Dooley, J.
¶ 1. What are the chances? This case poses the question of when that familiar rhetorical question is also a valid legal inference. Defendant was convicted of two counts of arson after four unexplained fires occurred at his rented house within an eight-week period. The trial court itself relied upon — and instructed the jury that it might rely upon — the fact that rare events like fires do not normally happen together in order to infer that defendant willfully started the fires. Although we agree that there are potentially valid inferences to be drawn from the multiplicity of fires, we cannot approve the trial court’s instruction. However, we affirm because the error does not rise to the level of plain error.
¶ 2. In late 2008 and early 2009, defendant and his family rented a single-family home with a detached two-car garage in Colchester. Between November 28, 2008, and January 18, 2009, the Malletts Bay Fire Department, as well as the Colchester Police Department, responded to four separate fires at that residence. These fires represented four of the ten fires to which [626]*626the Malletts Bay Fire Department responded in all of 2008 and 2009.
¶ 3. At around 1:00 p.m. on November 28, 2008, flames engulfed the detached garage at defendant’s residence. By the time the fire department arrived, the fire had broken through the roof and was blowing out of the top of the structure. The fire destroyed the garage, but the fire department was able to contain it and keep it from spreading further. A neighbor testified that before the firefighters arrived defendant was unsuccessfully attempting to put the fire out with a small kitchen fire extinguisher. Responding officials described defendant as pacing, very upset, and near tears. He stated that he had no idea how the fire had started — one of the officials testified that he explained: “I was spending the day watching Kiss videos and getting loaded.” Sure enough, firefighters made note of Kiss videos playing on the television, and a preliminary breath test of defendant revealed a blood-alcohol content (BAC) of .288. An area resident testified that, as he was watching the fire, defendant — visibly intoxicated —' approached him and said, “Good fire, huh?” and “So I did a good job, then, right?” Ultimately, the cause of the fire was considered “undetermined,” which was explained at trial to mean that authorities were “unable to find a definitive reason why that fire started,” because of the fire’s extensive destruction of any physical clues. Authorities investigated the possibility that the fire might have been caused by burning leaves in the backyard, by faulty electrical wiring in a garage light, or by an older refrigerator located in the garage, but no physical evidence was present to confirm a particular theory.
¶ 4. Approximately one month later, at roughly 6 p.m. on Christmas Day, the fire department responded to a fire in a clothes dryer in the basement of defendant’s residence. The dryer was an older model and was heavily loaded at the time of the fire. Defendant’s wife testified that she had started a large load of items in the dryer, including kitchen towels that defendant regularly used to clean out the deep fryer. At trial, the fire chief testified that dryers can and do catch fire, but that inspection of this dryer did not reveal any of the common malfunctions. Vegetable oils were found on some of the clothes, but no accelerants or other materials were present to suggest that the fire was deliberately set. Investigating officers said defendant told them that he was the only one home at the time, that he was doing laundry, and that he noticed smoke while he was watching [627]*627television. The fire chief noted signs of intoxication in defendant’s behavior, and defendant submitted to a preliminary breath test, which produced a BAC of .210. Once again, the cause of the fire was considered undetermined.
¶ 5. A little over one week later, at approximately 4 a.m. on January 5, 2009, a neighbor reported a fire on the enclosed porch of defendant’s residence. A Colchester police officer arrived at the scene before the fire department, and he was able to contain the fire with the fire extinguisher from his cruiser until the fire department arrived and fully extinguished it. The fire occurred in some odds and ends of pressure-treated lumber that were stacked on the porch, which was undergoing some renovation. At trial, the police officer testified that defendant’s family members were the only people in the house when he arrived, and a firefighter circled the house and found no footprints in the snow. The responding officer testified that defendant stated that he had smoked a cigarette on the porch around midnight and had seen nothing wrong. Empties showed that defendant had consumed nine 24-ounce cans of beer that night. Because of the prior fires, the state police fire investigator was called to the scene, and he ultimately concluded that the fire was “intentionally set.” This conclusion was not based on physical evidence of the cause of the fire but rather on the fact that “there were no other viable ignition sources in the area to explain how the fire started.” The investigator did consider the possibility that a cigarette might have caused the fire, having found cigarette butts on the porch and knowing that defendant and his wife were smokers. In fact, defendant’s wife testified that defendant smokes roughly a pack a day, that she smokes five to ten cigarettes a day, and that both of them generally used the porch as the area for smoking. Nevertheless, the investigator ruled out that cause based on defendant’s strong assertion that he had not dropped a cigarette on the porch that day. At trial, a defense expert disagreed with the decision not to classify the cause of the fire as “undetermined,” noting that there were easily ignitable materials and that “we have a man who drinks a lot, smokes all over the place.”
¶ 6. Finally, just after midnight in the early morning of January 18, 2009, the fire department responded to another fire at defendant’s residence. By the time they arrived, the firefighters found the south side of the house fully engulfed in flames. The fire department was unable to prevent the fire from completely [628]*628destroying the residence and all of its contents. Defendant, who had been home alone, had apparently been forced to break a window to escape the fire, and he had cut his hands on the glass, requiring medical attention. The only footprints found around the house were those of defendant evacuating. A police officer at the scene spoke to defendant and noticed strong signs of intoxication. Defendant told the officer that he had consumed “ten or so beers,” at which point the officer took defendant into protective custody, where a preliminary breath test indicated a BAC of. .155. At trial, a state police fire investigator testified that the fire started on the porch, just outside the kitchen. The investigator was able to rule out several possible causes for the fire, including electrical faults and environmental ignition, and he was able to determine that “it came down to human involvement somehow.” He explained that “something had to be either placed there or started there to make this fire go,” but he could not say whether it was intentional or not. Specifically, he noted that it was possible for a cigarette to have started the fire. The defense expert stated that the cause of the fire should be classified as “undetermined,” largely because it was not possible to rule out careless smoking as the cause of the fire.
¶ 7. The State charged defendant with four counts of first-degree arson, alleging that he “willfully and maliciously . . . caused to be burned [a] dwelling house ... or other outhouse that is parcel thereof ... in violation of 13 V.S.A. § 502.” Initially, the State also charged him with four counts of setting a fire to defraud an insurer in violation of 13 V.S.A. § 506. However, although the family had renters’ insurance that paid out on the first and fourth fires, defendant’s wife insisted that defendant was unaware of the policy because she handled matters of that nature. In response, prior to trial, the State dropped the four counts of setting a fire to defraud an insurer. Also prior to trial, the defense moved to dismiss all the charges for lack of a prima facie case, see V.R.Cr.P. 12(d), and alternatively to sever all the charges for separate trials. The court denied these motions as untimely filed.
¶ 8. At trial, the defense moved to dismiss all counts at the close of the State’s case. Defense counsel argued that, with regard to each count, the State had failed to provide sufficient evidence to prove that the fire was intentionally set and that intent could not be inferred from the other fires. In response, the State argued that the frequency of the fires — four in eight weeks — showed [629]*629that the fires were not accidental. After pressing the State, to no avail, for some legal authority to support its desired inference, the judge introduced sua sponte the so-called “doctrine of chances,” which the judge described as holding that “when you have an unusual number of events that don’t normally occur accidentally, the.jury can infer that they occurred intentionally, and that’s exactly what we have here.” The judge referred the attorneys to People v. Mardlin, 790 N.W.2d 607 (Mich. 2010), State v. Allen, 725 P.2d 331 (Or. 1986), and State v. Lizotte, 109 Vt. 378, 197 A. 396 (1938). On the basis of the doctrine of chances, the trial court denied the motion to dismiss.
¶ 9. The trial court also drafted a jury instruction based on the doctrine of chances. In full, the instruction read:
In this case, there is evidence of four different fires at Mr. Vuley’s home. The defense claims that each of these fires was undetermined.1 It is your job to examine each incident separately to determine whether the State has proved beyond a reasonable doubt each of the elements of arson or attempted arson for each of the fires.
You may consider the meaning of the evidence that four different fires occurred at Mr. Vuley’s house. The weight to be given to evidence that four fires occurred is left to your common sense and judgment. You may, but are not required to, apply the doctrine of chances, as follows. Based on ordinary common sense and human experience you may conclude that it is unlikely that this number of similar accidents would occur on the same property in a short period of time. Considered in isolation, one fire may be easily explained as an accident. However, when all similar incidents are considered collectively or in the aggregate, the doctrine of chances may create an inference of human design. You may conclude that the recurrence of similar incidents incrementally reduces the possibility of accident. The improbability of a [630]*630coincidence may cause you to conclude that the fires were intentionally set.
From the evidence of four separate fires, you may (but are not required to) infer that the four fires were not the result of accidental or natural causes, but were the willful act of some person.
At the charge conference, defense counsel objected to this proposed instruction. Defense counsel began by pointing out that, in the three cases cited by the court the day before, the other incidents were not the subject of simultaneously pending charges. When asked why this should make a difference, defense counsel argued that the instruction “relieves the State of its burden . . . to prove each and every element of each and every crime.” Ultimately, the court cut defense counsel off, stating ‘Well, we agree to disagree and I’m going to include it.” The defense renewed the objection following the jury charge but did not elaborate.
¶ 10. After deliberation, the jury acquitted defendant on the first two counts, which related to the earlier incidents, but convicted on counts three and four, which related to the later incidents. Defendant appeals, arguing that the trial court erred by (1) not granting the motion to dismiss, (2) giving the jury instruction on the doctrine of chances, and (3) refusing to grant the motions to sever and dismiss for lack of a prima facie case. We first consider the denial of the motions and then proceed to the issues involving the doctrine of chances.
I.
¶ 11. The circumstances surrounding the motions to dismiss and to sever are as follows. The information to commence this case was filed on September 30, 2009. Status conferences were held in October and November of that year, and then on January 26, 2010. At the January conference, the parties filed a stipulated schedule, with depositions to be conducted by May 1 and motions to be filed by May 15. The schedule could not be met, and at a May 11 status conference, the court extended the deadlines, setting the motion deadline at August 13 and ordering that the trial commence in September, with no further continuances. At a July 20 status conference, the parties reported that depositions had not been completed and they would not be ready for trial in [631]*631September. The court continued the trial date to November without resetting any other deadlines. At a status conference in October, the court set a new trial date of December 6. Thereafter, on November 17, defendant filed a motion to dismiss for lack of a prima facie case or, if the prosecution was to proceed, to sever all counts. The court denied the motion as beyond the deadline for motions — August 13 — set at the May status conference and never modified. Thereafter, the court denied a motion for reconsideration.
¶ 12. Defendant argues on appeal that the motion deadline had been impliedly extended, and it was an abuse of discretion to deny the motion for reconsideration. The two motions defendant filed in November were pretrial motions. See V.R.Cr.P. 12(b)(5), (d); Reporter’s Notes, 1984 Amendment, V.R.Cr.P. 12(d); V.R.Cr.P. 14(b)(4). Thus they were subject to the filing deadline of Vermont Rule of Criminal Procedure 12(c), which is the date of the status conference following arraignment, or twenty-eight days after arraignment if there is no status conference, “unless extended by the court.” V.R.Cr.P. 12(c). Failure to file motions by the applicable deadline in Rule 12(c) is a waiver of the right to file the motion unless the court grants relief from the waiver “for cause shown.” V.R.Cr.P. 12(f); see State v. Dann, 167 Vt. 119, 136, 702 A.2d 105, 116 (1997).
¶ 13. Defendant argues that there was no waiver here because the trial judge who conducted the July status conference, a different judge from the one who imposed the time deadlines, impliedly removed the deadline for filing motions. Defendant relies upon an interchange in the status conference in which defense counsel stated that all depositions could not be done by July 31, and the judge responded that she did not care when the depositions were done. Defense counsel then responded, “[I]t drives everything out. It drives the motion — and everything.” The exchange is vague, and defense counsel does not explicitly state that the motion deadline cannot be met. The court did not extend the motion deadline, either orally or in writing. In fact, the court continued the trial date, but did not address the motion deadline. We conclude that the lawyers who were subject to a written court-imposed deadline bore the burden of having it explicitly continued or removed. The failure to do so here meant that pretrial motions could not be filed after the deadline.
[632]*632¶ 14. Defendant urges us to reach the same result by holding that the court should have granted the motion for reconsideration. The motion for reconsideration was premised on the representation that the parties “asked the court for additional time ... to file motions” and the status conference judge “granted the joint request.” The court denied the motion for reconsideration indicating that there was no entry that the status conference judge had granted such a request and the motions were not timely filed. We note that there is no support in the record for the representation that the parties requested that the motion deadline be extended and none for the further representation that the court granted the request.
¶ 15. As defendant acknowledges, the trial judge’s ruling was discretionary, and we will overturn it only on a finding of abuse of that discretion. Dann, 167 Vt. at 137, 702 A.2d at 116. As we noted in Dann, the point of pretrial motions is that the court should confront them before the court addresses the merits of the case and the parties have commenced preparing for trial. Id. at 136, 702 A.2d at 116. We also found important in Dann that the only reason given for why the motion could not have been filed earlier was weak. Id. at 137, 702 A.2d at 116. Here, defendant gave no reason for why the motions could not have been filed earlier, and we can see no reason in the record. Defendant did not appeal to the discretion of the trial judge, but argued the point that there had been no waiver, the argument we rejected above. We find no abuse of discretion.
¶ 16. The remaining two claims on appeal directly involve the doctrine of chances. Before reviewing them, we first address the doctrine of chances. The doctrine of chances has evolved as a theory about the grounds for the admission of evidence. Evidence of other crimes or acts is not admissible to demonstrate a person’s character for the purposes of showing that a person acted in conformity therewith. V.R.E. 404(b). Against the backdrop of this prohibition, the doctrine of chances purports to offer an explanation of the logical relevance of other incidents that does not rely upon character-based reasoning. The classic articulation of the doctrine comes from Wigmore:
To prove Intent, as a generic notion of criminal volition or willfulness, including the various non-innocent mental states accompanying different criminal acts, there is [633]*633employed an entirely different process of thought. The argument here is purely from the point of view of the doctrine of chances, — the instinctive recognition of that logical process which eliminates the element of innocent intent by multiplying instances of the same result until it is perceived that this element cannot explain them all. . . . [T]he mind applies this rough and instinctive process of reasoning, namely, that an unusual and abnormal element might perhaps be present in one instance, but the oftener similar instances occur with similar results, the less likely is the abnormal element likely to be the true explanation of them. ... In short, similar results do not usually occur through abnormal causes; and the recurrence of a similar result (here in the shape of an unlawful act) tends (increasingly with each instance) to negative accident or inadvertence or self-defence or good faith or other innocent mental state, and tends to establish (provisionally, at least, though not certainly) the presence of the normal, ie. criminal, intent accompanying such an act ... .
1 J. Wigmore, Evidence § 302, at 611-12 (2d ed. 1923). The logic resonates with common sense. As one court succinctly explained, “The man who wins the lottery once is envied; the one who wins it twice is investigated.” United States v. York, 933 F.2d 1343, 1350 (7th Cir. 1991), overruled on other grounds by Wilson v. Williams, 182 F.3d 562 (7th Cir. 1999). An archetypal example of the doctrine is the English case of Rex v. Smith, 11 Cr. App. R. 229, 84 L.J.K.B. 2153 (1915), in which the defendant had three wives who all suspiciously drowned while taking a bath. The court held that the other two deaths were probative with regard to whether the charged death was an accident.
¶ 17. The doctrine of chances is regularly invoked in arson cases. In fact, Professor Imwinkelried illustrates the doctrine with a discussion of arson:
Suppose that the defendant is charged with arson. The defendant claims that the fire was accidental. The cases routinely permit the prosecutor to show other acts of arson by the defendant and even nonarson fires at premises owned by the defendant. In these cases, the courts invoke the doctrine of objective chances. The [634]*634courts reason that as the number of incidents increases, the objective probability of accident decreases. Simply stated, it is highly unlikely that a single person would be victimized by so many similar accidental fires in a short period of time. The coincidence defies common sense and is too peculiar.
1 E. Imwinkelried, Uncharged Misconduct Evidence § 4:1, at 4-6 to -9 (rev. ed. 2008). The cases upon which the trial court relied are examples of this reasoning. In Mardlin, for example, the Michigan Supreme Court held admissible in an arson case evidence of four previous home and vehicle fires — all of which involved insurance claims — associated with the defendant. 790 N.W.2d at 610-11. The present case poses the open question of whether Vermont law similarly approves the inference described by the doctrine of chances.2
¶ 18. Despite an illusory simplicity, the doctrine of chances faces a high hurdle: the need to demonstrate specific, criminal intent. Further, it must do so without engaging in propensity reasoning, which is the claim that because a person once engaged in certain criminal conduct, the person is more likely to do it again. A careful look at the doctrine of chances shows that it actually describes a number of different inferences, some of which [635]*635meet the challenge and avoid propensity reasoning and some of which do not. The conflation of these various modes of reasoning under the same name is the source of a great deal of confusion in the case law and the commentaries.
¶ 19. The first inference that is described as the doctrine of chances might be explained as follows: where independently improbable events occur together, it is likely that some common cause exists that explains them all. This is actually a two-step inference. First, one infers from the unlikelihood of multiple improbabilities coinciding that something abnormal must be present. Second, one derives from the posited abnormality an explanation of each individual incident. That is, one first infers that not all instances could be a matter of chance, and then infers that, if not all were a matter of chance, then each one is more likely to have been not a matter of chance. For example, if ten children in a home die from an unusual respiratory illness, one might infer that there is something in the house causing the illness. And this makes it more likely that any particular case of respiratory failure in the house is caused by this underlying source, as opposed to random misfortune. We might call this the probabilistic doctrine of chances.3 Lightning doesn’t strike twice — unless there’s a reason.
¶ 20. Defendant argues that the latter inference is not logically supported and based his motion to dismiss on this argument. He asserts that, even if the objective probability of all the incidents occurring together is very low, the probability of each individual incident is unaffected. Flipping heads ten times in a row is very [636]*636unlikely — less than one in a thousand — such that one might infer that there is something unfair about the coin or the coin flipper. But, as defendant correctly points out, this improbability does not establish the same certainty that any particular flip, say the second one, was unfair. Our certainty that the collection of events was not random will always be greater than our certainty that any individual event was not random. So far, we agree. It does not follow, however, that no evidentiary inference is possible as to individual flips. Although we cannot say that there was only a one-in-a-thousand chance that the second flip was fair, we can say that there is some evidence that it was not fair. Given that it is highly unlikely that all the flips were random, this makes it more unlikely that the second flip was random — not as highly unlikely, but more unlikely. This is because, insofar as we can infer the presence of some sort of irregularity, it makes it more likely that such an irregularity affected each individual instance. This is the second stage of the overall inference described above, and there is nothing improper about this reasoning. It does not relieve the State of its burden — it simply acknowledges that the presence of other associated events may suggest a defeasible hypothesis for explaining a particular, otherwise improbable event. After all, what are the chances?
¶ 21. Matters are somewhat more complicated, however, when this reasoning is applied more specifically to multiple human actions over time. Here, we reach the hurdle of determining intent. Explaining human action is what is involved in the second form of inference described as the doctrine of chances: where actions that might independently have been accidental occur together, that makes it more likely that the acts were not accidental.4 This reasoning is simply an application of the above-referenced probabilistic reasoning applied to human action, and it [637]*637has the same two-step form. First, one infers that all of the acts could not have been accidental — there must be something intentional about some of this person’s conduct — and then one infers that this intentionality likely explains each particular incident. For example, if a man shoots three different wives and asserts that each shooting was accidental, one might infer that not all of them could have been accidents and that, more likely, the man has intentionally killed all of them. As we already said, drawing this inference is logically supported. This is simply the probabilistic doctrine of chances applied to human behavior.
¶ 22. Nevertheless, we conclude that this second form of inference is generally impermissible. The barrier is that, although logically relevant, probabilistic reasoning about human behavior of the form necessary to the second step runs afoul of the prohibition against using propensity-based reasoning. Inferring from the implausibility of all occurrences being accidents that any particular occurrence was not an accident necessarily involves reasoning based on propensity.5 Consider the above example. From the series of shootings, one can infer that it is unlikely that the killings were all accidents. This is logically equivalent to the [638]*638conclusion that at least one of the killings was not an accident. But it would be an inference based on propensity to say that, because a man has intentionally killed a wife, he is therefore more likely to have intentionally killed this wife.6 For this reason, in In re S.G., we rejected an attempt — based on the same section in Wigmore — to introduce evidence of abuse of one child to show that an unexplained broken bone in another child was less likely to be an accident. 153 Vt. 466, 471, 571 A.2d 677, 680 (1990); see also State v. Lipka, 174 Vt. 377, 396 n.11, 817 A.2d 27, 43 n.11 (2002) (reiterating reluctance to adopt Wigmore rationale). This is not, of course, to deny that there is a logical chain of inference in such cases. The whole point of the propensity-based prohibition in Vermont Rule of Evidence 4047 is that it excludes evidence that is logically relevant. When one applies the probabilistic doctrine of chances to human behavior, one gets a logically valid inference, but one that is based on human propensity to act in conformity with other bad acts — precisely what the rule of evidence does not allow.8
[639]*639¶ 23. While the general prohibition against propensity-based evidence might, at first blush, suggest that the doctrine of chances cannot be used, there is a valid third and altogether different inference that goes under the header of the doctrine of chances: where someone has already committed an unusual act with bad consequences, it is more likely that subsequent acts of that type were not accidental.9 This inference is based on the understanding [640]*640that people learn from their mistakes. If someone has accidentally done something with terrible consequences, then that person becomes less likely to do that same thing accidentally a second time. The person is likely to have a heightened awareness of the consequences and therefore is less likely to be careless. For example, a man who accidentally shot his first wife will be less likely than an average person to accidentally shoot his second wife. As a result, one can infer that a subsequent shooting — especially one under distinctly similar circumstances — is less likely to be inadvertent and therefore more likely to be intentional. We might call this the psychological doctrine of chances. To see the difference, note that the probabilistic doctrine of chances would find it more likely that each improbable incident was not an accident, whereas the psychological doctrine of chances would apply to primarily explain subsequent incidents.10 Note that the psychological doctrine of chances provides no inference that the first incident was not an accident.
[641]*641¶ 24. There is nothing impermissible about the inference involved in the psychological doctrine of chances. Essentially, this reasoning treats the past incidents not as bad acts attributable to the defendants, but as events that would affect defendant’s knowledge and mental state. When introduced for this purpose, the previous fires serve essentially the same purpose as when prior bad acts are used to show intent or knowledge. See V.R.E. 404(b); In re S.G., 153 Vt. at 471, 571 A.2d at 680 (noting that, although impermissible to infer abuse, “the earlier incident would raise skepticism that the mother was unaware that S.G.’s symptoms were caused by a broken leg”).' In this form, the doctrine of chances is less about cosmic happenstance and more about the relevance of individual history to a person’s mental state. We have no difficulty concluding that the latter is an appropriate sort of evidence for a jury to consider.
II.
¶25. Against this backdrop, we consider the application of the doctrine of chances to this case. Courts are often confused by the difference between these three types of inference. This confusion is compounded where, as here, the factual situation introduces the possibility of all three forms of inference and there is no obvious motive for the alleged crimes. First, it is highly implausible that the four fires were simply unrelated freak accidents: an electrical problem, a dryer malfunction, combustion in some spare lumber, and so on. Something unusual was going on in defendant’s house. To the extent that defendant maintained that his household was just extremely unlucky, the doctrine of chances is a basis for rejecting this explanation. The doctrine provides a reason why the other fires are probative to the individual counts. This much, however, is compatible with the thought that the fires were nonetheless accidental — it does not get the State specific criminal intent.
¶ 26. One might use the doctrine of chances in the second way in order to suggest that even accidental ignition is implausible as an explanation for all of these fires. Defendant is surely not the only heavy-drinking smoker in Colchester, but his residence accounted in two months for forty percent of the structural fires over a two-year span. From this perspective, it is unlikely that they were all simply accidents — and therefore it is likely that some or all of them were not accidents. Although logically [642]*642relevant, this proposition cannot be used to infer that defendant is guilty of causing a specific individual fire. Such an inference would involve propensity reasoning: because defendant likely started some fire intentionally, he is an arsonist that started all fires intentionally. If the multiple fires were viewed as probative solely based on this reasoning, we must consider the result erroneous.
¶ 27. On the other hand, the psychological version of the doctrine might be used to shed doubt on any theory that the third or fourth fire was caused accidentally. Someone who has had two or three fires at his home in the. span of a month should be much less likely to toss around cigarettes cavalierly. While the chance of starting a house fire may generally be so remote that someone might fail to pay it adequate attention, it is hard to imagine that defendant, after experiencing multiple recent fires, could have been similarly unaware. In this way, the previous fires make it less plausible that this was mere carelessness. Thus, the earlier fires are probative of lack of accident. Of course, their probative value would have to be weighed by the jury against countervailing evidence. In particular, heavy drinking makes carelessness more likely. But this does not mean that the previous fires are not relevant to whether the ignition was intentional, only that their relevance must be weighed alongside other facts by the jury. In short, the other fires do provide some evidence regarding defendant’s mental state.
¶ 28. The jury heard the details of all four fires. The potential relevance of the multiple fires was brought out at various points. Over the objection of defense counsel, the State opened an entire line of questioning on redirect of the fire chief discussing the improbability of so many fires at one location, in which the fire chief explained, “One fire, completely believable. Two fires, -wicked bad luck. Three fires, no way. . . . [M]ost homes don’t burn once, let alone three times.” Another firefighter spoke about the peculiarity of three fires at one location, stating that “it was extremely unusual for there to be three fires at the same address.” Under the doctrine of chances, the jury did not have to be shielded from evidence of the multiple fires, and, in this case, the jury was not shielded.
III.
¶ 29. We turn now to the issues on appeal, which implicate the doctrine of chances.
[643]*643A.
¶ 30. The first issue is whether the trial court erred in denying defendant’s motion for a judgment of acquittal, relying upon the doctrine of chances. We conclude that the trial court did not err in denying the motion for a judgment of acquittal. We review de novo a motion for a judgment of acquittal. State v. Ellis, 2009 VT 74, ¶ 21, 186 Vt. 232, 979 A.2d 1023. The question is whether the State has produced evidence that could reasonably support a guilty verdict. See State v McBurney, 145 Vt. 201, 204, 484 A.2d 926, 928 (1984) (“[T]he sole issue raised by a motion for judgment of acquittal is whether the State’s evidence fairly and reasonably tends to show the defendants’ guilt, so that a jury would be justified in finding guilt beyond a reasonable doubt.”).
¶ 31. Reviewing the evidence in a light favorable to the State, we conclude that a jury could have concluded that defendant intentionally set the third and fourth fires.11 With regard to both fires, the state investigators essentially testified that they could rule out causes not involving human action, but that they could not completely rule out accidental causation. The inability to rule out accidental causation was the basis for the defense’s motion for acquittal.
¶ 32. We conclude that the State presented sufficient evidence to survive a motion for judgment of acquittal with respect to counts 3 and 4. There was evidence that the third and fourth fires were intentional — namely, the improbability of accident in light of the other fires. This inference, which could support a guilty verdict on these counts, is based on the psychological doctrine of chances, applicable given that defendant had experienced two recent fires at his house. One could reasonably expect him to be particularly vigilant against creating another fire. “A judgment of acquittal is proper only if the State has failed to put forth any evidence to substantiate a jury verdict.” State v. Turner, 2003 VT 73, ¶ 7, 175 Vt. 595, 830 A.2d 122 (mem.). We conclude that such was not the case here. The evidence of the other two fires may have been only circumstantial proof of intent and it may have been counterbalanced by the evidence of defendant’s drinking habits, but these are matters of weighing the [644]*644evidence appropriate for the jury. There was no error in permitting the jury to engage in this enterprise.
¶ 33. Defendant also contests the element of identity with respect to the third fire, arguing that the evidence was insufficient to prove that he was in fact the person who set this fire. In support of this assertion, defendant argued very cursorily in his motion for a judgment of acquittal that he was not home alone during the third fire as he was during the remaining fires.
¶ 34. We conclude that the State did not need to rely on the doctrine of chances or on an impermissible inference about defendant’s character to establish identity. There was sufficient evidence presented at trial from which the jury could permissibly infer that defendant was the person who started this fire. See State v. Kerr, 143 Vt. 597, 603, 470 A.2d 670, 673 (1983) (“[P]roof of facts includes reasonable inferences properly drawn therefrom.”). As stated above, defendant was physically present when all of the fires started, and he was home alone for three of the four fires. He made statements to a neighbor after the first fire, including “Good fire, huh?,” and “So I did a good job, then, right?” Defendant was unemployed and depressed. He was intoxicated when officials responded to each fire. During the third fire, there was evidence that defendant had consumed nine 24-ounce cans of beer. There was expert testimony that the third fire was intentionally set. There were no footprints in the snow around the house to support the notion that an outsider had set the fire. Defendant’s wife testified that she had had not been smoking in the area where the fire started. From this circumstantial evidence, the jury could reasonably infer that defendant set the fire in question.
¶ 35. While it is true that there were other family members at home during the third fire, “the State is not required to exclude every reasonable hypothesis of innocence in proving a case with circumstantial evidence.” State v. Warner, 151 Vt. 469, 472, 560 A.2d 385, 387 (1989). “So long as the jury by way of a process of rational inference could conclude beyond a reasonable doubt that defendant committed the acts for which he was charged, we will not disturb the jury’s verdict.” State v. Godfrey, 2010 VT 29, ¶ 18, 187 Vt. 495, 996 A.2d 237 (quotation omitted). We conclude that the State’s evidence met this standard here.
[645]*645B.
¶ 36. The propriety of the jury instruction presents a more difficult question. As an initial matter, the State points out that, although defense counsel objected to the jury instruction during the charge conference, defense counsel did not reassert the nature of the objection after the instructions had been read. “No party may assign as error any portion of the jury charge or omission therefrom unless he objects thereto before the jury retires to consider its verdict, stating distinctly the matter to which he objects and the grounds of his objection.” V.R.Cr.P. 30. Accordingly, the State contends that the objection to the jury instruction has not been preserved. Defendant replies that defense counsel’s statement, “I would like to renew my objection to the doctrine of chances,” was sufficient to satisfy the preservation requirements, particularly in light of the discussion at the charge conference.
¶ 37. We cannot agree. Our cases make clear the requirement of making an objection to jury instructions after the jury is charged. State v. Wheelock, 158 Vt. 302, 306, 609 A.2d 972, 975 (1992) (“[FJailure to object to an instruction after it is given to the jury is considered a waiver of any error even if the substance of the objection is made known before the jury charge.” (emphasis added)); see also State v. Bacon, 163 Vt. 279, 284, 658 A.2d 54, 59 (1995) (instruction error not preserved because “it was not specifically renewed after the charge.”); State v. Pelican, 160 Vt. 536, 538, 632 A.2d 24, 25-26 (1993). As we have explained, “The primary reason for the rule is to give the trial court one last opportunity to avoid an error.” Wheelock, 158 Vt. at 306, 609 A.2d at 975; see also Jones v. United States, 527 U.S. 373, 387-88 (1999) (applying identical Federal Rule of Criminal Procedure 30). While an objection at a charge conference may help explain a complying objection made after the charge is delivered to the jury, it is not a substitute for the complying objection.
¶ 38. As for defense counsel’s renewal of the objection with no explanation of the grounds after the jury was charged, it was insufficient. Rule 30 plainly demands that, before the jury retires, a party must “stat[e] distinctly ... the grounds of his objection,” not merely the objection itself. Here, defense counsel’s cursory renewal, without any statement of grounds, did not comply with Rule 30. Although counsel may have believed that the content of the objection was already well rehearsed, Rule 30 requires at [646]*646least a brief restatement. See State v. Martin, 2007 VT 96, ¶ 42, 182 Vt. 377, 944 A.2d 867 (rejecting defendant’s argument that “a general objection to the court’s instruction sufficiently preserved the issue for appeal”); Winey v. William E. Dailey, Inc., 161 Vt. 129, 138, 636 A.2d 744, 750 (1993) (decided under substantially identical civil rule and holding that “a blanket reference made after the charge to arguments made before the charge, even if allowed by the trial court,” is not sufficient for preservation).12 The deficiency in the language of the objection here is clearly shown by contrasting it with the language of the objection to the jury instructions in the recent case of State v. Kolibas, 2012 VT 37, 191 Vt. 474, 48 A.3d 610. In that case, the defendant attacked a criminal-intent instruction because it allowed a finding of intent if defendant intended to harm someone other than the actual victim. Defense counsel said after the instruction was delivered: “We object to the instruction of intent regarding the aggravated assault charges, that the State does not have to prove there was an intent to harm a specific individual.” Id. ¶ 12. We held that “[defendant made a detailed objection to the court’s instructions on transferred intent, and this issue is clearly preserved for review” because defense counsel stated a ground for the objection — that the instruction provided that the State did not have to prove there was an intent to harm a specific individual. Id. The objection in Kolibas would have been equivalent to the objection in this case if it had omitted the last phrase and said only that defendant objected to the intent instruction regarding aggravated assault.
¶ 39. Here, the deficiency in defense counsel’s objection is particularly significant because of what occurred at the charge conference. Defense counsel was clear at the charge conference that she opposed the giving of the doctrine of chances instruction but less clear on the grounds.13 She argued that it was inappro[647]*647priate to use the doctrine as a jury charge, rather than as a rule of evidence. She claimed the doctrine would allow the jury to find intent for each fire solely because there were four fires. She generally claimed that the instruction would relieve the State of its burden of proof and deny her client due process of law. She argued that the doctrine could rest on uncharged misconduct evidence but not on charged conduct because the latter had to be proven beyond a reasonable doubt. As she developed her objection, the trial judge interrupted and said that they — the judge and the defense attorney — would have to agree to disagree, ending the discussion. We concur with defendant that an argument truncated by the trial court is normally sufficient to preserve unstated grounds for objection. See People v. Boyette, 58 P.3d 391, 412 (Cal. 2002) (holding there was preservation where “[d]efense counsel made several objections that were cut off by the trial court before counsel could articulate fully the grounds of his objection”). However, the issue here is not whether defense counsel preserved the ground for her objection at the charge conference, but instead whether she preserved the grounds after the instruction was read to the jury. Not having been able to fully state the grounds at the charge conference reinforced the need to state them fully after the charge was given. Instead, defense counsel stated no grounds.
¶ 40. Because defendant failed to comply with Rule 30, we review only for plain error.14 See State v. Doleszny, 2004 VT 9, [648]*648¶ 10, 176 Vt. 203, 844 A.2d 773 (explaining that objections not properly renewed will be reviewed only for plain error).
¶ 41. At the outset, we conclude that the jury instruction on the doctrine of chances did constitute error. First and foremost, we conclude that the instruction, as given, was not a correct statement of the law as explained in this opinion. A jury charge will be upheld “[i]f the charge, taken as a whole and not piecemeal, breathes the true spirit of the law, and if there is no fair ground to say that the jury has been misled.” State v. Gokey, 136 Vt. 33, 36, 383 A.2d 601, 602 (1978); see also State v. Rounds, 2011 VT 39, ¶ 30, 189 Vt. 447, 22 A.3d 477 (holding that permissive-inference instruction constitutes error where it misstates legal rule involved allowing inference to be drawn that is not legally authorized). In this case, the jury instruction did not carefully explain the legitimate and improper use of the doctrine of chances. The instruction did not clearly delineate which of the three different versions of the doctrine was proper and did not specifically instruct that the jury could not use propensity reasoning to determine whether defendant was guilty with respect to a specific fire. Indeed, the jury could have understood that it was permissible to find guilt based on propensity. For example, the instruction states, “The improbability of a coincidence may cause you to conclude that the fires were intentionally set.” But this follows only if one infers from the fact that the fires likely were not all accidents to the conclusion that each fire was not an accident — an inference that makes logical sense, but only to the extent one believes human behavior tends to be consistent such that a criminal cause for one fire is likely to explain other fires. This is propensity reasoning. We do not dispute that the evidence of the other fires is relevant and admissible in this case. But we are not convinced that the jury instruction, as given, properly tracked the theory according to which it was admissible.
¶ 42. Second, even if the instruction had clearly delineated which of the versions of the doctrine was proper, the trial court should not have given any instruction on the doctrine of chances. The doctrine of chances is about the admissibility of evidence, which is not generally a matter in which the jury should be instructed. See [649]*649V.R.E. 104(a). It is illuminating that, although the doctrine of chances is a relatively well-worn legal concept, we have not found a single case in which it has been used as a basis for a permissive-inference jury instruction. The concern is that explaining why evidence has been admitted, and the inference that can be drawn from the evidence, risks encouraging the jury to give the evidence additional weight. For this reason, we recently found error in a jury instruction on the admission of evidence in State v. Brandt, 2012 VT 73, ¶21, 192 Vt. 277, 59 A.3d 141. As in Brandt, the charge here specially explained the admissibility ruling, describing to the jury how they could use the evidence. The jury had the evidence of the other fires; there was no reason for the trial court to highlight its potential significance. See State v. Tahair, 172 Vt. 101, 108-09, 772 A.2d 1079, 1085 (2001) (emphasizing difference between what jury might infer on its own and what it might conclude when evidence is highlighted by jury instruction).
¶ 43. Having concluded that the instruction constituted error, the question becomes whether it rises to the level of plain error. See State v. Lambert, 2003 VT 28, ¶ 13, 175 Vt. 275, 830 A.2d 9 (rejecting argument that flawed jury instruction was error per se). We recently itemized the elements of plain error with respect to a jury instruction:
Four factors guide our plain error analysis: (1) there must be an error; (2) the error must be obvious; (3) the error must affect substantial rights and result in prejudice to the defendant; and (4) we must correct the error if it seriously affects the fairness, integrity, or public reputation of judicial proceedings. When reviewing possible error in a jury instruction, we examine the instructions in light of the record evidence as a whole and determine if any error would result in a miscarriage of justice. Moreover, we review the instructions in their entirety. If the charge as a whole is not misleading, there is no plain error. This is a very high bar — we find plain error only in rare and extraordinary cases.
State v. Herrick, 2011 VT 94, ¶ 18, 190 Vt. 292, 30 A.3d 1285 (citations omitted). In Rounds, 2011 VT 39, ¶¶ 32-34, we recently found plain error because of a jury instruction that erroneously allowed the jury to use a permissive inference and misstated the [650]*650elements of that inference. We noted that “[t]he instruction as given allowed the jury to look past defendant’s theory and infer his criminal intent based on insufficient or nonexistent evidence, all to his direct prejudice.” Id. ¶ 33. We concluded that the instruction seriously affected the integrity of the jury verdict, noting that “[i]f the jury relied on the erroneous inference in determining the knowledge element of the crime, it failed to find defendant guilty of all elements of the crime beyond a reasonable doubt” and “the instruction provided the jury a path to convict defendant that was not supported by the evidence and was contrary to the statute’s dictates.” Id. ¶ 34.
¶ 44. We cannot find plain error in this case. It is not a rare and extraordinary case. Although the jury instruction could have resulted in prejudice for defendant and a miscarriage of justice, we can only speculate that it did, and, in fact, the jury verdict is consistent with the proper use of the doctrine of chances.
¶ 45. Our main concern with the jury instruction is that it allowed the jury to convict defendant based on propensity reasoning — because of the number of fires, defendant must be an arsonist and, as an arsonist, he intended to start each fire. But the split jury verdict is strong evidence that the jurors did not engage in such reasoning ■ — ■ otherwise they would have convicted defendant for each fire. The acquittal of the charge from the first incident is particular evidence that the jury saw the doctrine correctly, since the psychological doctrine of chances could not be used to create an inference that defendant’s act was intentional the first time. Thus, conviction for the latter two fires is in line with the third, and valid, version of the doctrine of chances. It also suggests that the jury was influenced by the State’s expert testimony which found a human cause for the last two fires, but not for the first two.
¶ 46. While there are significant similarities between this case and Rounds, there are significant differences. In Rounds, the permissive inference was statutory, and there was a dispute over whether the basic facts were shown sufficiently to allow the jury to find the inferred intent. We held that Vermont Rule of Evidence 303 required that the basic facts be proved beyond a reasonable doubt when the inferred intent is an element of the offense and the jury instruction failed to require this standard of proof. Id. ¶ 27. More importantly, we further held that the key basic fact was not proved beyond a reasonable doubt as a matter [651]*651of law so the statutory permissive inference should not have been included in the jury instruction. Id. ¶ 29. The errors were of constitutional dimension because the permissive inference allowed the State to prove its case without proof beyond a reasonable doubt. Id. ¶ 32.
¶47. While we hold here that the court should not have given the permissive-inference instruction, the concern is not of constitutional dimension. Instead, our primary concern is that the jury might have given too much weight to what is properly understood as an evidence-admission rule. Again, the split jury verdict shows that the jury did not give undue weight to the instruction.
Affirmed.