EASTERBROOK, Circuit Judge.
Armando Chaidez, Manuel Chavira, and Lilia Silva (Chavira’s step-daughter) were convicted of conspiracy to possess heroin, cocaine and marijuana, with intent to distribute. Chaidez and Chavira were also convicted of possessing these drugs with intent to distribute them (the jury acquitted Silva of this charge), and Chaidez was convicted on a third count of possession with intent to distribute cocaine. The Chicago police found most of the drugs in a stash house leased by Chavira and Silva, after obtaining Silva’s consent to search the house. The principal question is whether the detention leading to the discovery of the drugs was a “reasonable” seizure.
I
The Chicago police received a tip (from a source “proven reliable in the past”) that Chaidez was a large-scale heroin dealer. The tipster provided no details. A check revealed that Chaidez’ name came up in a drug investigation involving two others unrelated to this case. The police decided to set up a “moving surveillance” of Chaidez, a procedure involving multiple vehicles in radio contact.
Surveillance began at 8:30 a.m. when Chaidez left his home. He carried a small plastic bag; no agent could see its contents. Chaidez got into the passenger side of a car and was driven to a west side restaurant, La Fonda del Requeredo. Chai-dez spoke briefly with the driver, then got out of the car and entered the restaurant through the back door. He emerged minutes later through the front door. Still carrying a plastic bag, Chaidez walked to his car, a white Cadillac parked nearby. He drove circuitously for awhile, going south, then west, then north. After gassing up his car, he eventually settled on north, driving to a north side lounge, La Hacienda. According to police testimony, drug dealers frequent both La Fonda del Requeredo and La Hacienda.
Chaidez stayed at La Hacienda for less than five minutes, emerging with two unknown associates. Agent Guadalupe Rodriguez observed the three “looking around a lot at cars going by. More than the normal way most persons do.” He assumed they were looking for surveillance (wisely, but unsuccessfully, as it turns out). After a brief conversation Chaidez again drove evasively, frequently changing directions. He managed to grope northward, however, and stopped in the middle of a side street. Chavira and Silva then entered the picture.
Chavira and Silva spoke to Chaidez through the window of the Cadillac. They then got into their own car (a Pontiac) and drove off, with Chavira driving. Chavira and Silva drove to a gas station; Chaidez followed. Chaidez got out of his car and walked over to the Pontiac to converse briefly with Chavira. No one bought fuel, although Chavira checked the air in his tires. They left after a few minutes, driving in different directions. The surveillance stayed with Chaidez.
Chaidez drove east for a few minutes. He then abruptly made a U-turn and proceeded west. He entered the Kennedy Expressway, where he varied his speed considerably (driving at the speed limit for awhile, then switching to the right lane and dropping to 40 m.p.h., then repeating the process). The police had some trouble with this technique, for they either had to pass or to change speed suddenly and reveal the tail. Chaidez turned off on the Edens Expressway, which he left at Lake-Cook Road. Here the Pontiac rejoined Chaidez’ Cadillac.
The two cars drove, Pontiac in the lead, to a house on Weiland Road. All three defendants went inside. Chaidez again held a small plastic bag in his hand. More than a half hour later, Chaidez left the house (sans plastic bag, this time) and drove north, followed in five minutes by Chavira and Silva. Both cars stopped in a parking lot. Chaidez made a telephone call, and then the defendants drove off again “at a high rate of speed”. The police, thinking their cover blown, decided to [1196]*1196stop the cars. They did so by blocking the narrow road in both directions by turning their cars sideways. The five agents got out of their cars, guns drawn but pointed downwards, and proceeded to interrogate the defendants.
Agent Rodriguez, the only Spanish-speaking agent at the scene, went to the passenger side of the Pontiac to question Silva. He identified himself and asked Silva her name. She answered truthfully, and he asked if she lived in the Weiland Road house. She said she rented it for her father, and had come to do his laundry. He read her Miranda warnings, which she understood, and then asked if she would consent to a search of the house. She said yes, so he took her over to a police van, where she signed a consent to search form. At some point the agents conducted a pat-down search. Agent Rodriguez radioed to another agent at the house that consent had been obtained and the search could begin. The police found more than just dirty linen.
What happened to Chaidez and Chavira is less clear, and less relevant. If the seizure of Silva was justified, and her consent valid, then all three are sunk. Silva’s consent cannot be the fruit of an illegal seizure of Chaidez or Chavira, since that seizure did not lead to the questioning of Silva. Wong Sun v. United States, 371 U.S. 471, 83 S.Ct. 407, 9 L.Ed.2d 441 (1963); Ker v. Illinois, 119 U.S. 436, 7 S.Ct. 225, 30 L.Ed. 421 (1886); Frisbie v. Collins, 342 U.S. 519, 72 S.Ct. 509, 96 L.Ed. 541 (1952). Still, the treatment of Chaidez and Chavira sheds light on the environment surrounding Silva, and Chaidez does challenge a subsequent search of his own apartment, so we glean what we can from the record. Chaidez and Chavira were ordered out of their cars and “searched”. Whether it was a pat down or a more intrusive search is not clear. Chaidez’ license was taken away. Both cars were searched, and the keys were taken from the ignition. (The road was blocked in both directions anyway.) The police found nothing in any of these searches.
All three were detained for the 10 or 15 minutes it took to search the house. When the drugs were found, all three were arrested. Chaidez was asked to consent to a search of his apartment. After he expressed concern that his family would be frightened by a search, Agent Rodriguez agreed to go to the apartment himself to calm the family. Chaidez then consented; the search turned up more drugs.
The district court held that Terry v. Ohio, 392 U.S. 1, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1966), supports the stops. We review the district court’s findings of fact deferentially. The standard of review for the conclusion that the seizures were reasonable, either because probable cause existed or for some other reason, is in transition in this circuit. Several cases hold that review is de novo. See United States v. Ingrao, 897 F.2d 860, 862 (7th Cir.1990); United States v. Jaramillo, 891 F.2d 620, 626 (7th Cir.1989); United States v. Sophie, 900 F.2d 1064, 1072 (7th Cir.1990). Recently doubts have been expressed, United States v. Malin, 908 F.2d 163, 169-70 (7th Cir.1990) (concurring opinion). Ordinarily application of law to fact is reviewed for abuse of discretion, and the fact-specific determination that behavior was “suspicious enough” to permit the intrusion is little different in quality from, say, a fact-finder’s conclusion of negligence. The Government does not ask for deferential review in this case, so we conduct our own analysis without taking sides on the proper approach.
II
A
The Fourth Amendment provides that searches and seizures shall not be “unreasonable”. The Supreme Court often treats a search without probable cause as “unreasonable”, drawing on the requirement in the second clause of the Fourth Amendment that no warrant may be issued without probable cause. But this starting point is riddled with exceptions. For a stop and search of the person, Terry requires only “reasonable suspicion”. Administrative [1197]*1197searches, Camara v. Municipal Court, 387 U.S. 523, 87 S.Ct. 1727, 18 L.Ed.2d 930 (1967), inventory searches, South Dakota v. Opperman, 428 U.S. 364, 369-76, 96 S.Ct. 3092, 3097-01, 49 L.Ed.2d 1000 (1976), school searches, New Jersey v. T.L.O., 469 U.S. 325, 341, 105 S.Ct. 733, 742-43, 83 L.Ed.2d 720 (1985), border searches, United States v. Montoya de Hernandez, 473 U.S. 531, 541, 105 S.Ct. 3304, 3310-11, 87 L.Ed.2d 381 (1985), drugtesting programs, National Treasury Employees Union v. Von Raab, 489 U.S. 656, 109 S.Ct. 1384, 103 L.Ed.2d 685 (1989), and searches incident to arrest, Chimel v. California, 395 U.S. 752, 89 S.Ct. 2034, 23 L.Ed.2d 685 (1969), for various reasons all require less than probable cause; in each case the Court emphasized that the fundamental question is reasonableness. If some factor makes it reasonable for the police to act on suspicion short of probable cause, then the Fourth Amendment does not ignore that factor and insist on probable cause. As the Court in Camara noted, “reasonableness is still the ultimate standard.” 387 U.S. at 539, 87 S.Ct. at 1736.
The extent of the intrusion is one such factor. It is “common sense that if the Fourth Amendment is intended to strike a balance between the interest of the individual in being left alone by the police and the interest of the community in being free from the menace of crime, the less the interest of the individual is impaired the less the interest of the community need be impaired to justify the restraint.” United States v. Serna-Barreto, 842 F.2d 965, 966 (7th Cir.1988). Consideration of the extent of intrusion abounds in modern Fourth Amendment doctrine. Stops that do not entail detention need not be justified by any suspicion. E.g., Michigan v. Chesternut, 486 U.S. 567, 108 S.Ct. 1975, 100 L.Ed.2d 565 (1988). Searches incident to arrest may be justified by the reduced marginal intrusion of searching a defendant already in custody. See United States v. Robinson, 414 U.S. 218, 237-38, 94 S.Ct. 467, 477-78, 38 L.Ed.2d 427 (1973) (Powell, J., concurring); Wayne R. LaFave & Jerold H. Israel, Criminal Procedure § 3.5 at 145 (1985). The Court based Terry itself on the fact that a protective search is a “brief, though far from inconsiderable, intrusion upon the sanctity of the person”. Terry, 392 U.S. at 26, 88 S.Ct. at 1182. Recently the Court upheld automobile checkpoints where the police made stops without any individual suspicion, in part because of the minimal “intrusion resulting from the brief stop at the sobriety checkpoint”. Michigan v. Sitz, — U.S.-, 110 S.Ct. 2481, 2487, 110 L.Ed.2d 412 (1990); see also United States v. Martinez-Fuerte, 428 U.S. 543, 558, 96 S.Ct. 3074, 3083, 49 L.Ed.2d 1116 (1976); United States v. Brignoni-Ponce, 422 U.S. 873, 879-80, 95 S.Ct. 2574, 2579-80, 45 L.Ed.2d 607 (1975). Once the intrusion amounts to a full custodial arrest, however, the police need probable cause. Dunaway v. New York, 442 U.S. 200, 99 S.Ct. 2248, 60 L.Ed.2d 824 (1979). The scale extends in both directions. If an intrusion is greater than a traditional arrest, probable cause is not enough. Winston v. Lee, 470 U.S. 753, 105 S.Ct. 1611, 84 L.Ed.2d 662 (1985) (surgery).
These cases describe a continuum in which the necessary degree of confidence increases with the degree of intrusion. A “stop” without limiting the suspect’s freedom requires no suspicion; a brief detention calls for reasonable suspicion; an arrest requires probable cause; invasive techniques such as surgery require more. What if the intrusion lies somewhere between Terry and arrest, neither a “brief, investigatory” stop nor a traditional arrest, where the defendant is handcuffed, trundled into a paddy wagon, carted • to the station, fingerprinted, and held in a 12' x 8' cell? One answer would be to deny that there is a “between” — to insist that all encounters must be either Terry stops or arrests. Yet circumstances defy such simple categorization, and if a line must nonetheless be drawn it will be arbitrary, with nearly identical cases on opposite sides. Trying to force a continuous world into two categories is not only impossible but also unnecessary when the text of the Constitution calls for inquiry into “reasonableness”. See Illinois v. Gates, 462 U.S. 213, 103 S.Ct. 2317, 76 L.Ed.2d 527 (1983) (probable [1198]*1198cause determination is based on “totality of circumstances”); United States v. Cortez, 449 U.S. 411, 417, 101 S.Ct. 690, 694-95, 66 L.Ed.2d 621 (1981) (same for reasonable suspicion determination); United States v. Sharpe, 470 U.S. 675, 682-83, 105 S.Ct. 1568, 1573-74, 84 L.Ed.2d 605 (1985) (scope of stop is question of reasonableness). Why abandon the search for reasonableness when the intrusion falls between arrest and stop?
Pigeonholing is no boon for defendants: it has put considerable pressure on the limits of the Terry doctrine. Both the permissible reasons for a stop and search and the permissible scope of the intrusion have expanded beyond their original contours, in order to permit reasonable police action when probable cause is arguably lacking. See, e.g., Sharpe (20 minute detention); Hayes v. Florida, 470 U.S. 811, 817, 105 S.Ct. 1643, 1647, 84 L.Ed.2d 705 (1985) (fingerprinting at the site of arrest); Serna-Barreto (detaining at gun point); United States v. Glenna, 878 F.2d 967 (7th Cir.1989) (handcuffing the suspect); United States v. Taylor, 716 F.2d 701, 709 (9th Cir.1983) (same); George E. Dix, Nonarrest Investigatory Detentions in Search and Seizure Law, 1985 Duke L.J. 849 (1985) (noting that, unless the defendant is taken to the stationhouse, the limits on nonarrest detention are unclear after Sharpe and Hayes). Increasing the threshold probability when the intrusiveness increases ensures that privacy interests will be protected (the “reasonable suspicion” threshold from Terry is low) without hampering reasonable investigative techniques by the police. Another circuit has suggested this approach, United States v. Quinn, 815 F.2d 153, 158 (1st Cir.1987), and we now adopt it. Stops too intrusive to be justified by suspicion under Terry, but short of custodial arrest, are reasonable when the degree of suspicion is adequate in light of the degree and duration of restraint.
Considering the extent of intrusion makes the calculus by the police marginally more complicated. In some tension with the trend toward a broadly-conceived inquiry into reasonableness is the desire to create rules easily implemented by the police. See United States v. Robinson, 414 U.S. 218, 94 S.Ct. 467, 38 L.Ed.2d 427 (1973); New York v. Belton, 453 U.S. 454, 101 S.Ct. 2860, 69 L.Ed.2d 768 (1981); Michigan v. Summers, 452 U.S. 692,101 S.Ct. 2587, 69 L.Ed.2d 340 (1981). But “probable cause” and “reasonable suspicion” are themselves standards rather than rules, so the existence of a middle ground does not blur a rule that is now sharp. As the Court said in Sharpe, “[m]uch as a ‘bright line’ rule would be desirable, in evaluating whether an investigative detention is unreasonable, common sense and ordinary human experience must govern over rigid criteria.” 470 U.S. at 685, 105 S.Ct. at 1575. The principal Supreme Court cases holding that an intrusion is just too much to be justified by less than probable cause almost invariably involve the trappings of a traditional arrest. See Dunaway (suspect taken to stationhouse); Hayes (same); Florida v. Royer, 460 U.S. 491, 103 S.Ct. 1319, 75 L.Ed.2d 229 (1983) (suspect taken to interrogation room). The only exception is United States v. Place, 462 U.S. 696, 103 S.Ct. 2637, 77 L.Ed.2d 110 (1983), involving a 90-minute detention of luggage to arrange for a dog-sniff. But Place emphasized that the length of the detention of the luggage was unnecessary, id. at 709, 103 S.Ct. at 2645-46, and hence unreasonable, a conclusion consistent with our approach.
B
The detention in this case was more intrusive than an ordinary Terry stop. All the agents had their guns drawn, which places it at the “outer edge” of investigatory stops, see Serna-Barreto, 842 F.2d at 968; United States v. Ocampo, 890 F.2d 1363, 1369 (7th Cir.1989); the cars were searched; the agents took the keys and prevented movement; Silva was given Miranda warnings and taken to a police van to sign a consent-to-search form; and the purpose of the questioning seemed as much to be to obtain consent to search the home as to inquire into possible wrongdoing. Still, it was not an arrest (until heroin was found, at which point all agree an arrest [1199]*1199was permissible): Silva was detained only briefly, asked a few questions, and apparently not physically restrained.
What quantum of probability is necessary to justify the seizure? For a typical investigatory stop, “th[e] level of suspicion is considerably less than proof of wrongdoing by a preponderance of the evidence.” United States v. Sokolow, 490 U.S. 1, 109 S.Ct. 1581, 1585, 104 L.Ed.2d 1 (1989). For a full arrest, Gates speaks of a “fair probability”, 462 U.S. at 238, 103 S.Ct. at 2332, but does not define it. Some older Supreme Court cases seem to imply a “more-likely-than-not” standard for arrests, see Mallory v. United States, 354 U.S. 449, 77 S.Ct. 1356, 1 L.Ed.2d 1479 (1957); Johnson v. United States, 333 U.S. 10, 68 S.Ct. 367, 92 L.Ed. 436 (1948); but see Gerstein v. Pugh, 420 U.S. 103, 121, 95 S.Ct. 854, 867, 43 L.Ed.2d 54 (1975) (probable cause determination “does not require the fine resolution of conflicting evidence that a reasonable-doubt or even a preponderance standard demands”). Courts have on occasion explicitly rejected the preponderance standard. See United States v. Cruz, 834 F.2d 47, 50 (2d Cir.1987); Browne v. State, 24 Wis.2d 491, 129 N.W.2d 175, 180 (1964). Sound quantification of levels of probability, hard enough to come by after a trial, is not a plausible demand of police conducting an investigation. “Probable cause” is a flexible idea, responding to (among other things) the gravity of the offense, and hence the need for thorough investigation. Llaguno v. Mingey, 763 F.2d 1560, 1566 (7th Cir.1985) (in banc). (This is further support for our sliding-scale approach to the relation between the level of suspicion and the degree of intrusion.)
Had the police stopped and detained only Chaidez, there could be no doubt that their behavior was reasonable. Chaidez’ morning consisted of evasive and circuitous driving, brief and furtive meetings, scurrying in and out of restaurants known to be frequented by drug dealers, always carrying a plastic bag used in the drug trade to transport drugs. The police observed all this, knowing that a reliable informant had claimed Chaidez was a major heroin dealer, and that Chaidez’ name had come up in other drug investigations. We have permitted similar intrusions on less evidence. See, e.g., United States v. Sophie, 900 F.2d 1064, 1072-73 (7th Cir.1990); Ocampo, 890 F.2d at 1368-70. See also United States v. Williams, 876 F.2d 1521 (11th Cir.1989) (evasive driving); United States v. Espinosa, 827 F.2d 604 (9th Cir.1987) (surreptitious behavior); Sibron v. New York, 392 U.S. 40, 66, 88 S.Ct. 1889, 1904, 20 L.Ed.2d 917 (“deliberately furtive actions and flight at the approach of ... law officers are strong indicia of mens rea”). Something was afoot; perhaps it’s best to ask the probability question the other way ‘round— what was the chance that Chaidez was engaged in behavior unrelated to drugs? We think it was trivially small. Chaidez offers no plausible explanation of his behavior, and we can’t think of an innocent one.
The key is Silva, however, for she consented to the search of the house. Was it reasonable to detain her along with Chai-dez? Chavira’s and Silva’s association with him (two rendezvous plus his visit to their house) justifies stopping them as well. If the police see a dealer sell drugs on the street to five people in a row, surely when the sixth comes up, any reasonable threshold of probability is surpassed as to that sixth person. Still, there is some resistance to drawing inferences from “mere association” with suspected criminals. Both
[1200]*1200United States v. Di Re, 332 U.S. 581, 68 S.Ct. 222, 92 L.Ed. 210 (1948), and Ybarra v. Illinois, 444 U.S. 85, 100 S.Ct. 338, 62 L.Ed.2d 238 (1979), imply that probable cause must be “individualized” in the sense that association with criminals or criminal places is not of itself enough. Some courts have held that entry into a known drug house does not by itself justify a detention, even if it is likely that the person is entering to consummate a drug transaction. Compare United States v. Clay, 640 F.2d 157 (8th Cir.1981), with United States v. Patterson, 885 F.2d 483 (8th Cir.1989). This discounting of probabilistic evidence, however, is of questionable validity after Gates and Sokolow. In Sokolow, the Ninth Circuit divided evidence indicating drug trafficking into two categories, facts describing “ongoing criminal activity”, and “personal characteristics” of drug couriers, and insisted that at least one fact from the first category be shown to achieve “reasonable suspicion”. The Supreme Court responded, 109 S.Ct. at 1586:
The rule enunciated by the Court of Appeals, in which evidence ... is divided into evidence of ‘ongoing criminal behavior,’ on the one hand, and ‘probabilistic’ evidence, on the other, is not in keeping with ... our decisions. It also seems to us to draw a sharp line between types of evidence, the probative value of which varies only in degree.
In order to distinguish “mere association” from other evidence of criminal activity, more is needed than the conclusory assertion that “person-specific” evidence is preferable to “probabilistic” evidence. No facts are “inherently” probative; apparently innocent events may add up to strong suspicion, and apparently damning facts may be innocent. All inferential processes are probabilistic. The likelihood that a gun-shaped bulge in a jacket pocket means a gun is just that — a likelihood, not qualitatively different from the likelihood that standing beside a drug dealer connects one to drugs. Just ask the person whose sunglasses case produced the bulge. Even a confession signed in blood just increases the probability of guilt; nothing in the legal process is certain. Acknowledging the statistical nature of inferential processes may well make them more accurate. See Branion v. Gramly, 855 F.2d 1256 (7th Cir.1988); Daniel R. Shaviro, Statistical-Probability Evidence and the Appearance of Justice, 103 Harv.L.Rev. 530 (1989). Perhaps drawing a line between a suspect’s status (e.g., where he lives) and his actions makes some sense for reasons independent of statistics, but frequenting drug houses or associating with drug dealers falls into the latter category by any definition. See United States v. Rodriguez, 869 F.2d 479, 483 (9th Cir.1989). The critical question should be how reliable is the inference drawn from a given fact. The point of Ybarra is that presence in a bar where contraband is likely to be does not amount to probable cause; presence is not irrelevant.
Most but not all of the evidence that Silva was involved in criminal activity piggybacks on evidence of Chaidez’ complicity. The statement in Ybarra, 444 U.S. at 91, 100 S.Ct. at 342, that “a person’s mere propinquity to others independently suspected of criminal activity does not, without more, give rise to probable cause to search that person” is inapplicable for two reasons: here we have more, and we don’t need probable cause. Silva and Chavira met Chaidez and drove in tandem with him to a gas station. They took a different route from Chaidez but met him on Lake-Cook Road. It’s common to tell a friend you’ll meet him at your house and to drive separately. But the Pontiac was observed in front of Chaidez’ Cadillac before reaching the house. Why separate and meet up again on the road other than as a counter-surveillance tactic? When the three entered the house, Chaidez was still carrying the plastic bag; when he emerged he no longer had the bag in his hand. Finally, right before the stop, the police thought they had been “made” and that both cars were trying to flee. Flight is relevant to whether a detention was reasonable. Sibron, 392 U.S. at 66-67, 88 S.Ct. at 1904-05; United States v. Clark, 743 F.2d 1255, 1260 (8th Cir.1984).
[1201]*1201Because the detention fell short of an arrest, the principal facts (Silva’s connection to Chaidez on a day when he was probably conducting his drug business, together with her strange rendezvous with him on the road and the perceived sudden flight) justified the seizure. The police did what courts tell them to do: they conducted an extensive surveillance instead of just acting on a tip, they waited until their observations indicated illegal activity, they made the stop only because it was necessary — the bad guys were getting away. These agents behaved reasonably.
Ill
A. Two other Fourth Amendment challenges remain. Chaidez contends that the search of his apartment, to which he consented after the search of the stash house on Weiland Road turned up drugs, was the fruit of his illegal detention. Because Chaidez’ detention was reasonable, this issue drops out. Chavira takes a different tack. He maintains that Silva did not have authority to consent to the search of the house (his house). (Silva parrots this argument, but it can’t help her; whether or not she had authority to allow an intrusion into her stepfather’s privacy, she certainly could surrender her own privacy interest, which she did by consenting. See United States v. Fuesting, 845 F.2d 664, 671 (7th Cir.1988).)
Chavira emphasizes that nothing Silva said to Agent Rodriguez indicated that she had authority to consent to the search of the house. Police may search pursuant to a consent if they reasonably believe that the person consenting has authority to consent — if the person has apparent authority. Illinois v. Rodriguez, — U.S.-, 110 S.Ct. 2793, 111 L.Ed.2d 148 (1990); United States v. Rodriguez, 888 F.2d 519, 523 (7th Cir.1989). Silva answered “no” when asked whether she lived in the house and said that she was there only to do laundry. Chavira insists that it cannot be reasonable to infer authority from such responses, and we agree. The district court found that Silva had authority “based on her representation [to Agent Rodriguez] that she had rented the property for her father.” But this cuts the other way: it should have put Agent Rodriguez on notice that Chavira and not Silva resided in the house. Renting a place “for” someone else may create a sub-lease. A landlord does not have authority to permit a search of his tenant’s leasehold, Chapman v. United States, 365 U.S. 610, 81 S.Ct. 776, 5 L.Ed.2d 828 (1961), and the same holds for a tenant and his sub-tenant. Use of and access to the property are the touchstones of authority, United States v. Matlock, 415 U.S. 164, 171 n. 7, 94 S.Ct. 988, 993 n. 7, 39 L.Ed.2d 242 (1974), and, for all Agent Rodriguez knew, Silva had neither.
Apparent authority is sufficient but not necessary; if Silva had actual authority, that is enough. We know of no case where the defendant successfully used lack of apparent authority as a defense even though actual authority was present. Cf. Feguer v. United States, 302 F.2d 214, 248-50 (8th Cir.1962) (characterizing such a position as “intriguing” but rejecting it because no privacy interest of defendant was invaded). An argument does exist for looking only to apparent authority and ignoring actual authority: the purpose of the exclusionary rule is to deter unreasonable police conduct, so if the police unreasonably, but correctly, believe the third party is authorized to consent, suppression might serve the purpose of deterring future unreasonable police conduct in the majority of cases where the police turn out to be incorrect. Probable cause similarly is viewed ex ante, so that a search does not become “reasonable” because of what it turns up. All questions under the Fourth Amendment are resolved objectively. See Horton v. California, — U.S. -, 110 S.Ct. 2301, 2308-09, 110 L.Ed.2d 112 (1990).
Nonetheless, the argument is unpersuasive for two reasons. First, the Supreme Court said in Rodriguez that “[i]f [apparent authority does not exist], then warrant-less entry without further inquiry is unlawful unless actual authority exists.” 110 S.Ct. at 2801 (emphasis added). Second, this statement is not an accident but is consistent with the theory of consent [1202]*1202searches. Consents define the extent of the privacy interest a person seeks to assert. If someone who actually has a privacy interest surrenders that interest voluntarily, he may not later claim that the police should have refused the offer. Having accepted a lower degree of privacy, the suspect has surrendered the foundation for an argument under the Fourth Amendment that the police invaded an area he sought to preserve. The Fourth Amendment applies in the first place only to things persons seek to hold in confidence. So actual authority over the premises is enough to permit a search. What of the interests of the co-tenant? The underpinning of third-party consent is assumption of risk, Frazier v. Cupp, 394 U.S. 731, 740, 89 S.Ct. 1420, 1425, 22 L.Ed.2d 684 (1969); Matlock, 415 U.S. at 171 n. 7, 94 S.Ct. at 993 n. 7 (1974). One who shares a house or room or auto with another understands that the partner may invite strangers — that his privacy is not absolute, but contingent in large measure on the decisions of another. Decisions of either person define the extent of the privacy involved, a principle that does not depend on whether the stranger welcomed into the house turns out to be an agent or another drug dealer.
Silva paid the electricity and telephone bills for the house, both of which were in her name. The police found women’s clothing in the bedroom, and although the clothes were not directly linked to Silva, it is a natural inference that they were hers. And the owner of the property testified at trial that he rented the house to both Silva and Chavira. Silva is a lessee, and ordinarily a lessee has every right to permit a search of the leased premises. Matlock; United States v. Main, 598 F.2d 1086, 1092 (7th Cir.1979). Silva had the power to admit strangers to the house; she admitted the police; because the police were there at the sufferance of someone entitled to admit them, the evidence they obtained is admissible against everyone. Chavira responds to these facts by seizing on a key footnote in Matlock, 415 U.S. at 171 n. 7, 94 S.Ct. at 993 n. 7: “The authority which justifies the third-party consent does not rest upon the law of property, with its attendant historical and legal refinements, ... but rests rather on mutual use of the property by persons generally having joint access or control for most purposes.” Chavira, repeating his arguments against apparent authority, contends that while Silva may have been a lessee, she did not live in the house, and went there only to do laundry. The problem is that while Agent Rodriguez might have believed this, the jury did not believe it, and thus, so far as legal processes can determine truth, her story is not true.
The meanings of “mutual use” and “joint access” are far from clear, compare United States v. Heisman, 503 F.2d 1284 (8th Cir.1974) with State v. Campbell, 326 N.W.2d 350 (Iowa 1982), but this case is not close. The utility bills in Silva’s name, the lease in her name, and women’s clothing in the house all belie Chavira’s claim that Silva wasn’t using the house. There is no evidence that Chavira had more access to the house or used it more often than she; this was a stash house, with the electricity and plumbing turned off, apparently used for storing drugs and drug records. Neither the judge nor the jury believed Silva's tale about the limited purpose for which she visited the house. If she had a different purpose in going there, she had actual authority to consent to a search.
B. Chavira contends that his conviction must be reversed because drug records bearing his fingerprints were erroneously admitted against him. The drug records were ten years old and pertained to transactions unrelated to those charged in the indictment. Since the defendant denied knowingly possessing the drugs, the records are admissible under Fed.R.Evid. 404(b) to prove such knowledge. Chavira complains, however, that at trial the Government portrayed the records as pertaining to contemporaneous transactions; he cries “unfair prejudice” and essentially invokes Rule 403. He did not make such an objection at trial, however, and so has waived the argument. United States v. Carroll, 871 F.2d 689, 691 (7th Cir.1989) (specific ground for objection must be identified at trial). The objection he did make [1203]*1203was a vague assertion that the drug records were irrelevant, not that the Government was mischaracterizing them. (In fact, Judge Alesia was puzzled about the nature of the objection, repeatedly asking Chavira’s counsel to be more specific. We are puzzled also by his claim in this court: Agent Dailey testified at trial that the records contained “old balances”, hardly a mischaracterization.)
C. Finally, Silva contends that there was insufficient evidence to support her conviction. Although the government introduced no direct evidence linking her to the conspiracy, Silva did support the operation by paying the bills for the stash house. She was seen going in and out of the house with Chaidez and Chavira moments before the police discovered contraband in plain view. Her explanation for her presence was incredible. The inferences, when viewed in the light most favorable to the prosecution, are strong enough to survive our limited review: a rational jury could be convinced beyond a reasonable doubt of Silva’s involvement. United States v. Draiman, 784 F.2d 248, 251 (7th Cir.1986).
Defendants’ remaining arguments are insubstantial. Because the police behaved reasonably in gathering evidence, and because no reversible error occurred at trial, the judgments are
Affirmed.
The district court found that “the agents conducted a protective search of the defendants", apparently including Silva. The only testimony at the suppression hearing on this point came from Officer Rodriguez, who did not mention that Silva was searched. The parties do not press this discrepancy, however (in fact, the government admits in its brief that the "defendants" were patted down), so neither do we. In line with the district court’s findings, we assume that the police conducted a brief protective search of Silva but otherwise did not restrain her physically.