Introduction
Day 1 delivered. Three winners from seven races, a double that returned over £200 on a £20 stake, and a day where the filters demonstrated exactly why we trust them. But let’s be honest — the result sheet tells only part of the story. Winston Junior ran a cracking race into 2nd in the Fred Winter, Jagwar was touched off in the Ultima, and Lulamba was beaten in the Arkle by a Mullins horse the data had flagged but the filter threshold narrowly excluded. Near misses sting. The data, though, held its shape throughout.
And then there was Lossiemouth. She sprinted clear up that famous Cheltenham hill in the Champion Hurdle and it was something to behold — powerful, relentless, exactly the kind of performance the convergence data had pointed toward. TPR top 4, hot trainer, market confirmation. The model did not find her by accident. That is what five years of festival data looks like when it fires.
Day 2 brings the Queen Mother Champion Chase, the Brown Advisory Novices Chase, and the Champion Bumper — three races that between them cover the full spectrum of what our framework does well. The novice chase and the Queen Mother are exactly the categories where Filter E and Filter A have produced their best festival returns. The bumper is where we put the speed figures down and let the breeding data lead instead. We have clear views on all three.
🏇 Today’s Card at a Glance
Seven races across Wednesday, with Grade 1 racing the dominant theme again:
- 13:20 – Turners Novices Hurdle (Grade 1) – 2m5f, 22 runners
- 14:00 – Brown Advisory Novices Chase (Grade 1) – 3m1f, 16 runners
- 14:40 – BetMGM Cup Handicap Hurdle – 2m5f, 24 runners
- 15:20 – Glenfarclas Cross Country Chase – 3m5.5f, 14 runners
- 16:00 – BetMGM Queen Mother Champion Chase (Grade 1) – 2m, 10 runners
- 16:40 – Grand Annual Challenge Cup Handicap Chase – 2m, 20 runners
- 17:20 – Weatherbys Champion Bumper – 2m, 22 runners
The BetMGM Queen Mother Champion Chase is the centrepiece — the championship two-mile chase that sits at the very top of our Grade 1 chase category, the single most profitable race type in our entire five-year dataset. Ten runners, a clear convergence qualifier, and a race the model has very strong feelings about. The Brown Advisory at 14:00 is a 16-runner novice chase where Gordon Elliott’s stable has sent a horse the data rates highest. And to close the day, the Champion Bumper — twenty-two lightly-raced horses where breeding figures take over from speed figures as the primary analytical tool.
🎯 How We Approach The Data
Our analysis is built around the TPR (Total Performance Rating) – a machine learning-derived metric that combines a horse’s speed figures, going preferences, track record, and trainer momentum into a single forward-looking score. Across our database, horses ranked in the top three on TPR win approximately 55% of all races. The top four account for around 67%. At five consecutive Cheltenham Festivals, the convergence filters built on TPR have returned positive ROI in every graded race category except heavy ground and handicap hurdles.
The supporting metrics you will see throughout:
- T-1 through T-5 – Class-adjusted speed figures from the last five runs. T-1 is the most recent and carries the most weight. Where a figure was posted matters as much as what it says – a T-1 of 160 at Cheltenham on Good To Soft in January is very different from the same number at a flat track on Good ground in October.
- GoingTPR – Average speed rating on today’s ground conditions. A GoingTPR of 0 means a horse is unproven on this going – a meaningful flag on a day when the track demands genuine versatility from every runner.
- TrackTPR – Cheltenham-specific performance. The undulations, the climb to the finish, the drying effect on the far side – this track asks questions that not every horse can answer, and TrackTPR tells us who has answered them before.
- Trainer Momentum – Comparing a trainer’s 1-month average against their 12-month baseline. Day 1 reinforced this forcefully — every hot trainer runner hit the frame or won. We apply this overlay rigorously throughout Day 2.
One strategic note before we begin: our five-year dataset draws a sharp distinction between race types. Novice graded races return 66.7% winners for our highest-confidence filter. Grade 1 chases return 61.1%. Handicap hurdles return 5.6% for SP favourites and barely more for anything else. Day 1 confirmed all three of those findings in real time. We apply different frameworks to different races and we tell you which is which throughout. There is no single approach that works everywhere at Cheltenham — the data is clear on that — and intellectual honesty about the limits of the model is as important as confidence where it is genuinely warranted.
Let’s get into the card.Let’s get into the card.
Race 1: 13:20 Turners Novices Hurdle (Grade 1)
2m5f | Novice Hurdle | Grade 1 | Good To Soft | 22 runners
Race Analysis
Twenty-two runners over two and a half miles on Good To Soft ground — stamina and Cheltenham experience will matter as much as raw class. No horse in the field satisfies our full Filter E criteria, but the convergence of data and market on a single horse is clear enough in a novice Grade 1.
The model’s highest-rated horse this morning is actually SKYLIGHT HUSTLE, which heads the TPR standings at 64. But it goes off at 14/1, placing it 8th in the market — and under our rules, when the top-rated horse is dismissed that comprehensively by the betting, the historical record is unambiguous. Horses rated first by the model but ranked 4th or worse in the market have won just 2.3% of festival races over five years. SKYLIGHT HUSTLE is a pass. That drift also tells us something useful — the market has had overnight to look at this race and it has actively moved away from our TPR leader.
That brings us to NO DRAMA THIS END, which is TPR rank 2 at 62 and has shortened to 2/1 overnight. The T-1 of 161 came at Newbury on Good in late December, “soon led – not fluent 4 out – ridden briefly after last – nudged along and kept on well final 110yds – comfortably.” The T-2 of 167 — the stronger of the two figures and the best class-adjusted speed figure in the field’s recent form — came a month earlier at Sandown on Soft, “headway and led 2 out – clear run-in – easily.” Both wins, both front-running performances.
The Cheltenham angle seals it. NO DRAMA THIS END won at this course and over today’s exact trip on Soft in November — “travelled strongly – led last – went clear run-in – impressive” — giving a TrackTPR of 88, the fourth best course figure in the field. The GoingTPR of 67 looks soft but is misleading, dragged down entirely by a bumper run at last year’s festival where the horse was bumped and weakened on unsuitable ground. The hurdle figures on Soft read 167 and 108. There is no genuine going concern here.
The main dangers are SKYLIGHT HUSTLE and KING RASKO GREY (7.5/1), though neither represents a confident case at their prices. SKYLIGHT HUSTLE posted a T-1 of 155 when landing a Grade 1 at Leopardstown on Good To Soft in December, “led approaching last – ran on well,” and the Elliott yard is marginally above its baseline at 68 against a 12-month figure of 67. But every career run has been over two miles and the DistTPR is 0. The step to two and a half is an unresolved stamina question — and the market’s move to 14/1 overnight suggests the betting public agrees. KING RASKO GREY has Paul Townend aboard, always a factor at this festival, but a T-1 of just 56 from a third at Leopardstown on Heavy in February — “lost position towards finish” — has damaged the model profile considerably. TPR rank 8 and that recent figure make it hard to justify at 15/2.
ACT OF INNOCENCE (8/1) is consistent but has a GoingTPR of 0, meaning no proven form on today’s ground, and the T-1 of 144 trails the top two in the market. BOSSMAN JACK’s T-2 of 200 from Ffos Las on Soft is eye-catching but was achieved in a significantly weaker contest, and the T-1 of 113 from Chepstow showed a notable drop in quality. Paul Nicholls’ yard is marginally below its seasonal average — one-month TPR of 80 against a 12-month baseline of 85 — the one mild cloud over an otherwise clear selection.
The selection is NO DRAMA THIS END. TPR rank 2, market leader at 2/1, proven Cheltenham winner over today’s trip. The model’s top-rated horse has drifted to 14/1 overnight — when the market makes that kind of statement, you listen to it. Model and market are aligned on the horse that matters.
🎯 SELECTION
NO DRAMA THIS END
Race 2: 14:00 Brown Advisory Novices Chase (Grade 1)
3m1f | Novice Chase | Grade 1 | Good To Soft | 16 runners
Race Analysis
The Brown Advisory is the championship novice chase of the week — three miles and a furlong on Good To Soft ground, a test of jumping, stamina and class in equal measure. The TPR hierarchy is clear and the market agrees, which in a Grade 1 novice chase is the strongest convergence signal available.
ROMEO COOLIO leads the ratings at 78 — five points clear of second-placed KOKTAIL DIVIN at 73, with WESTERN FOLD a more distant third at 54 — and is the 2/1 market favourite. The Filter A qualification holds up firmly: TPR rank 1, GoingTPR rank 3 with a genuine figure of 148, and SP rank 1. Gordon Elliott’s yard is running hot with a one-month TPR of 68 against a 12-month baseline of 67, and Jack Kennedy takes the ride. The T-2 of 235 — the standout class-adjusted figure in this entire field — came at Leopardstown on Good To Soft in December, where he “pushed along after 2 out – went second and 1 length down last – ridden and challenging final 110yds – led towards finish.” The T-1 of 112 from Leopardstown on Soft in February requires context: ROMEO COOLIO won that race but “idled run-in – ridden and reduced lead towards finish – just did enough” at a short price. This is a horse that does what is asked and no more, not one in decline.
The one question mark is DistTPR = 0. All of ROMEO COOLIO’s chasing form has come between 2m1f and 2m4f — today’s 3m1f is new territory. Kayf Tara is a reliable stamina influence and the Fairyhouse win in November over 2m4f, “led going easily 2 out – clear approaching last – eased towards finish – impressive,” suggests a horse with plenty in reserve. The trip rise is a legitimate flag rather than a dealbreaker.
The most credible alternative on the data is FINAL DEMAND (9/2), which holds the best GoingTPR in the field at 163, a TrackTPR of 108, and Cheltenham experience — third in last year’s novice hurdle there on Good To Soft, “pressed leader home turn – not fluent last – outpaced run-in.” Townend takes the ride and Mullins’ yard is the hottest in training (87 vs 82 baseline). But the T-1 of 73 from Leopardstown in February is a hard number to ignore — “Led – not fluent 9th – headed just before 3 out – not fluent when pressed leader 2 out – soon lost ground – weakening when lost second run-in.” That was a genuine weakening performance, not an easing down, and at TPR rank 6 the model’s case relies almost entirely on the going and course credentials rather than current form figures.
WENDIGO (7/1) has a TrackTPR of 124 from a Cheltenham run last March where the comment notes the horse was “hampered and stumbled badly just after 2 out – headway approaching last – stayed on” — an excuse that upgrades that run considerably. The T-2 of 192 from Kempton on Good is strong. But a T-1 of just 48 from a soft Ayr win — “went clear before 5 out – easily” in a race where he was clearly superior to inferior opposition — and a TPR rank of 12 make this a supplementary interest at best.
KAID DAUTHIE (11/2) has Walsh aboard and market support the model does not justify. TPR rank 10, GoingTPR of 73, no Cheltenham course form, and a PU at last year’s festival in the novice hurdle — “stumbled badly 1st – tailed off when pulled up after 2 out.” Pass.
The selection is ROMEO COOLIO. Filter A qualifier, dominant TPR leader, hot Elliott yard, model and market fully aligned. The trip rise is the only caveat in an otherwise clear-cut case.
🎯SELECTION
ROMEO COOLIO
Race 3: 14:40 BetMGM Cup Handicap Hurdle
2m5f | Handicap Hurdle | Class 1 | Good To Soft | 24 runners
Race Analysis
Twenty-four runners in a Premier Handicap Hurdle is as tough as it gets at this festival. Our backtesting is blunt on this category — it is the hardest race type to crack, SP favourites win less than 6% of the time, and even our strongest filters produce modest returns. We are working with minimum stakes here and complete honesty about what the data does and doesn’t support.
The first job in any race is to identify which horses our model rates most highly — that is the TPR rating, a number between 0 and 100 generated by machine learning that combines all available data into a single score. Then we overlay going form: GoingTPR measures how well a horse has performed specifically on today’s ground conditions, in this case Good To Soft. A GoingTPR of 0 means the horse has never shown it can handle this going — a genuine unknown. Finally we look at market price, because when our ratings and the market disagree, the market usually knows something we can’t see.
Two horses fall at the first hurdle. FORTY COATS leads our TPR standings at 86 but the market has it at 10/1 — that puts it outside the top three in the betting, which in our five-year dataset represents a hard veto. The top-rated horse that the market dismisses has won just 2.3% of the time at this festival. Pass. STORM HEART is the 9/2 market favourite but has a GoingTPR of 0 — its runs on Good To Soft are non-existent, with all form coming on Soft or Heavy ground. Backing a handicap hurdle favourite with no proven form on today’s specific going is precisely the pattern our data most strongly warns against. Pass.
That brings us to KATEIRA. There is no clean Filter A qualifier in this race — the top three on TPR all fail the going or market criteria — but KATEIRA is the closest thing to a convergence case the field offers. She sits at TPR rank 4 (score 49), GoingTPR rank 2 (figure 121, meaning she has genuinely performed well on Good To Soft ground), and a market price of 13/2 that puts her third in the betting. When the model and the market are pointing at the same horse in a race this difficult, that alignment deserves respect even if the case isn’t watertight.
The speed figures tell a complex story. Think of these as standardised performance scores — 100 is par for the class, higher is better. The T-2 of 185 was posted at Musselburgh on Good To Soft in January, but a check of the form reveals she finished eighth that day and “weakened before 2 out” — the figure is high but the run was a non-performance, not the kind of evidence you build a case on. The most recent run — T-1 of 121 from Ascot on Good To Soft in February — was also below expectations, with the jockey reporting the horse “was never traveling.” That phrase matters. It is not an excuse like poor ground or interference; it is the rider saying something simply wasn’t right on the day. Two below-par runs in a row demand acknowledgment. Dan Skelton’s yard is also running slightly below its seasonal average — one-month TPR of 75 against a 12-month baseline of 78 — so the trainer momentum signal is mildly negative rather than positive.
What keeps KATEIRA in contention is that the underlying quality is clear when she is right — the T-5 of 172 from Wetherby on Good in November, “led clearly before 3 out – clear before 2 out – pushed out run-in,” shows a horse capable of a high-class performance. And the going credentials are genuine, not an artefact. If the yard has got the horse back to its best, the price of 13/2 represents fair value for the best convergence case available in a 24-runner field.
GUARD DUTY is worth a mention for interested readers. The T-1 of 196 from Doncaster on Soft in January is the best recent speed figure in the entire field — “led before last – kept on” — and the T-3 of 183 at Newbury on Soft confirms this is a horse capable of big numbers on testing ground. At 18/1 it sits outside our TPR top four and the trainer is running cold (Lavelle 68 vs 81 baseline), so it does not qualify under our framework. But those figures are hard to ignore entirely at the price.
The selection is KATEIRA — the horse in this field where our model, going form, and the market most closely align, even if the convergence is looser than we would like. The recent running concerns are real and this is minimum-stakes territory, but in a 24-runner handicap hurdle, three independent signals pointing to the same horse is as much as we can ask for.
🎯SELECTIONS
KATEIRA
Race 4: 15:20 Glenfarclas Cross Country Chase
3m5f | Cross Country Chase | Good To Soft | 14 runners
Race Analysis
The Glenfarclas Cross Country Chase is unlike anything else at the festival. Run over Cheltenham’s specialist banks course — a winding, undulating circuit of ditches, banks, and unusual obstacles that has nothing in common with the conventional chase track — it rewards horses who know the course intimately and have the temperament to handle the demands. Speed figures and ability ratings still matter, but proven form over this specific course matters more here than in virtually any other race at the meeting.
Our TPR rating — the machine learning score that synthesises all available performance data into a single number — throws up an interesting quirk here. It places LATENIGHTPASS top of the field at 35, yet the market has that horse at 34/1. Under our rules, when the model and the market disagree that sharply, we trust the market — that pattern has produced just 2.3% winners across five years of festival data. LATENIGHTPASS is a pass. The model’s overall TPR can be unreliable in specialist races like this because it weights all form equally, when in reality cross country form and conventional chase form are almost different sports.
That context is important for understanding FAVORI DE CHAMPDOU. On TPR alone it ranks eighth in the field, which looks alarming at first glance. But here is what that number is failing to capture: on January 24th of this year, at this racecourse, over this exact cross country course, on Soft ground — conditions almost identical to today — FAVORI DE CHAMPDOU won in a canter. The form comment reads “prominent – led just after 3 out – pushed along and went clear before last – comfortably.” The class-adjusted speed figure from that run is 200. That is not just the best figure posted by any horse in today’s field — it is comfortably ahead of the next-best recent cross country figure, FINAL ORDERS’ T-1 of 164 from the same trial. The reason the TPR sits at rank 8 overall is that the model is averaging in a fall in December 2025 (FAVORI took a keen hold and came down at the 13th) and some weaker form from earlier in the season. Strip those out and focus on the two most recent cross country runs — a win at Leopardstown on Good to Soft in late December and then that dominant Cheltenham win in January — and the picture is transformed.
Think of the T-1 through T-5 figures as a horse’s last five exam scores. The most recent — T-1 — carries the most weight. FAVORI’s T-1 of 200 came from winning on this course in January. That is directly applicable to today in a way that almost no other piece of form in the race can match. Going credentials are also genuine: a GoingTPR of 97 ranks second in the field, confirming this is a horse that performs on Good to Soft ground rather than merely tolerating it. Gordon Elliott’s yard is running hot — the one-month trainer metric of 68 above the 12-month baseline of 67 — and Jack Kennedy in the saddle tells you the stable has maximum confidence in its chance.
The most credible rival is STUMPTOWN. The defending champion won this race at last year’s festival — T-2 of 148 from a Cheltenham Cross Country win in March 2025 — and also won here in December 2024 with a T-3 of 140. The course form is extraordinary: TrackTPR of 139, the best in the field, built on two dominant wins at this venue. But STUMPTOWN has not run for 150 days. The reason matters — a pulled up at Aintree in April that the trainer attributed to unsuitable ground combined with a vet finding lameness in the right hind. That is a significant injury flag, and returning after five months off without a prep run in a Grade-quality race over the most demanding course at the festival is a significant ask. If STUMPTOWN comes back to its very best it has the course form to be competitive, but there is a meaningful question over whether it can return to that level after the injury. The Gavin Cromwell yard is also running below its seasonal average right now, which adds to the concern.
FINAL ORDERS ran in the same January Cheltenham cross country trial that FAVORI won so impressively, and a T-1 of 164 tells you it was competitive — the comment puts it disputing second before the last before weakening run-in. That is solid recent course form. A T-2 of 127 from winning over this course in December 2025 adds to the credentials. At 13/2 it has a legitimate each-way case, particularly if you have concerns about backing FAVORI at 3.5. The same cold Cromwell yard applies, but the course form is undeniable.
FAKIR DOUDAIRIES deserves a mention because the career averages look good — GoingTPR of 124 and TrackTPR of 130, both ranking highly. But the recent form is a serious worry. The last two runs read: 4th at Punchestown on Soft to Heavy in February with “weakened gradually approaching last,” and 13th at Cheltenham in December with “always behind.” Two consecutive poor efforts on a horse with a high career average usually means the career peak is behind it. The T-3 of 176 from Clonmel back in November 2024 shows what FAKIR was once capable of, but the evidence from the last two runs suggests that horse no longer exists at those levels. Pass.
The selection is FAVORI DE CHAMPDOU. A T-1 of 200 from winning at this venue in January, on this course, on similar going — there is no more relevant piece of form in the entire field. The market agrees at 3.5, the yard is running hot, and the jockey booking confirms the stable’s intent. The conventional TPR rank of 8 reflects the model’s limitations in specialist races rather than any genuine weakness in the case.
🎯SELECTION
FAVORI DE CHAMPDOU
Race 5: 16:00 BetMGM Queen Mother Champion Chase (Grade 1)
2m | Grade 1 Chase | Good To Soft | 10 runners
Race Analysis
The Queen Mother Champion Chase is the race our data loves most. Grade 1 chases over two miles at this festival are where the convergence model has produced its most reliable results over five years — horses that top the ability rankings, have proven going and course form, and carry market confidence win at a rate that makes straightforward selections justified. Small, quality fields where class tells. This is exactly that.
FOUND A FIFTY leads our overall TPR standings but goes off at 22/1 — a significant gap between what the model sees and what the market thinks. That kind of disagreement has a near-zero win record in our dataset, and the most recent form comment explains why: the vet reported the horse was “blowing hard post race” after finishing fourth at Navan just 11 days ago. Pass without hesitation.
The two horses our filters lock onto are MAJBOROUGH and LEAU DU SUD — both qualify under our strongest convergence criteria. But one of those cases is significantly stronger than the other.
LEAU DU SUD has been off the track for 95 days, and the last run before that break told a straightforward story: at Sandown on Good to Soft in December, it “weakened after 2 out.” No interference, no jumping error, no ground excuse — the horse simply ran out of petrol in a race where IL ETAIT TEMPS won decisively. The course form at Cheltenham is genuine — a win here in November on Soft — but returning from three months off having weakened without excuse last time makes LEAU DU SUD difficult to trust at 5.5. The going credentials (GoingTPR of 135, second in the field) keep it technically in the picture, but the substance behind that average is thinner than the number implies.
MAJBOROUGH is the selection. The case is built on four months of consistent, improving form: a figure of 141 here at last year’s festival despite “bad mistakes at both the second last and last fence” according to the jockey — still finishing third having been in the lead; a win at Punchestown in May; third at Leopardstown in December where it was “a bit short of room run-in”; and then a dominant Leopardstown win in February on Soft where it “went well clear run-in – impressive.” Back-to-back figures of 161 and 160, the two most recent runs both clean performances. The going form rating of 143 leads the field — this is a horse that actively performs better in testing conditions rather than tolerating them. The Cheltenham course average of 147 (second in the field) is built on that festival run last year, which is directly applicable form. Mullins’ yard is running at its best level of the season — the one-month metric of 87 is comfortably above its 12-month average of 82. Mark Walsh is in the saddle. It goes off as odds-on.
The jumping does deserve a mention. Multiple runs carry references to jumping left and occasional mistakes. The Cork run in December included a “bad mistake 2 out – soon headed.” But the arc is clearly positive — the February win was described as impressive and clean. At six years old, this is a horse still sharpening its technique, and the trajectory of the form suggests it is doing exactly that.
The most interesting alternative in the race is IL ETAIT TEMPS at 7/2. The T-1 of 0 represents a fall at Ascot in January — the horse was going well in second when it “hung right and fell 2 out.” That is an unambiguous excuse. Before it: won Sandown, produced a figure of 258 winning at Clonmel in November — the standout raw speed figure in any recent run by any horse in this field — and won at Sandown and Punchestown before that. An unblemished record across completed starts. Paul Townend takes the ride, which is the second Mullins booking when Mark Walsh has MAJBOROUGH — that tells you the pecking order, but it also confirms how seriously the stable regards both horses. At 7/2, IL ETAIT TEMPS is a legitimate each-way play for those who want more price.
The selection is MAJBOROUGH. Two 160-plus figures in succession, best going credentials in the field, a dominant recent win, the hot Mullins yard, Walsh up. In a Grade 1 two-mile chase at Cheltenham, when all the signals converge on the same horse this clearly, the data says back it.
🎯SELECTION
MAJBOROUGH
Race 6: 16:40 Grand Annual Handicap Chase
2m | Premier Handicap Chase | Good To Soft | 20 runners
Race Analysis
The Grand Annual is a wide-open 20-runner handicap over the minimum trip, and like most big-field handicap chases at this festival, it resists the neat analysis that works so well in the graded races. Our data is direct on this: in races like this, the market is a better starting point than our ability rankings alone, and the best approach is to find horses where the data and the market are both pointing in the same direction.
That said, two horses in particular demand attention — and they represent very different kinds of interest.
The elephant in the room is RUBAUD, which our model rates second in the field and which sits at 14/1. A figure of 188 at Kempton in October, 178 at Kempton on Boxing Day, 175 at Wincanton on Heavy last month — those are genuinely elite numbers. The problem, and it is a real one, is that every single one of those runs came over hurdles. Today is a return to fences, and RUBAUD’s two completed chase starts — both from last season — tell a different story: “outpaced after 3 out – weakened run-in” at Punchestown in May, “weakened before last” at Aintree in April. The hurdle figures show a horse of real quality; the chasing record shows a horse that has yet to transfer that quality to the bigger obstacles. The market, at 14/1, is priced precisely on that uncertainty — and under our rules, when a highly-rated horse is drifting that far from the market’s top three, the historical record says pass. We hold that line, but at 14/1 the potential upside if the ability does translate makes it an understandable each-way flutter for those who want to take the risk.
The selection the data actually supports is VANDERPOEL — the model’s top-rated horse in this field. Where RUBAUD’s big numbers are over hurdles, VANDERPOEL has been building its case properly over fences: a clean win at Ascot on Soft in December, “in touch – headway to lead 2 out – soon clear – ridden out,” posting a figure of 161; followed by a win at Sandown in January on Good, “headway before 3 out – led before 2 out – ridden out,” figure of 148. Back-to-back chase wins, Ben Pauling’s yard running slightly above its seasonal average. At 7/1, it sits third in the market — that combination of model and market alignment is what we want to see in a handicap like this. It hasn’t been seen in public for 67 days, which is worth noting, but there is no injury excuse in the record — it simply hasn’t run since Sandown.
One word of caution on the going: VANDERPOEL’s one poor run this season came at Uttoxeter on Good to Soft in October — “jumped left on occasions – outpaced from 3 out – weakened after 2 out.” Today is Good to Soft, the same description. The win at Ascot came on Soft; the Sandown win on Good. That Uttoxeter run, combined with the jumping left notes, means this is not a certain horse on today’s ground — its best form has tended to come on slightly more testing conditions than Good to Soft offers.
JOUR DEVASION is worth knowing about at 11/1. Three chase wins in a row through autumn and winter, capped by a dominant Chepstow success in December where it “went further clear after 4 out – comfortably,” producing the best single recent figure of any horse in the field (232). The going credentials are exceptional — GoingTPR of 137, best in the entire field. The reason it sits outside our primary framework is that it ranks 12th in our overall ratings and has no Cheltenham form at all. In a 20-runner handicap at the festival’s most unique track, no course form is a meaningful gap. But the in-running chase form is the freshest and most consistent of any horse at a double-digit price.
BE AWARE heads the market at 4/1 but sits fifth in our ratings — a gap that would normally prompt caution. Avoid. RELIEVED OF DUTIES has genuine Cheltenham course form but its last two runs read “always behind” and “weakened quickly before last.” Whatever produced those performances, it cannot be relied upon here.
The selection is VANDERPOEL — top-rated on the model, two clean chase wins, market respect at 7/1, Pauling yard in decent form. The each-way case on RUBAUD is for those who believe the hurdle figures will transfer to fences, with the understanding that today is very much an open question on that front.
🎯SELECTION
VANDERPOEL / RUBAUD
Race 7: 17:20 Weatherbys Champion Bumper
2m | NH Flat | Good To Soft | 22 runners
Race Analysis
The Champion Bumper is the race where our standard analytical toolkit has to work hardest. With most runners having just one or two starts to their name, the form book offers almost nothing to separate them. What it does offer is breeding — and this is one area where EquiAnalytix goes considerably deeper than most.
Every sire and damsire in our database carries a SireTypeTPR figure — a measure of how well that bloodline’s offspring have historically performed in this specific race type. For NH Flat races in particular, where individual run data is thin, that number becomes one of the most informative signals available. We know which sire lines reliably produce horses built for the unique demands of a bumper: the flat pace, the stamina test, the absence of obstacles to hide behind. The critical word there is reliably — a SireTypeTPR figure is only meaningful when it is built from multiple runners across multiple seasons. A sire with one bumper runner simply reflects that runner’s own debut figure back at you, which tells you nothing you didn’t already know. We apply that filter rigorously here.
The race is led in the market by LOVE SIGN DAUNOU at 9/2 — Mullins trained, Patrick Mullins aboard. A debut win at Naas on Heavy in January, “made all – went well clear from 2f out – canter,” is as smooth a first run as you will see. The Goliath Du Berlais sire line carries a bumper-specific breeding figure of 93 built on genuine multi-runner evidence — a proper NH specialist sire. This is a legitimate contender and the Patrick Mullins booking typically signals the stable’s most fancied runner.
THE MOURNE RAMBLER at 17/2 cannot be dismissed lightly. Colin Keane taking a cross-channel booking in a festival bumper is a confident stable signal, and the T-1 of 164 from a Leopardstown win on Good to Soft in December — “midfield – good headway – went second 1f out – ridden to lead final 110yds – won going away” — is the highest debut figure in the field by some distance. The sire Well Chosen carries a bumper breeding figure of 66 which is modest, but the performance on the day was not modest at all.
The selection is KEEP HIM COMPANY. A T-1 of 174 posted at Leopardstown on Good to Soft in December is the strongest recent figure in the race, producing a GoingTPR of 117 that most rivals cannot match. The Walzertakt sire line carries a bumper-specific breeding figure of 86, built on multiple runners — genuine independent evidence of an NH Flat specialist sire. Elliott’s yard is running marginally hot at 68 against a 12-month baseline of 67, and Jack Kennedy takes the ride. Form, breeding, and jockey all point in the same direction. One significant caveat: the TrackTPR is zero — KEEP HIM COMPANY has no Cheltenham form at all, having won at Fairyhouse and Leopardstown. In a race where certainty is impossible, the combination of the field’s best recent figure and strong sire credentials is as good as the data allows, but the lack of course form is a genuine unknown.
The each-way case sits with QUIRYN at 8/1. The Sottsass sire line carries a bumper-specific breeding figure of 108 — the highest of any sire in this field, above even Goliath Du Berlais. Sottsass is a newer sire transitioning from the Flat to jumps, so that figure deserves some caution on sample size, but it cannot be dismissed. What backs it up is the dam Quezon Sun with a progeny average of 143, the strongest dam evidence in the field. The debut win at Naas on Soft in January was done well within the horse — “travelled strongly – led over 2f out – ridden briefly and went clear inside final furlong – readily.” And then there is Paul Townend. Mullins runs four horses in this race. The world’s best jumps jockey is on QUIRYN, and in a bumper it carries weight. At 8/1 each-way the place component alone makes this a worthwhile supporting play.
🎯SELECTION
KEEP HIM COMPANY / QUIRYN
📋 Day 2 Summary: Selections at a Glance
| Race | Time | Selection |
|---|---|---|
| 1 | 13:20 | NO DRAMA THIS END (12) |
| 2 | 14:00 | ROMEO COOLIO (10) |
| 3 | 14:40 | KATEIRA (14) |
| 4 | 15:20 | FAVORI DE CHAMPDOU (2) |
| 5 | 16:00 | MAJBOROUGH (8) |
| 6 | 16:40 | VANDERPOEL (9) / RUBAUD (3) |
| 7 | 17:20 | QUIRYN (21) / KEEP HIM COMPANY (8) |
Conclusion
That wraps up our Day 2 coverage of the Cheltenham Festival 2026. Wednesday’s card is headlined by MAJBOROUGH in the Queen Mother Champion Chase — our highest-confidence selection of the day, a Filter E qualifier in the festival’s most profitable race type. ROMEO COOLIO in the Brown Advisory and the KEEP HIM COMPANY / QUIRYN combination in the Champion Bumper complete the core plays. The double of ROMEO COOLIO and MAJBOROUGH is where the day’s structure sits.
Day 1 gave us three winners including the mighty Lossiemouth sprinting clear up the hill in the Champion Hurdle. The model is in good shape. Now we find out if it can back that up.
Through our partnership with Tote UK, we will be back tomorrow for Day 3, with the Ryanair Chase, the Stayers’ Hurdle, and the Pertemps Final among the headline acts. The festival rolls on.
Get the Full Data in Our App
Access every racecard, live analytics and the complete blog on your phone: