← Blog · Results · 24 August 2026 · 6 min read
BMW Championship and Nexo Championship 2026: how our picks actually finished
Eight selections published on 17 August, three days before the first round. One of them placed. Here is every finish, unedited, with two disclosures about the data behind them.
By Doug Dinwiddie. Founder of ProPlace. Former DP World Tour caddie.
Both events finished on Sunday 23 August. ProPlace published four selections in each before a ball was struck, eight in total, and one of the eight placed. It was a losing week on a flat each-way stake.
Below is every one of those eight, beside the finish it produced. Nothing has been edited, reordered or removed.
ProPlace is a statistical research product, not a tipster service and not a bookmaker. 18+ | BeGambleAware.org
The receipts
Every selection in both tables was published on 17 August 2026 at 18:00 UTC, three days before the first round. Nothing below has been edited since it went out.
BMW Championship, PGA Tour, 20 to 23 August 2026
| Player | Conviction (out of 15) | Model place probability | Market implied place probability | Finish | Placed |
|---|---|---|---|---|---|
| Xander Schauffele | 14 | 38.2% | 21.7% | 26th | No |
| Ludvig Aberg | 14 | 37.7% | 21.7% | 16th | No |
| Cameron Young | 13 | 33.2% | 20.0% | 37th | No |
| Si Woo Kim | 13 | 30.0% | 16.7% | 33rd | No |
Nought from four placed.
Nexo Championship, DP World Tour, 20 to 23 August 2026
| Player | Conviction (out of 15) | Model place probability | Market implied place probability | Finish | Placed |
|---|---|---|---|---|---|
| Nathan Kimsey | 12 | 19.7% | 9.1% | 30th | No |
| Marcus Armitage | 10 | 20.5% | 12.5% | 38th | No |
| Ewen Ferguson | 9 | 26.3% | 20.0% | 24th | No |
| Matthew Jordan | 9 | 21.6% | 14.3% | 2nd | PLACED |
One from four placed.
To a flat ten pound each-way stake the eight selections returned minus ninety pounds across the two events. Matthew Jordan finishing second was the only return.
Past performance is not a guide to future results.
What the week says, and what it does not
Our two highest conviction picks of the week, Schauffele and Aberg, both finished inside the top 26 and neither placed. In a small no-cut field, finishing well and placing are very different things, and this week is a clean illustration of that. If that distinction is new to you, what a no-cut event changes about a week covers it in more detail.
The one that paid was one of the lowest conviction picks on the board. Jordan was a 9 out of 15. That is not a point in the model's favour and we are not going to present it as one. Over eight selections in one week, conviction ordering tells you nothing.
Eight selections is also far too small a sample to conclude anything about the model, good or bad. A week is noise. We write these up because the record only means something if the bad weeks sit in it alongside everything else, not because a single card settles an argument in either direction.
The full archive of published selections is on the results page, and the event by event breakdown sits on performance.
Two disclosures about the data behind this page
Neither of these changes a finish above. Both change how much weight the numbers beside those finishes deserve, so they belong here rather than in a footnote.
One, the prices are estimated. Every price behind this week's eight selections is model-estimated rather than a captured bookmaker price. Our records mark them as estimated, with the source recorded as a simulated outrights model and the verification flag not set. The market implied place probabilities in the tables above are therefore our best reconstruction of the market, not a quoted price you could have taken. If you want the mechanics of that conversion, the implied probability definition sets it out.
Two, the stored place terms are wrong. Our stored place terms for both events read eight places at one fifth the win odds. For a small no-cut field that is too generous a reading and we believe it is an error in our records. The honest consequence is straightforward: because no BMW Championship selection finished inside eight places, and the one place we recorded finished second, no result on this page changes under stricter terms. The metadata is wrong. The settled outcomes are not. We are correcting the terms and flagging it here rather than quietly fixing it. The place terms entry explains what those fractions do to a return.
If you want to look at the work
A complete card from the current field is published on the free card page, and every published selection, this week's included, stays on the record.
Doug Dinwiddie. Founder of ProPlace. Former DP World Tour caddie.
Past performance is not a guide to future results. ProPlace is a statistical research product, not a tipster service and not a bookmaker. 18+ | BeGambleAware.org
Frequently asked questions
- Did ProPlace pick a winner at the BMW Championship 2026?
- No. ProPlace published four selections for the BMW Championship on 17 August 2026 and none of them won or placed. The four finished 16th, 26th, 33rd and 37th.
- How many ProPlace selections placed in the week of 20 August 2026?
- One of eight. Matthew Jordan finished second at the Nexo Championship. The other seven selections across the two events did not place, so the week was a loss on a flat each-way stake.
- Were ProPlace's BMW and Nexo picks published before play?
- Yes. All eight were published on 17 August 2026 at 18:00 UTC, three days before the first round on 20 August, and none of them has been edited since.
- Where can I see every ProPlace selection, including the losing ones?
- The full archive sits on the results page at proplace.golf/results, with the event by event breakdown on the performance page. Losing weeks stay published in the same place as the rest.