Residual Credit for Guilty Pleas Entered After the Trial Has Begun:
Commentary on R v Szczerbacz [2025] EWCA Crim 1576

1. Introduction

R v Szczerbacz [2025] EWCA Crim 1576 is a Court of Appeal (Criminal Division) authority on a narrow but recurrent issue in sentencing practice: how to apply the Sentencing Council’s guideline on reductions for guilty pleas where the defendant pleads guilty after the trial has already begun.

The appeal was formally against sentence, but in substance it focused on a single, short point of principle: whether a sentencing judge may simply refuse any credit for a guilty plea because it was entered on the second day of trial, or whether the judge must still consider whether some residual reduction is justified in light of the benefits actually produced.

The Court of Appeal held that the sentencing judge erred in principle by treating the late stage of plea as automatically excluding any reduction. Reaffirming and refining earlier authorities (Mason, Ray Carroll, Ali), the Court emphasised that even once a trial has started there remains a possible (though usually small) reduction in sentence – dependent on the real, case-specific benefits of the plea.

In doing so, the judgment provides important guidance on:

how to read and apply the Sentencing Council’s guideline on Reduction in Sentence for a Guilty Plea;
the proper scope of judicial discretion when pleas are entered during trial; and
the distinction between utilitarian credit for a plea and other mitigating factors such as remorse.

2. Factual and Procedural Background

2.1 The offence

The appellant, Dawid Szczerbacz, stabbed the victim, Patric Szabo, in the area between the neck and collarbone with a kitchen knife (paras 3–6). The stabbing occurred at the home of Ms Monica Wilk and her partner, long-standing friends of Mr Szabo, who had allowed the appellant to sleep on their sofa after he became homeless.

Key factual points:

There had been prior tension between the appellant and the victim, including accusations about the use of the appellant’s phone (paras 4–5).
The appellant had earlier threatened to stab Mr Szabo (para 5).
On 24 September 2023 a knife went missing; Ms Wilk later found a knife in the appellant’s pocket, and had seen him with a knife from the kitchen (para 5).
After midnight, when Mr Szabo returned and knocked at the locked front door, the appellant opened it with a kitchen knife in his hand and stabbed him without warning (para 6).

Fortunately, prompt medical treatment avoided a collapsed lung, but the seriousness of the injury plainly justified a charge of a grave violent offence.

2.2 The investigation and charges

The appellant denied responsibility when arrested and interviewed (para 7). He suggested that someone else must have attacked the victim. He was charged with:

attempted murder; and
wounding with intent to cause grievous bodily harm (a section 18 offence).

On 26 February 2024 he served a Defence Statement denying:

any arguments with the victim;
any threats to stab him; and
the stabbing itself.

The Defence Statement also required the attendance of key witnesses, including the victim and Ms Wilk (para 8), thereby confirming that a contested trial on the issue of guilt was anticipated.

2.3 The trial

Trial preparation was fully completed: edited interviews, agreed facts, special measures applications and other evidential rulings (para 9). The jury were sworn, and the first day was taken up with procedural matters; no opening or evidence took place (para 9).

On the second day of the trial, 29 March 2024, the appellant changed his plea:

He pleaded guilty to wounding with intent.
No evidence was offered on the attempted murder count, and an acquittal was directed (para 10).

2.4 Sentencing at first instance

The judge, having considered a pre-sentence report, a psychiatric report and the appellant’s Polish antecedents, found that the statutory criteria for “dangerousness” were met (para 11). He imposed an extended sentence under the dangerousness provisions:

8 years’ custody; plus
5 years’ extended licence;
total extended sentence: 13 years (para 2).

The judge gave no reduction for the guilty plea, explaining that:

“Given the late stage at which your guilty plea was entered, there can be no credit or reduction of your sentence as might have been the case had you pleaded guilty earlier in the proceedings.” (para 12)

He also found the appellant to be “in no sense remorseful” and criticised him for continuing to deny responsibility and to blame his lawyers (para 12).

2.5 The appeal

The appellant did not challenge:

the dangerousness finding; or
the overall structure and length of the sentence, save as it might be adjusted for guilty plea credit.

The single ground of appeal was that the judge erred in law in refusing to allow any reduction in sentence for the guilty plea on day two of the trial (para 12).

On appeal, Mr Greenhalgh (for the appellant) argued that:

the plea meant the victim and other witnesses did not, in fact, have to give evidence;
interpreters and other trial resources could be released; and
the three-day listing could be used for other cases;

and that these features justified some level of reduction (para 13).

3. Summary of the Judgment

The Court of Appeal (Lewis LJ giving the judgment) allowed the appeal to a limited extent. The key conclusions were:

Error of principle: The sentencing judge’s statement that there “can be no credit” once a plea is entered at such a late stage was wrong as a matter of principle. Even where a plea is entered during the trial, the Sentencing Council guideline and authorities such as Mason and Ray Carroll make clear that some reduction may still be appropriate (para 19).
Residual discretion even during trial: The guideline language (“reduced further, even to zero, if the guilty plea is entered during the trial” – para 15) envisages a range from a small reduction down to zero; it does not create a default rule of “no credit” once the trial has begun (paras 15–18).
Case-specific assessment: Having recognised this error, the Court undertook its own assessment. It acknowledged that:
- the trial was short (three days), and fully prepared;
- no plea indication was given before the second day;
- the investigation and pre-trial preparation had already been completed;
- the trial was underway, but no evidence had yet been led; and
- the plea spared the victim and Ms Wilk from actually testifying, though they had prepared to do so (para 20).
The Court assessed that there was a small but real utilitarian benefit.
Quantum of reduction: The Court concluded that “some limited reduction” was justified, but “well below 10%” (para 21). It fixed a modest reduction of “about 5% (or 4 months)” (para 21).
Re-sentencing outcome:
- The extended sentence was quashed and re-imposed in a slightly shorter form.
- Custodial term reduced from 8 years to 7 years 8 months (4 months’ reduction).
- Extended licence period of 5 years left unchanged (para 21).
- Total extended sentence: 12 years 8 months (7y 8m custody + 5y licence).

The judgment therefore affirms that a plea entered on the second day of trial can still attract some credit, even where the defendant has maintained a defence until that point and denied guilt in interview and in his Defence Statement.

4. Legal Framework

4.1 Statutory provisions

Two provisions of the Sentencing Act 2020 frame the Court’s reasoning:

Section 73 – Reduction in sentence for guilty pleas
Section 73 requires the court, when sentencing a defendant who has pleaded guilty, to take into account:
- the stage in the proceedings at which the offender indicated the intention to plead guilty; and
- the circumstances in which the indication was given.
This section embeds the principle that timing matters: the earlier the plea, the greater the potential reduction.
Section 59 – Duty to follow sentencing guidelines
Section 59 requires courts to follow Sentencing Council guidelines unless doing so would be contrary to the interests of justice. As a result, the Sentencing Council guideline on Reduction in Sentence for a Guilty Plea has near-mandatory force.

4.2 The Sentencing Council guideline: Reduction in Sentence for a Guilty Plea

The Court summarises the central rationale of the guideline (paras 14–15):

to encourage defendants who are going to plead guilty to do so as early as possible;
because an acceptance of guilt typically:
- reduces the impact of the crime on victims;
- saves victims and witnesses from having to give evidence; and
- promotes the public interest by saving time and money.

Key structural elements include:

Maximum one-third reduction where a guilty plea is indicated at the first reasonable opportunity (usually the first appearance in the magistrates’ court).
Maximum one-quarter reduction if the plea is entered after the first stage of proceedings but before the day of trial.
Maximum one-tenth reduction if the plea is given on the first day of trial but before the trial “starts” in the sense of evidence being called.
Plea during trial: the guideline provides that the maximum reduction “should be reduced further, even to zero, if the guilty plea is entered during the trial” (para 15).

The phrase “even to zero” is the pivotal wording in this case. The Court in Szczerbacz, following Mason and Ray Carroll, reads this as conferring a range of discretion (0–10%) once the trial has started; it does not mandate nil credit simply because evidence has begun (or is about to begin).

5. Precedents Cited and Their Influence

5.1 R v Mason [2022] EWCA Crim 1830; [2023] 2 Cr App R (S) 8

In Mason, the appellant faced serious drugs charges and initially pleaded not guilty. On the first day of trial, applications for adjournment (on medical and evidential grounds) were refused. The judge was informed of a possible resolution and allowed time for discussions (para 16). The jury were sworn and the Crown opened the case; a body-worn video was shown in evidence. The following morning — the second day of trial — the appellant pleaded guilty.

The trial judge in Mason stated there was “no credit” for the plea because the trial had already started (save for one count treated differently), imposing 16 years’ imprisonment (para 16).

The Court of Appeal set out and applied the guilty plea guideline in detail. Its key observations, quoted in Szczerbacz at para 16, include:

“The Sentencing Council has left much to the good judgment of trial judges as to the precise level of credit once the trial has started. There may be proper reason to allow no credit once a trial has started, or to allow some credit, albeit less than 10%. The stage at which the trial has reached may be material. Each case turns on its own facts. This court will not lightly interfere with the exercise of judgment on the facts by a trial judge adopting the correct approach set out in the guideline.”

However, in Mason:

the trial was only in its early stages (no witnesses had given oral evidence apart from the video);
the judge acknowledged there was a “public saving” as a result of the plea (para 17); but
the judge nonetheless treated himself as unable or unwilling to reflect that saving by any reduction.

The Court of Appeal found this was too rigid. It allowed a modest reduction of six months, from 16 years to 15 years and 6 months (para 17).

In Szczerbacz, Mason is used to reinforce the proposition that:

even after the trial has started, some credit may be appropriate;
the amount is fact-sensitive and may be very small, but should reflect genuine utilitarian benefits;
an inflexible “no credit once trial has begun” stance is inconsistent with the guideline.

5.2 R v Ray Carroll [2024] EWCA Crim 779

Ray Carroll reviewed Mason and earlier authorities, including Ali (see below). It confirmed that the guideline’s reference to reductions “even to zero” during trial does not remove the possibility of some credit.

Critically, Ray Carroll held that:

The weight of the evidence against the defendant is not a relevant factor in deciding whether a reduction for a late plea is appropriate (para 18). In other words, it is wrong to deny or reduce credit merely because the case against the defendant was overwhelming.
The judge’s task is to consider:
- the stage at which the plea was made; and
- “all the circumstances” – i.e. the extent of real savings in time, resources, and trauma to witnesses.

In that case, the trial judge erred by treating the strength of the prosecution case as relevant to the availability of credit. The Court of Appeal corrected that error and granted a six-month reduction (from 7 years 9 months to 7 years 3 months) (para 18).

In Szczerbacz, Ray Carroll supports two important principles:

Even a late plea can justify modest credit.
The focus must remain on utilitarian benefits, not on how easy conviction would have been had there been a trial.

5.3 R v Ali [2022] EWCA Crim 1884; [2023] 1 Cr App R (S) 52

Ali is referred to in Ray Carroll as part of the line of authority recognising that the guideline allows for a range of reductions below 10% when a plea is entered during a trial (para 18). Although the present judgment does not detail Ali’s facts, its doctrinal relevance lies in:

confirming that the guideline’s wording is to be read as conferring discretion, not as a rigid formula; and
emphasising that the stage of proceedings and the real-world consequences of the plea are the primary considerations.

Szczerbacz therefore fits into and slightly sharpens an existing jurisprudential trajectory: Ali –> Mason –> Ray Carroll –> Szczerbacz.

6. Analysis of the Court’s Legal Reasoning

6.1 Identifying the error of principle

The heart of the appeal lies in para 19, where the Court contrasts the sentencing judge’s approach with the proper interpretation of the guideline and the case law. The judge had said:

“Given the late stage at which your guilty plea was entered, there can be no credit or reduction of your sentence…”

The Court held that this was wrong “as a statement of principle” (para 19). The correct position is:

Once the trial has started, the guideline permits reduction to be “reduced further, even to zero” (para 15); that means the possible range is from less than 10% down to 0%, not that the default is 0%.
It is an error to assume that the mere fact of a plea on day two, without more, excludes the possibility of any reduction.

In other words, the sentencing judge’s comments show that he treated the start of the trial as a bright-line cut-off. The Court of Appeal, relying on Mason and Ray Carroll, makes clear that the guideline requires a nuanced and case-specific assessment even at this late stage.

6.2 Application of the guideline to the specific facts

Having identified an error of principle, the Court then applied the correct approach itself (paras 20–21). Its analysis can be broken down as follows.

6.2.1 The stage of proceedings

Relevant features:

The trial was listed for three days (para 20).
All preparation and pre-trial work had been done (para 20).
The jury were sworn, but on day one the case was not opened and no evidence was taken (para 9).
On the morning of the second day, before any witnesses had given evidence, the plea was entered (para 10).

The trial was therefore “underway” in a formal sense, but substantively at a very early stage. The judge had not yet heard from the victim or other witnesses; the prosecution’s case had not yet been developed in front of the jury.

6.2.2 The extent of utilitarian benefit

The Court then examined the three familiar rationales for guilty plea credit (para 15) and assessed the extent to which they were engaged here (paras 20–21):

Reduction of impact on the victim (para 15(a), para 20):
The Court found “nothing to suggest the appellant’s guilty plea in any way reduced the effect of the crime on the victim” (para 20). The plea did not change the underlying harm caused by the stabbing.
Saving victims and witnesses from testifying (para 15(b), para 20):
This factor was engaged:
- The victim and Ms Wilk had expected to give evidence up to the morning of day two.
- Because of the plea, they were spared the ordeal of entering the witness box and having to relive events in front of a jury (para 20).
This is a significant human benefit, even though they had already experienced the stress of preparing to testify.
Saving public time and resources (para 15(c), para 20):
Here, the Court was candid that the saving was “some, albeit very little” (para 20):
- Investigation and trial preparation were already complete.
- The trial had begun, but had not yet used court time for evidence.
- Stopping on the morning of day two produced some limited economy in court time.

Overall, the Court characterised the utilitarian benefits as modest but real:

little saving in terms of public resources; but
a more meaningful benefit in sparing the victim and a key witness from giving live evidence.

6.3 The quantum of reduction: “about 5%”

Having concluded that some reduction was justified, the Court then considered its size. The key passage is para 21:

“In the circumstances there is a benefit in encouraging defendants to admit their guilt. There is scope for some limited reduction in such circumstances, albeit well below 10%. In all the circumstances of this case, given the limited benefits of the trial coming to an end sooner so that no further resources needed to be spent on the trial and given that the witnesses did not have to testify, we would make a modest reduction of about 5% (or 4 months) in the sentence.”

Two points deserve emphasis:

Position within the 0–10% band:
The Court explicitly situates the reduction at “well below 10%” and expresses it approximately (“about 5%”). This illustrates the discretionary, non-mechanical nature of the assessment.
Application to the custodial term only:
The 4-month reduction is applied to the 8-year custodial component of the extended sentence, reducing it to 7 years 8 months; the 5-year extended licence is left untouched (para 21). This reflects the established principle that guilty plea credit goes to the punitive, not the protective, element of an extended sentence.

The result is:

Original: 8 years’ custody + 5 years’ extended licence = 13 years.
Substituted: 7 years 8 months’ custody + 5 years’ extended licence = 12 years 8 months.

6.4 Relationship with earlier authorities

Szczerbacz does not break new conceptual ground, but it:

clarifies and reinforces the line of reasoning from Ali, Mason and Ray Carroll;
illustrates how a very late plea (day two of trial) can nonetheless yield a small credit; and
corrects a still-prevalent misreading of the guideline as imposing a near-automatic “no credit once trial has started” rule.

In practical terms, it shows that:

The guideline’s “even to zero” language is not a threshold rule but an end-point on a spectrum: 10% (first day, before evidence) –> small but >0% (early in the trial) –> 0% (in appropriate cases where the plea is so late that it yields no real benefit).
Appeals will still be subject to the usual restraint (“This court will not lightly interfere…”, per Mason), but where a judge has misstated the principle (as here) the Court will intervene.

6.5 Credit for plea vs. lack of remorse

An important, if implicit, aspect of the reasoning concerns the distinction between:

utilitarian credit for a guilty plea; and
personal mitigation arising from genuine remorse.

The sentencing judge observed that the appellant was “in no sense remorseful” and continued to deny responsibility, even attempting to blame his lawyers (para 12). Those observations go to personal mitigation, not to the utilitarian rationale for plea credit.

The Court of Appeal, while not criticising that assessment, nevertheless allowed a reduction based on the concrete benefits that flowed from the plea: less court time, and spared testimony for the victim and a key witness. This underscores that:

Guilty plea reduction is awarded for its practical consequences (savings, avoidance of distress), not as a reward for moral contrition.
A defendant may receive plea credit even if found to lack remorse; conversely, remorse without early plea may act as separate mitigation but will not attract guideline credit.

7. Complex Concepts Explained in Plain Terms

7.1 Extended sentences and “dangerousness”

An extended sentence is imposed where:

the offender is assessed as “dangerous” under the Criminal Justice Act 2003 regime (reflected in the Sentencing Act 2020); and
the court concludes that an ordinary determinate sentence would not adequately protect the public.

It has two parts:

Custodial term – the period of imprisonment that reflects the seriousness of the offence and previous offending.
Extended licence – extra time on licence after release, during which the offender can be recalled for breach or further offending.

Guilty plea reductions normally apply only to the custodial term, because the licence is a public protection measure, not a punishment. Szczerbacz is a textbook example: the 8-year custodial term was reduced; the 5-year licence remained unchanged.

7.2 “Credit” for a guilty plea

“Credit” (or reduction) for a guilty plea is not a discretionary indulgence; it is a structured response to:

the earlier acceptance of guilt;
the savings in time and money; and
the avoidance of distress to witnesses.

The Sentencing Council guideline sets default maximums:

Up to one-third off for the earliest pleas.
Up to one-quarter for later, but still pre-trial, pleas.
Up to one-tenth for pleas on the first day of trial before evidence.
0–10% for pleas during trial, to be assessed case by case.

The key message from Szczerbacz is that even where the plea is very late, the judge must still ask: “Did this plea save anything, and if so, how much?”

7.3 The “first stage of proceedings”

The guideline’s “first stage of proceedings” typically means:

in magistrates’ courts: the first hearing where the defendant can reasonably be expected to indicate a plea;
for either-way or indictable-only matters proceeding to the Crown Court: generally, the first appearance in the Crown Court where a plea is taken (e.g. the Plea and Trial Preparation Hearing, PTPH).

Pleas indicated at this earliest realistic point receive the largest discount (normally one-third) because they maximise all three utilitarian benefits. Szczerbacz illustrates the other end of the spectrum.

7.4 “Interests of justice” and departure from guidelines

Section 59 of the Sentencing Act 2020 allows a court to depart from a Sentencing Council guideline only if it would be contrary to the interests of justice to follow it. The Court in Szczerbacz does not suggest any such departure; instead, it corrects the judge for misapplying the guideline.

This underlines that the guideline is the starting point and default framework. Discretion operates mainly within the ranges it specifies (e.g. deciding whether to give 0%, 3%, or 8% within the 0–10% “during trial” band), rather than by ignoring the guideline altogether.

8. Impact and Implications

8.1 For sentencing judges

The judgment sends clear signals for first-instance judges:

Avoid bright-line cut-offs.
It is an error of law to say or assume that “once the trial has started, there can be no credit.” The guideline contemplates a range down to zero; it does not switch credit off automatically.
Record the reasoning.
Sentencing remarks should:
- identify the stage at which the plea was entered;
- consider each of the potential benefits (victim impact, witnesses, resources); and
- express the selected percentage or approximate adjustment.
Recognise small but real benefits.
Even where preparation is complete and the trial has begun, if:
- no witnesses have yet given evidence; or
- the plea significantly shortens the trial;
a modest reduction (as in Mason and Szczerbacz) may be required.
Distinguish utilitarian credit from remorse.
Lack of remorse may legitimately affect other aspects of sentence (e.g. aggravating/mitigating balance), but should not be conflated with the separate question of whether the plea saved time or spared witnesses.

8.2 For defence practitioners

For defence lawyers advising clients on plea timing, Szczerbacz reinforces several points:

Earlier is always better.
The difference between a first-opportunity plea and a second-day-of-trial plea is enormous: from around one-third to perhaps as little as 5% — or even 0% — depending on the case.
There may still be marginal benefit in a late change of plea.
If a client insists on contesting the case until trial but then wishes to plead, it is still correct to advise that:
- a plea before any witness is called is more likely to gain some credit;
- credit diminishes as more of the trial is expended; but
- even at a relatively late stage, a plea might justify a small reduction, particularly if it shortens cross-examination or avoids vulnerable witness evidence.
Appeal prospects.
Where a sentencing judge expressly states that “no credit is possible” merely because the trial has started, Szczerbacz provides a clear appellate foothold: that is a misdirection in principle.

8.3 For victims and witnesses

The judgment takes seriously the strain placed on victims and witnesses who expect to testify, even if the plea comes at a late stage.

It recognises that being spared the ordeal of giving evidence is a meaningful benefit, even where anxiety has already been suffered in anticipation.
It underscores that the criminal justice system values their well-being as a distinct component of sentencing calculus, not just as an afterthought.

8.4 For the wider law of sentencing

Doctrinally, Szczerbacz:

consolidates the reading of the guilty plea guideline as flexible but structured;
illustrates the Court of Appeal’s willingness to correct categorical misstatements of principle, while otherwise respecting the trial judge’s evaluative judgment;
clarifies that post-commencement plea credit lies on a sliding scale rather than being an all-or-nothing concept.

Over time, this should promote greater consistency in sentencing practice and reduce the frequency of appeals based on rigid, guideline-inconsistent statements such as “no credit is available once the jury is sworn.”

9. Conclusion: Key Takeaways

R v Szczerbacz confirms and refines the law on guilty plea reductions in the specific, but common, situation where the defendant pleads guilty during trial. The core lessons are:

No automatic bar after trial begins.
The fact that a trial has started does not automatically exclude the possibility of any reduction for a guilty plea. The Sentencing Council guideline permits reductions “even to zero” – implying a range, not a rule.
Fact-sensitive assessment within a 0–10% band.
Once evidence has begun (or is about to begin), the available reduction is generally somewhere between 0% and 10%, with the exact figure depending on:
- how much of the trial has been used;
- whether witnesses have already testified;
- what savings (if any) in time and resources are achieved; and
- the extent to which the plea spares victims and witnesses from giving evidence.
Separation of utilitarian credit and remorse.
Credit is awarded for the practical benefits of the plea, not as a moral reward. Lack of remorse does not automatically preclude plea credit if there are still tangible savings.
Application to extended sentences.
Credit should be applied to the custodial element of an extended sentence, not to the protective licence period. Here, a modest 5% reduction (4 months) was applied to an 8-year custodial term.
Importance of accurate sentencing remarks.
Sentencing judges must accurately reflect the guideline and explain why a particular level of reduction within the relevant band is appropriate. Statements that “no credit is possible” simply because of the stage reached will invite successful appeal.

In essence, Szczerbacz cements a nuanced, principled approach: while defendants who delay their pleas until the trial is underway should expect only very modest reductions, the system must still recognise and measure any genuine late-stage benefits their plea provides.

Residual Credit for Guilty Pleas Entered After the Trial Has Begun: Commentary on R v Szczerbacz [2025] EWCA Crim 1576