. A disagreement arose between McDonald and Strus regarding the extent of the damage to the house caused by the first fire. McDonald voiced his opinion that the house was a total loss, but Strus did not agree.

McDonald was in his home on May 12 to let workers in between 9:00 a.m. and 9:30 a.m. He returned at about 11:30 or 11:45 a.m. and remained in the building until approximately 1:00 or 1:15 p.m. McDonald did not report seeing any smoke or smelling anything smoldering or burning at that time. At the scene of a fire that erupted later on May 12, however, McDonald told Colpaert that he might have seen a wisp of smoke or steam while he was home that afternoon. McDonald claimed that he did not report it at that time because he thought that his mind was playing tricks on him. Except for the WFD firefighter who detected and extinguished a hot spot the evening of May 10, no one who had either been in the building or who had viewed the outside of the building between the two fires reported seeing, smelling or otherwise sensing smoke or burning material in the building prior to the report of the fire. Strus, the insurance adjuster, reported that McDonald had called him at approximately 3:30 p.m. on May 12, and "was very adamant that the dwelling be considered a total loss."

Brian Funches, the mail carrier on the route that included McDonald's home reported having observed what he described as a gray Blazer containing a man and two children in the driveway of McDonald's home at about 4:30 p.m. on May 12, approximately a half hour before a fire was reported.

At approximately 5:08 p.m. on May 12, the WFD was dispatched to McDonald's home because of the subsequent fire. Colpaert commanded at the fire scene with Smith's assistance, along with other firefighters from Winnetka and from surrounding communities. Upon arriving at the scene, the WFD found that a large fire had vented itself through the roof toward the center of the structure. The fire was extinguished by approximately 8:50 p.m. Colpaert appointed WFD Captain Dale Solberg as the lead investigator of the cause and origin of the May 12 fire. At the scene of the fire, McDonald asked a number of the firefighters what had caused the fire, but they responded that they did not know. He may also have inquired about the WFD's potential liability if the fire were a rekindle.

Because Colpaert wanted a thorough and complete cause and origin investigation, he requested that additional fire investigators from other fire departments report to McDonald's home on May 12, to conduct a cause and origin investigation. This procedure is known as requesting a "fire investigator box." The State Fire Marshal was also called to conduct a cause and origin investigation. The fire investigator box for the May 12 fire dictated that an investigator from each of Northbrook, Northfield and Winnetka participate in the cause and origin investigation of the May 12 fire. These individuals and Mitchell Kushner, a Special Agent assigned to the Division of Arson Investigation for the Office of the Illinois State Fire Marshal (OISFM), arrived at the scene on May 12, 1999. Kushner, who happened to be a good friend of Solberg, had received extensive training in fire investigation, including investigation into the origins and causes of fires.

In October 1999, Solberg wrote the following email to an online community of arson investigators, in which he colorfully discussed the decision to call in outside investigators:

Let me take it one step further. Your department has a fire in a residential structure. Approximately 48 hours after the first fire, a second fire happens. Everyone is starting to think rekindle. Everyone around is thinking the same thing. The structure has a value of over $1,000,000.00. The homeowner says to the Fire Chief, "who is responsible for this, you?" He insinuates that your department could not extinguish the fire right the first time and that you should be responsible for the damage. I called investigators from 4 surrounding communities including the Office of the State Fire Marshal — Division of Fire Investigation. I put in charge an investigator from another community so as not to prejudice the investigation. Their conclusion — the fire was set — not a rekindle. Save face for my department.1

Pl. Rule 56.1(B) Statement ¶ 92. Between 8:50 p.m. on May 12 and 12:35 a.m. on May 13, Solberg, Kushner and Michael Roeder, another WFD fire investigator, investigated the fire. Roeder, however, testified that he did not actually participate in determining the cause and origin. The WFD was assisted by Wayne Leucht and Bernie Arends, investigators from Northbrook, as well as Tom Burke of Northfield. The investigators were assisted by an evidence technician from the Winnetka Police Department, a photographer from the WFD, and an accelerant sniffing dog (Nikki) from OISFM. Samples taken from the home to test for accelerants proved negative, although the absence of accelerants would not rule out the possibility that the fire was deliberately set by someone lighting combustibles, such as paper.

Solberg first sought to determine the place of the origin of the May 12 fire by examining the remains of McDonald's home, from the area of least damage to the area of most damage, and by talking with eyewitnesses at the scene. Solberg also took into consideration the effect of the May 10 fire. Before leaving McDonald's home, Solberg concluded that although the attic area was damaged at least to the same extent as the breakfast nook, the origin of the May 12 fire was a stud pocket at the base of the west wall of the breakfast nook, a wall that continues down to the lowerlevel bar. He formed this conclusion based on fire and burn patterns, firefighting tactics and witnesses' statements.

Kushner also conducted an investigation of the May 12 fire by surveying the entire scene. He testified that he consulted with firefighters and investigators who had been on the scene at the May 10 fire in order to differentiate damage caused by the two fires. Domingo Hernandez, a worker at McDonald's home who witnessed the outbreak of the second fire, was interviewed by Kushner and others on the 12th, after the fire. Solberg Dep. at 32; Kushner Dep. at 49. According to Hernandez, he was working in the back east corner of the backyard with several other individuals who thought that they smelled smoke. They walked up to the house and saw smoke and fire through the plywood of a boarded-up window in the south window of the breakfast nook area. Solberg Dep. at 32-33. However, in a subsequent declaration, Hernandez averred that he saw flames coming from the southeast corner of the roof over the breakfast nook and that he did not see flames or smoke coming from the south window of the breakfast nook.2 It was Hernandez who telephoned his boss, and his boss telephoned the fire department. Hernandez says that approximately half hour elapsed from the time he observed smoke until the time he called his boss.

Kushner, along with other investigators, noted an area of heavy charring and burning in the west wall of the break-fast nook. This wall had no drywall remaining on it and consisted solely of exposed stud beams. Kushner and other investigators focused on this area as a possible point of origin of the second fire. Kushner was told that the break-fast nook sustained smoke damage in the first fire but did not sustain fire damage. He investigated the breakfast nook's west stud wall and structural members and examined the area above the stud pocket of the west wall of the breakfast nook by shining a light on the area above the stud wall. No light was visible from above, indicating to him that no burn material had fallen from another area down into the base of the stud pocket. Based on the burn patterns, firefighting tactics and witnesses' statements, Kushner also concluded that the May 12 fire originated at the base of the west wall of the breakfast nook and that the fire had spread from the stud pocket of the breakfast nook into the bar area. The Northbrook investigators agreed that this was the point of origin because of the extensive damage to the west wall of the breakfast nook and because it appeared to be the lowest point of burning in the specific area which they had examined.

After determining the point of origin of the second fire, Solberg and Kushner undertook the investigation of the cause of the May 12 fire. The cause of any fire may fall into one of three broad categories: accidental, incendiary and undetermined. The WFD's expert has written in a publication entitled, Fires and Explosions: Determining Cause and Origins, "Firemen and investigators ... should be extremely careful about attributing the fire to any cause whatsoever unless they have corroboration and are absolutely sure of the cause." Pl. Rule 56.1(B) Statement at ¶ 53. He has further explained that "[t]o properly and effectively establish the one and only true cause, the investigator must examine, evaluate, and rule out all other possible causes." Id. There is "no presumption that a fire has been intentionally set. On the contrary, the presumption of innocence which belongs to the accused carries with it a presumption that the fire is of accidental or providential origin." Id. at ¶ 54. Additionally, one of the textbooks in WFD's library warns, "[i]f doubt exists about the cause of a fire, an investigator must rule out all possibilities of accidental fires before drawing conclusions, forming opinions, or proceeding with an arson investigation and trying to determine an incendiary cause." Id. at ¶ 57.

Solberg and Kushner ruled out the possibility that gas, electricity or lightning caused the May 12 fire because the utilities at McDonald's home had been turned off following the May 10 fire, and the weather service had reported no lightning strikes in the area. They ruled out the possibility that smoking materials caused the fire because witnesses reported that no one was smoking in the house and because there was no evidence of smoking materials in the area of origin. They also ruled out the possibility of spontaneous combustion because there was no evidence of spontaneous combustion in the area of origin.

Solberg and Kushner ruled out the possibility that a rekindle of the May 10 fire caused the May 12 fire because they concluded that the area of origin of the second fire was physically remote from the area of the first fire and because there was no burning in the area of origin of the second fire which remained from the first fire. Moreover, in the experience of Smith, Solberg and Kushner, no structure had rekindled after the extinguishment of a fire at the same structure approximately 48 hours earlier. The possibility of a rekindle never entered Burke's mind because of the 48-hour time span between the two fires. On May 20, 1999, the local newspaper, Winnetka Talk, quoted Smith as saying, "It's not a rekindle. Rekindles don't happen 49 hours after the first fire." Def.'s 56.1 Statement ¶ A78; Smith Dep. at 104-09. At his deposition, Smith testified that he was speaking generally and not specifically about the May 12 fire. Id. Solberg admitted that the mere fact that a second fire flares as long as two days after the first fire does not, in and of itself, enable an investigator to rule out a rekindle.

McDonald seems to argue that smoldering cellulose insulation may have fallen down from the attic, thereby igniting a fire near the breakfast nook. Solberg and Kushner claim, however, that they also ruled out the possibility that combustible material had dropped down into the stud bay area of the breakfast nook to cause the May 12 fire because the structural materials above the open stud bay were intact following the May 10 fire and, thus, could not have fallen down. Moreover, no burned material, such as wood members or insulation, was present in the stud bay following the second fire. Solberg, Kushner and John Agosti, a fire investigator for Atlantic Mutual, all concluded that the May 12 fire was incendiary. Northbrook's investigators agreed, and Burke, Northfield's firefighter, considered the May 12 fire suspicious.

3. Involvement of the Bureau of Alcohol, Tobacco and Firearms

On May 17, 1999, Kushner and Solberg decided to call Agent Glowski of the Bureau of Alcohol, Tobacco and Firearms (ATF) to relate their findings regarding the cause of the May 12 fire and to suggest that the ATF might want to look into it. Solberg testified that he contacted the ATF to tap into its resources and expertise. It is apparently not uncommon for local fire departments to write a preliminary report that a fire is incendiary and then call in the ATF, and even McDonald's expert did not criticize the WFD or the State Fire Marshal for contacting the ATF under the circumstances present in this case.

Kushner spoke with Glowski about the May 12 fire, briefly summarizing what he found and expressing his conclusions. He told Glowski that he thought the May 12 fire was an arson, and mentioned the conversation McDonald had had with Strus shortly before the fire broke out. Solberg also informed Glowski that he did not believe that the May 12 fire was a rekindle and that he believed the fire was incendiary. He told Glowski that the opinions of the investigators as to the cause and origin were preliminary. Neither Solberg nor Kushner mentioned any suspects in connection with the fire. The ATF got involved in the investigation because Kushner indicated that the fire was "a possible arson." Def. 56.1 Statement at ¶ A60.

Once the ATF was called into the investigation, the ATF, itself, determined the scope and nature of its involvement. Solberg provided the ATF with information that he had, but the ATF was free to conduct its own cause and origin investigation if it felt that appropriate. No one from the WFD objected to the ATF's conducting a cause and origin investigation, and no one told the ATF not to conduct such an investigation. Within one or two months of the ATF's involvement, Glowski enlisted Special Agent John Mirocha, an ATF cause and origin specialist, to review the cause and origin of the fire. Solberg was "open and forthright" and complied with everything that the ATF asked of him. Id. at ¶ A67. Neither Kushner nor the WFD ever expressed to Glowski any animus toward McDonald or his family. After reviewing photographs, sketches and plans of the house and discussing the two fires with Solberg, Mirocha agreed with the cause and origin determination Solberg and Kushner had made.

Glowski interviewed McDonald on June 15, 1999. This was the first time that McDonald learned that the investigators believed that the cause of the May 12 fire was incendiary. During the ATF's investigation, Glowski determined that David Curtis, McDonald's brother-in-law, owned a gray Blazer and that he was in the area of the May 12 fire at the time the fire began, although he did not work or reside in the area. When Glowski interviewed Curtis he appeared to Glowski to be extremely nervous. Glowski served Curtis with a grand jury subpoena. Curtis, however, did not appear before the grand jury. In mid-to- late 1999, an Assistant United States Attorney spoke with McDonald's counsel about an investigation by the U.S. Attorney's Office into a possible arson at McDonald's home and into McDonald's possible involvement in the arson.

On January 5, 2000, Winnetka issued a nuisance violation to McDonald, ordering him to demolish his home. Apparently, the official who issued the nuisance violation, had consulted with the WFD before issuing it. Neither Solberg nor Smith informed this official that the ATF might need to get into the house at some point to physically inspect it in the course of the ongoing investigation.

In March 2000, McDonald's counsel and a private cause and origin investigator employed by McDonald made a presentation to the U.S. Attorney's Office and the ATF, expressing his opinion that the May 12 fire had been a rekindle. After this presentation, cause and origin experts went to the scene of the house. The ATF concluded, based upon its examination of the scene, that, while it could not eliminate the possibility of a rekindle, it also could not rule out that the May 12 fire was intentionally set "by someone piling up available combustibles in that stud bay and lighting it." Id. at ¶ A70. The ATF did not conclude that the May 12 fire was a rekindle, nor did it eliminate arson as a cause of the May 12 fire. The ATF's uncertainty regarding the cause of the May 12 fire was due not to any questionable quality of the previous fire investigation but, at least in part, to the significant alteration of the scene in the months since the May 12 fire.

When the ATF shared its opinion with Solberg and Colpaert, Solberg refused to change Winnetka's National Fire Incident Reporting System (NFIRS) report to reflect that the cause of the May 12 fire was indeterminate. Colpaert ordered him to do so, but Solberg refused and Colpaert made the change himself. Glowski wanted to continue to investigate following the ATF's visit to McDonald's home because the occupants of the gray Blazer seen at the residence half an hour before the May 12 fire had not been identified. The Assistant U.S. Attorney who was investigating the case telephoned McDonald's counsel, however, and informed him that the investigation would be dropped.

The defendants involved in this litigation apparently did a questionable job of documenting McDonald's fires. On June 24, 1999, the WFD conducted a Post-Incident Analysis (PIA) of the May 10 fire. The WFD Standard Operating Procedure (SOP) # 3.02.01 requires each PIA that is conducted to be documented by completing a PIA Fact Sheet, a Post-Incident Worksheet and a Post-Incident Summary. In this case, however, there is no record of what transpired at the PIA. Colpaert testified that the WFD SOPs are simply guidelines — not requirements.

Less than a week after the PIA, however, shift commander John Gaughan ordered a videotape from an outside vendor entitled, "Overhaul: Loss Control," which demonstrated, inter alia, "the practice of searching through an extinguished fire scene to find hidden fires that could rekindle." Pl. Rule 56.1(B) Statement ¶ 107. In the following weeks, WFD firefighters who were present at the May 10 fire watched the video — some more than once.

Under Illinois law, fire departments are obligated to report all fire incidents to the Office of the State Fire Marshal (OSFM). The NFIRS reporting system is used to meet this obligation. The NFIRS incident report for the May 12 fire, dated May 13, was completed and signed by Solberg. The report lists the "Ignition" factor as "Incendiary" and the "Type of Action Taken" as "Investigation Only." Id. at ¶ 132. Although the NFIRS handbook calls for the "Officer in Charge" section of the document to bear the date that the officer creates the report, Solberg did not sign the report on May 13, and he could not recall when he created it. The NFIRS report was also apparently amended multiple times, but only the most recent amended version could be recovered.

An NFIRS report for the May 12 fire was also prepared by the City of Highwood. Highwood responded to the May 12 fire after the WFD asked for other municipalities to assist. Highwood was not dispatched to the fire scene. Instead, Highwood reported to the WFD fire station for station coverage. Highwood's original NFIRS report lists the cause of fire as a rekindle. Highwood Chief Ron Pieri, however, testified that he changed the original report because it did not correctly reflect Highwood's role in the May 12 fire. The Highwood firefighter who had prepared the original report testified that he was in the kitchen of the WFD when he had a conversation with an unidentified WFD dispatcher. Based on this brief conversation with the dispatcher, the Highwood firefighter prepared the report and marked rekindle as the cause of the May 12 fire. When Pieri changed the report, he did not explain to its original author that he was going to change it, did not mark the change field, did not change the date and did not put his name on the new version. The revised version of Highwood's NFIRS report did not include a cause determination.

A WFD SOP requires that the fire investigator "[p]repare the necessary forms, sketches and report to record the facts determined by his investigation." (SOP 401). This is known as a "Fire Investigation Report" (FIR) and is not the same thing as a NFIRS report. Although Solberg prepared a FIR for the May 10 fire, he did not prepare a FIR for the May 12 fire because he intended to rely on Kushner's report and did not have time to prepare a separate one.

The OSFM requires that its investigators complete a single-page initiation form that is intended to provide the basic information about the time, date, location and cause of the fire incident. After the form is completed, the investigator provides a copy of the form to the OSFM. On May 13, Kushner began filling out an initiation form for the May 12 fire. He completed the form on or about May 17; however, for reasons unknown, neither Kushner nor the OSFM was able to produce this form. Kushner did, however, prepare an investigatory report, which is dated May 27.

On May 25, 2000, McDonald brought suit against the defendants in the Northern District of Illinois, bringing an equal protection claim and a state law claim for intentional and reckless infliction of emotional distress. McDonald argued that "[f]or ten months, [he] lived in a state of constant dread, fearing what would happen to his wife, children, including a disabled child, and livelihood." Pl. Br. at 10. The district court denied the defendants' motion to dismiss on May 2, 2001. However, it granted the defendants' motion for summary judgment on October 25, 2002. This appeal followed.

We review a grant of summary judgment de novo. See Wyninger v. New Venture Gear, Inc., 361 F.3d 965, 974 (7th Cir.2004). In doing so, we construe all facts in favor of the non-moving party. See id.; Rogers v. City of Chicago, 320 F.3d 748, 752 (7th Cir.2003). However, "we are not required to draw every conceivable inference from the record." Bell v. Duperrault, 367 F.3d 703, 707 (7th Cir.2004), quoting McCoy v. Harrison, 341 F.3d 600, 604 (7th Cir.2003). Inferences that are supported by only speculation or conjecture will not defeat a summary judgment motion. Id. Summary judgment is proper "if the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to judgment as a matter of law." Fed.R.Civ.P. 56(c) (2003).

This Court has recognized equal protection claims brought by a "class of one," although we have acknowledged that it is difficult to succeed with such a claim. See Levenstein v. Salafsky, 164 F.3d 345, 353 (7th Cir.1998) (citing Esmail v. Macrane, 53 F.3d 176, 179 (7th Cir.1995)). A class of one equal protection claim may be brought where (1) the plaintiff alleges that he has been intentionally treated differently from others similarly situated and (2) that there is no rational basis for the difference in treatment or the cause of the differential treatment is a "totally illegitimate animus" toward the plaintiff by the defendant. See Vill. of Willowbrook v. Olech, 528 U.S. 562, 564, 120 S. Ct. 1073, 145 L. Ed. 2d 1060 (2000); Nevel v. Vill. of Schaumburg, 297 F.3d 673, 681 (7th Cir.2002); but see Bell, 367 F.3d at 709-10 (Posner, J., concurring); and Hilton v. City of Wheeling, 209 F.3d 1005, 1008 (7th Cir.2000) (rejecting the "no rational basis" approach because "[i]f a merely unexplained difference in police treatment of similar complaints made by different people established a prima facie case of denial of equal protection of the laws, the federal courts would be drawn deep into the local enforcement of petty state and local laws.").3

In the present case, we find that McDonald has failed to identify someone who is similarly situated but intentionally treated differently than he. There is no precise formula to determine whether an individual is similarly situated to comparators. See Barrington Cove, LP v. R.I. House & Mortg. Fin. Corp., 246 F.3d 1, 8 (1st Cir.2001). As a general rule, whether individuals are similarly situated is a factual question for the jury. See Harlen Assoc. v. Vill. of Mineola, 273 F.3d 494, 499 n. 2 (2d Cir.2001). However, a court may properly grant summary judgment where it is clear that no reasonable jury could find that the similarly situated requirement has been met. Id.; Bell, 367 F.3d at 709-10 (affirming district court's grant of summary judgment where the plaintiff failed to raise a triable issue as to whether he was "similarly situated" to comparators); Purze v. Vill. of Winthrop Harbor, 286 F.3d 452, 455-56 (7th Cir.2002) (same).

It is clear that similarly situated individuals must be very similar indeed. Purze, 286 F.3d at 455 (holding that in order to be considered "similarly situated," comparators must be "prima facie identical in all relevant respects"). For instance, in the employment context, we have said that "[m]ore evidence than the mere fact that other employees were not discharged for at best arguably similar misconduct must be demonstrated ...." Hiatt v. Rockwell Int'l Corp., 26 F.3d 761, 771 (7th Cir.1994). Where the plaintiff claims that he was disciplined more harshly than others, he must show that he is "similarly situated to persons outside the protected class with respect to performance, qualifications, and conduct." Radue v. Kimberly-Clark Corp., 219 F.3d 612, 617 (7th Cir.2000) (citation omitted). "This normally entails a showing that two employees dealt with the same supervisor, were subject to the same standards, and had engaged in similar conduct...." Id. at 617-18; see also Cruz v. Coach Stores, 202 F.3d 560, 568 (2d Cir.2000) (holding that for purposes of a claim of selective enforcement of a policy calling for termination of employees engaged in fighting or verbal assault, a plaintiff who had engaged in fighting was not situated similarly to those who engaged in a verbal assault).

We have imposed on plaintiffs an equally high burden with regard to establishing someone who is similarly situated in the context of "class of one" equal protection claims. See Bell, 367 F.3d at 708-09 (individuals were not similarly situated in a "class of one" equal protection case where individuals submitted applications for pier extensions at different times, requested different extensions, or requested to replace existing structures rather than build new ones); Purze, 286 F.3d at 455 (individuals were not similarly situated in a "class of one" equal protection case where the individuals submitted different variances than the plaintiff requested, submitted their plats during different time periods, or had requests granted by different and previous Boards); see also Manley v. Thomas, 255 F. Supp. 2d 263, 267-268 (S.D.N.Y.2003) ("[T]he number and variety of factors bearing on the seriousness of the underlying offense and the likelihood that an offender will be a danger to the community make it impossible to conclude, on the bases of the sketchy data presented, that petitioner has been singled out from among all homicide offenders for disparate treatment."); Wantanabe Realty Corp. v. City of New York, 2003 WL 21543841, at *14, 315 F. Supp. 2d 375, 395 (S.D.N.Y.2003) ("[Plaintiff] would have to show that other property owners whose structures for some reason were unwanted by high City officials were not also subjected to sham inspections and politically driven demolition processes.")

In the present case, McDonald attempts to meet the "similarly situated" requirement by arguing that the WFD had an established procedure that it followed when investigating every fire in Winnetka other than his. Reply Br. at 4 ("Defendants deviated from these norms only in their investigation of the second fire at McDonald's home and thus treated McDonald differently from every other Winnetka homeowner at whose residence defendants investigated a fire."). This procedure, argues McDonald, is to rule out all possible accidental causes before deeming a fire arson. Id. Therefore, McDonald argues that he is similarly situated to every resident who had a residential, structural fire, but that he was treated differently because only in his case did the WFD fail to rule out all non-arson causes before making an arson determination.4 Id.

There are several problems with this argument. First, McDonald has failed to present any competent evidence that the WFD did not rule out all possible accidental causes, including rekindle, in this case. In fact, the evidence upon which McDonald relies demonstrates that the WFD did rule out rekindle. For instance, McDonald relies heavily on Solberg's email as a sort of "smoking gun." See supra, Slip op. at 995-96. However, this email suggests that the WFD initially considered rekindle as a possible cause but ultimately eliminated it. McDonald further relies on Smith's public comments right after the second fire, in which he discussed one reason that he believed the fire was not a rekindle. Additionally, McDonald presents "cause and origin" reports, which McDonald claims were edited to remove references to rekindle. If true, these deletions would also suggest that rekindle had been considered at one point but was ultimately eliminated. Both Solberg and Kushner testified that the investigative team eliminated rekindle. Kushner Dep. at 107-09, 176; Solberg Dep. at 89, 156. Their testimony was consistent with contemporaneous ATF reports of conversations each member had had with the ATF. See Glowski Dep., Ex. 9 (ATF Report 001) ("Kushner did not believe the second fire was a rekindle from the first fire."); Ex. 10 (ATF Report 002) (noting that Solberg "eliminated all possible causes for the start of the fire. He also said it was not rekindle from the first fire."). Bernie Arends, a fire investigator from Northbrook, agreed. See Arends September 25, 2002 Decl. ("It was the collective conclusion of the six investigators at the scene on May 12 that there was no reasonable possibility that fire rekindled in the stud pocket. There was therefore no reason to conclude that a rekindle caused the May 12 fire.").

Further, McDonald has presented no evidence of any particular procedure which should be used to "rule out" rekindle. In other words, McDonald does not identify any additional steps that the WFD would have had to take to officially rule out rekindle.5 This makes it difficult, if not impossible, for McDonald to effectively argue that, as a procedural matter, the WFD failed to rule out rekindle. It also supports the conclusion that McDonald's real argument is not procedural (i.e., that the WFD failed to rule out rekindle) but substantive (i.e.,that they improperly ruled out or should not have ruled out rekindle).6

Second, even assuming McDonald had presented some competent evidence that WFD failed to rule out rekindle, he presented no evidence of anyone "similarly situated" but treated differently. Instead, McDonald presents seven "cause and origin" reports prepared by the WFD for previous fires that it investigated (some dating back almost twenty years). Not surprisingly, these reports suggest that the WFD believed that it properly eliminated all non-arson causes in prior fires. This evidence is of little relevance, however, because it does not establish whether the WFD actually eliminated all non-arson causes in prior investigations. In other words, they do not show whether these comparators were actually treated differently. The homeowners in those cases may have been just as displeased with the WFD's investigation as McDonald is in the present case. It must be remembered that the WFD argues that it eliminated all non-arson causes for McDonald's fire as well. Therefore, even if a jury could find that the WFD failed to rule out a non-arson cause in this case, there is no way that the jury could determine whether there were similar problems with its other investigations. The simple fact that the WFD thinks it did a thorough job with respect to its earlier investigations is no more dispositive than its belief that it did a thorough job in the present case.

These other investigations are not similarly situated to McDonald's for another reason. None of these other investigations involved possible rekindle. Eliminating rekindle as a possible cause of a fire is not as simple or objectively demonstrable as eliminating other causes. For instance, one can eliminate lightning as a cause simply by checking the weather as recorded by a weather service. See Solberg Dep. at 104. Similarly, natural gas can be eliminated as a possible cause where, as in this case, the gas had been turned off well before the fire. Id. at 101; Pl. Rule 56.1(B) Statement, Ex. 37 (Kushner's May 27, 1999 Report). Smoking materials can be eliminated if nobody was in the house smoking. See Solberg Dep. at 102. Eliminating rekindle, however, is a more subjective and imprecise task, and as this case demonstrates, reasonable experts can differ on the matter. Therefore, the possibility that the defendants may have been able to eliminate other accidental causes in previous investigations would not make it surprising that they "failed" to rule out rekindle with the same degree of precision in this case.

McDonald's real argument in this case is that the WFD improperly ruled out rekindle — not that it failed to rule out rekindle. As discussed supra, McDonald has presented evidence only of the former and not the latter. Moreover, it would be irrelevant to an equal protection claim that the WFD intentionally failed to rule out rekindle if it would have been proper to rule it out anyway. See Nevel, 297 F.3d at 681 ("[I]f the government would have taken the action anyway, the animus will not condemn the action.").7

The fact that McDonald must ultimately show that the WFD improperly ruled out rekindle in this case is helpful in understanding the "relevant respects" in which comparators must be similarly situated. In order to identify in what respects comparators must be similarly situated, we may draw on the well-established law regarding "selective prosecution" since such claims illustrate principles applicable here. See Levenstein, 164 F.3d at 353 ("Levenstein's allegation that his [injuries] were caused not by allegations that he had committed acts of sexual harassment, but rather arose from the sheer vindictiveness the University officials felt toward him ... falls within the context of selective prosecutions prohibited by the Equal Protection clause."); Cobb v. Pozzi, 352 F.3d 79, 99 (2d. Cir.2003) (discussing the narrow difference between a selective prosecution claim and an equal protection "class of one" claim).

The cases discussing selective prosecution have made it clear that in order for an individual to be similarly situated for such purposes, the evidence against the comparator must be "as strong [as] or stronger" than that against the person arguing there has been an equal protection violation. See United States v. Smith, 231 F.3d 800, 810-11 (11th Cir.2000) (citing United States v. Armstrong, 517 U.S. 456, 465, 116 S. Ct. 1480, 134 L. Ed. 2d 687 (1996)); United States v. Monsoor, 77 F.3d 1031, 1034 (7th Cir.1996) (defendant could not show that he was situated similarly to his uncharged accomplices for purposes of making a selective prosecution claim where the government lacked sufficient evidence to charge the other individuals); United States v. Hayes, 236 F.3d 891, 895 (7th Cir.2001) ("[I]n order to obtain discovery on such a claim, a defendant must at least produce some evidence that similarly-situated defendants of other races could have been prosecuted but were not."); United States v. Davis, 339 F.3d 1223, 1228 n. 3 (10th Cir.2003); United States v. Serafino, 281 F.3d 327, 331 (1st Cir.2002); United States v. Hastings, 126 F.3d 310, 315 (4th Cir.1997) ("[D]efendants are similarly situated when their circumstances present no distinguishable legitimate prosecutorial factors that might justify making different prosecutorial decisions with respect to them.") (internal quotation marks omitted). Therefore, in this case, McDonald must present other investigations conducted by the defendants in which the evidence of rekindle (or another non-arson cause) was of a similar or lesser weight and the evidence of arson was of a similar or greater weight, and that, nonetheless, the defendants did not deem the fire to be arson and did not refer the matter to the ATF.

McDonald has failed to identify a similarly situated individual most fundamentally because in none of the cases identified by McDonald was the evidence of a non-arson cause of a similar or lesser weight or the evidence of arson "as strong or stronger" than in McDonald's situation. See Smith, 231 F.3d at 810-11. Unlike the present case, in each of the WFD cases cited by McDonald, the cause of the fire was the obvious and admitted negligence of the homeowners or their agent.8 In the first case, the housekeeper admitted to starting the fire accidentally while working near the stove. Pl. Rule 56.1(B) Statement, Ex. 29. In the second case, the fire was started when the homeowners were barbecuing, and they removed ten charcoal briquettes from the grill and put them in a brown paper bag in their kitchen. Id. at Ex. 30. In the last case, the fire was started when the homeowner threw a rag soaked with paint thinner into a trash can in his garage which unfortunately also contained approximately three gallons of saw dust. Id. at Ex. 31.9 Based on the investigation reports, the WFD did not even consider arson as a possibility in any of these cases, likely because the causes of the fires were apparent. In each of these cases, the homeowners suggested their own accidental involvement in the fire and unlike the present case, in each of these cases there was no evidence presented of any fact contraindicating the possible non-arson cause.

While in the present case there may have arguably been evidence of rekindle, this cause certainly was not obvious. Even the ATF ultimately was unable to conclude that the second fire was or was not a rekindle. In fact, no party has ever conclusively determined the cause of this fire to be rekindle. First, the WFD determined that the point of origin of the second fire was a portion of McDonald's home that was unaffected by the first fire, seemingly ruling out rekindle. Regardless whether this determination was proper, the fact that the WFD articulated a basis for eliminating the non-arson cause itself distinguishes this case from the other investigations.

Second, the subsequent fire started more than 48 hours after the first fire. While a rekindle can apparently occur up to three days after an initial fire, such an occurrence must be rare, as none of the investigators in this case had ever experienced a rekindle occurring after such a significant gap in time. Agent Glowski of the ATF had never investigated such a rekindle and knew of no rekindle that occurred 48 hours after a fire. Glowski Dep. at 204. Kushner, in fact, did not believe that a rekindle could occur after such a time gap, at least given the facts of this case. See Kushner Dep. at 155; see also Arends September 25, 2002 Decl. at 11 ("In my 30 years of experience in the fire service, I have never experienced a rekindle anywhere near 48 hours after the first fire. Rekindles are very unusual (I have seen it happen only twice in my career) and rekindles occurring more than one or two hours after the first fire are especially extraordinary.") Third, during the 48-hour gap between the fires, a variety of individuals had been in and around the residence at different points and none reported to the WFD any smoke, smoldering or anything else that would suggest that the fire from May 10 was still burning. Fourth, the WFD was confident that the earlier fire had been completely extinguished.

Additionally, unlike the other investigations identified by McDonald, in the present case, there were what could be called suspicious circumstances surrounding the fire. Most notably, McDonald had multiple conversations with his insurer after the first fire, in which he was apparently very adamant in his attempts to convince the insurer to declare his home a total loss — a request which the insurer refused. The last of these conversations occurred shortly before the second fire began. Additionally, McDonald made a visit to the remains of his home shortly before the second fire.

It is also noteworthy that in the other investigations relied upon by McDonald, the investigating officer and the officer who prepared the investigative report was Michael Roeder. Roeder, however, did not participate in determining the cause and origin of the May 12 fire. Roeder Dep. at 204. To the contrary, it was Solberg's job to investigate the May 12 fire. Id. "This omission alone probably precludes a showing of similarity because when different decision-makers are involved, two decisions are rarely similarly situated in all relevant respects." Radue, 219 F.3d at 618 (internal quotation marks and citations omitted); Purze, 286 F.3d at 455 (noting that individuals were not similarly situated where, inter alia, they had their plat requests granted by different and previous Boards). In this case, for instance, Roeder may simply be a more thorough, competent or conservative investigator than Solberg, and this alone may explain any alleged difference in McDonald's treatment.10

Further, none of the cases identified by McDonald involved second fires, so there was naturally no discussion of rekindle in the corresponding reports. As discussed, it can be more difficult to precisely rule out rekindle than other accidental causes. Moreover, unlike the other cases, in the present case, an Illinois State Fire Marshal investigated the same fire and reached the same conclusion as did the WFD. Similarly, the ATF investigated and was unable to rule out arson as a possible cause. Finally, as noted supra, there is no way for this Court or a jury to determine whether the WFD's investigation of these other fires was as sharply criticized as its investigation here. We know nothing more about these investigations than what the WFD printed in its reports.

At this point we believe it is important to step back a bit to avoid losing sight of the forest for the trees. The reason that there is a "similarly situated" requirement in the first place is that at their heart, equal protection claims, even "class of one" claims, are basically claims of discrimination. See Olech, 528 U.S. at 564, 120 S. Ct. 1073 ("[T]he purpose of the equal protection clause of the Fourteenth Amendment is to secure every person within the State's jurisdiction against intentional and arbitrary discrimination") (citations omitted); City of Cleburne v. Cleburne Living Ctr., 473 U.S. 432, 439, 105 S. Ct. 3249, 87 L. Ed. 2d 313 (1985) ("The Equal Protection Clause of the Fourteenth Amendment ... is essentially a direction that all persons similarly situated should be treated alike."); E.E.O.C. v. Elrod, 674 F.2d 601, 604 (7th Cir.1982) ("[Protection against] discriminatory government conduct ... is the very essence of the guarantee of `equal protection of the laws' of the Fourteenth Amendment."). Even if McDonald was wronged here, we do not believe that he has shown the wrong to be discriminatory in nature. See Bell, 367 F.3d at 711-12 (Posner, J., concurring) ("[I]rrational differences in treatment having nothing to do with discrimination against a vulnerable class abound at the bottom rung of law enforcement."). Every time an actor commits a tort, he may be treating the victim differently than he frequently treats others, simply because tortious conduct is by nature a departure from some norm. Nonetheless, the purpose of entertaining a "class of one" equal protection claim is not to constitutionalize all tort law nor to transform every claim for improper provision of municipal services or for improper conduct of an investigation in connection with them into a federal case. Id. ("It is highly unlikely that the Supreme Court intended in Olech to open the door to such cases."). Therefore, we believe a meaningful application of the "similarly situated" requirement is important to avoiding a distortion of the scope of protection in "class of one" equal protection claims. In this case, we find that McDonald has failed to present evidence of a similarly situated individual and we AFFIRM the district court's grant of summary judgment on this basis.11