Another week, another review, and this time it is WinnerOdds turn. Winner Odds is a Value Betting software, but it is somewhat special. It identifies Value Bets in tennis with the help of an AI algorithm and delivers the tips to the end customer on a regular basis. WinnerOdds quotes a historical ROI from all advised picks at around 9% with a decent volume of monthly delivered picks:
Sounds good, so where’s the catch? Dive with me into my review of WinnerOdds and find out.
WinnerOdds is on the market since 2016. The man behind WinnerOdds is Miguel Figueres, who operates the service from Valencia, Spain and is, according to his own Twitter description:
Engineer of roads, canals and ports. Associate professor of the UPV. Specialized in Neural Networks and Artificial Intelligence. CEO and creator of @WinnerOdds
(Google Translate from Spanish; I am being told “Engineer of roads, canals and ports” is just Google speak for civil engineer)
Actually, Miguel contacted me some time ago asking if I would be interested in reviewing the product. I had a glance at the website, found the idea interesting and decided to give it a go. I got a trial access in order to try the software and place some bets and I was all set.
Discussing the matter further, we decided to enter into an affiliate agreement, meaning I would get a share from every software purchase you make if you are being redirected from my website. Regardless, I would do my best to deliver an honest review and point out all the benefits and drawbacks of using WinnerOdds as compared to alternative services.
What is WinnerOdds all about?
WinnerOdds is a very straightforward kind of service. It delivers Value Bets for tennis and tennis only. This in itself sounds familiar, as there are already a couple of players on the market, such as Trademate Sports, RebelBetting and BetBurger doing just that, and for a lot of sports too. So how does WinnerOdds distinguish itself from those services?
As it turns out, WinnerOdds has sort of a special method in identifying Value Bets. According to their website:
“The Artificial Intelligence algorithm used by WinnerOdds estimates the real probabilities of the tennis matches even before the bookmakers publish their odds.”
That sounds fancy. But frankly, I was a little skeptical when I first read that. Coming from the world of online sports betting, one has seen all kinds of wild promises and advertising slogans. What if that is just another Technical Value Betting software only making it sound like they do things differently? So I decided to dive a bit deeper into the methodology before I start collecting bets.
How does it work?
Now, we should have a realistic expectation as to how far would WinnerOdds reveal their methodology, as this should be expected to be the products main trade secret. However, we get a few useful insights. In particular, if you have a look at the How It Works page, you find out that WinnerOdds is running an AI method called Neural Networks on a database of 400,000 tennis matches. Using the dataset the algorithm updates itself dynamically and learns to calculate winning probabilities ever more precisely.
The winning probabilities are turned into odds and compared to the market to identify Value Bets. Interestingly, this happens before the market for a match is up, so at the moment the odds are published, if there is a discrepancy, you should expect a value bet to appear.
Furthermore, the algorithm takes into account the line movement in order to identify profitable trends. Now, this second part sounds an awful lot like a classical Technical Value Betting, so before placing my first bet, I decided to make a check. Is it just a ‘soft books being too slow to adjust their line’ kind of thing?
Putting a bet to the test
To answer this, I took the first advised Value Bet and it was offering value in several soft books. Next, I had a look at my sharp book’s accounts and Betfair to compare. Surprisingly, I found out the game wasn’t on offer in any of them. Clearly, WinnerOdds wasn’t comparing with the sharp’s line. This become all the more clear since among the next bets I found out some of them were apparently offering value in the sharp books and exchanges as well.
So this was definitely something else. WinnerOdds does seem to employ an original methodology to calculate the correct odds, resulting in value bets you would be unlikely to find in any other Value Betting tool.
This is an important point. At the end of the day, that is the main advantage of WinnerOdds to (somewhat) comparable services. That and the higher yield of 9.17%, which is to my knowledge currently unmatched on the market for this kind of service. Of course, the 9.17% yield promise was something that would not go untested.
That is when I started collecting my bets.
My WinnerOdds bets
I have done paper trading, but made sure to manually check the bet availability at the advised books every now and then.
I have not had a case of a bet that has disappeared at the moment I went to the bookmakers. In fact, the bets appeared to survive much longer as compared to a soft book bet forming an arbitrage pair with a sharp book or an exchange. This only made sense, since the bets were only available at WinnerOdds. This certainly wouldn’t have been the case with the value leg of an arbitrage pair, since such pairs are identified by a number of software tools, the above mentioned RebelBetting, BetBurger and Trademate Sports being only a few of them.
On the other hand, I found a relatively lower amount of total bets I could place as compared to a regular Arbitrage / Technical Value Betting service. This should not be a surprise, since WinnerOdds focuses on tennis only. This also means you total number of tips would be strongly dependent on the tournament schedule for the year.
Now, I have managed to collect a little more than 200 bets for the length of my trial, but I would advise you not to use that as an indication. I must admit I was not terribly efficient at marking my bets in WinnerOdds, for reasons that probably would not interest you too much. The number WinnerOdds quotes is around 1000 bets per month, which of course will be heavily dependent on the time of the year, as already explained.
Realistically you would only manage to catch a portion of those, since even though the bets last longer, after they attract a certain volume they will nevertheless disappear.
All in all, the bet availability was lower than in most tools, but satisfactory.
Yield / ROI
I had chosen a starting bank of 5000 units. I have achieved a yield of 14.98% with 222 pics. The stats look like this:
Now, of course we are not going to get too excited about that yield. From such a low number of bets it does not tell you that much. However, looking at some of the reviews WinnerOdds has already received from blogger colleagues, we get:
… and the list goes on. Given the above, combined with my own experience, I do believe the historical quoted yield of 8.98% is achievable.
Moreover, WinnerOdds are pretty transparent about their previous picks. You can download a .csv file with a complete pick history since the inception of the service and double check its content yourself.
Historical yield vs User historical yield
An important note should be made about the fact, that the user history shows a yield of only 5.3% since inception of the software. These are the records of the bets that users of the service actually marked as placed. That is, of course, a significant difference from the overall historical yield and from the average return recorded by the few reviews out there (including mine).
The reason for this is clear. When WinnerOdds records historical yield, or when I or other reviewer do paper trading, we always take the highest price on offer. Not the case with actual users, who don’t have access to all the covered bookmakers, because the one or the other does not accept customers from their jurisdiction or has limited them already.
What can you do?
Therefore, the fewer accounts you have, the less would you be able to take the best price on an advised pick. At the end of the day, whether you would be closer to the general average or the historical user average depends strongly on the number of accounts you have at your disposal. Presumably, if you bet from a punter-friendly country such as the UK and have all your accounts in check, you would achieve a better yield (and turnover) than if you bet from somewhere else. Furthermore, Spanish houses seem to be well-represented in the list of supported bookmakers, so it should be easier for you if you can get down there too.
However, in case you only have a few of the covered bookmakers at your disposal, you will achieve lower yield and turnover. This is certainly something to keep in mind, which doesn’t have much to do with the service, but is more of a general smart punter’s issue.
You can find the full list of supported bookmakers here.
Where does the yield come from?
I have broken down the achieved yield by tournament, the presumption being that smaller tournaments would offer higher yield. I am getting the following results:
Indeed the presumed effect is in place. The ATP tournament returns around 4% yield at best odds. ITF and Challenger meanwhile give you more than 9% ROI and a larger number of picks. Women’s tennis offers a reasonable 7.18%.
So, as you see WinnerOdds extracts more value from smaller competitions, although it is still profitable at large ones. This is not unusual – large markets tend to be more efficient than small ones. That also means you will have smaller limits on the picks that give you a higher edge, but that is just part of the punter’s life.
It is a Value Betting strategy, so the average caveats apply. The volatility is serious and you might lose a significant portion of your bank before turning a profit. Having a reasonable staking plan and being psychologically prepared for the inevitable drawdowns should help you with that.
Features & bet placement
The website has a nice and user-friendly design. In the screenshot above you see a number of neat stats that help you analyze the performance of your bets better. In the ‘Results’ section you could also check how your bets were graded:
You get the option to hide picks and you receive suggestions on the amount to stake. The suggestions are based on a fractional Kelly criterion with f = 0.125, but the ‘to win’ amount is capped at 2% of your bankroll and there is also a bottom limit set at 0.75%. The exact formula is:
Bet size = Max (0.75 / (Odds – 1); Min (2 / (Odds – 1); (0.5 + 0.125 * (Odds / MPO – 1) * 100) / (Odds – 1)))
You could also turn on notifications for new bets, which arrive to you per your registered e-mail.
Soft and Sharp Books
Now, I had some concerns finding out my first bet is advised in a soft book, since we are all aware that soft books are chasing out winners the moment they identify them. Those concerns were somewhat comforted after I found out the service identifies value bets at sharp books too. In my sample, around 10% of the bets have been made in sharp books and exchanges.
How often do Sharps offer the best odds?
The complete pick record shows that in only 3.73% of the advised picks the two sharp books (Pinnacle and SBOBet) were offering the highest odds with an aggregated advised stake of 2.68% of the all bookmaker total. The difference between those figures probably comes from the fact that the identified edge in sharps and exchanges was slightly lower compared to other books, therefore the advised stake there was slightly smaller too, which is to be expected.
A total of 157 picks with have been placed ar SBOBet with an aggregated advised stake of 217,94 units for a total profit of 4,603 units or a yield of +2,11%. 785 picks with Pinnacle were advised at a total stake of 1622,506 units for a profit of 16,155 units or an ROI of +1,00%. For those of you with preference for tables:
This is of course a rather low number and indicates that, if you want to get the best odds possible on the advised picks, in most cases you will have to place your bet at a soft book. Also, those records alone are inconclusive for the profitability of the service in sharps.
Is the service EV+ in sharps as well?
However, that shouldn’t trick you into thinking you can only place that many bets with Pinnacle and SBOBet. Rather, these were the cases where those two were offering the best price of all bookmakers. If we had a look at all the cases where the sharps were not offering the best odds, but were still above the minimum advised odds, we would get a much longer series and a clearer picture. Luckily, WinnerOdds has such records too – sadly, just Pinnacle and no SBOBet, but Pinnacle has way more bets anyway – and have sent it to me upon request. Let’s get to work!
All Pinnacle picks with value odds
Indeed, looking at all the picks where Pinnacle provided any value (and not just the highest value of all books), we get a much more telling total of 3076 bets. The yield of those is… *drum roll*… 0.79%! So we still have a slightly positive ROI but these time on a much longer series of bets. Now, the yield is still rather low so one might argue the results are still inconclusive. Before getting into technical details here, I’d rather answer one much more important question, namely:
Does WinnerOdds beat the Pinnacle Closing Line?
I have proceeded as follows: for every game where a Pinnacle bet was advised, the Pinnacle Closing Odds. These closing odds were adjusted for the margin at the time of placing the bet, giving the True Pinnacle Closing Odds:
Pinnacle Closing Odds * Margin = True Pinnacle Closing Odds
Finally, the Advised Odds at time of the bet were divided by the True Pinnacle Closing Odds and 1 was subtracted to find out the Expected Value for this bet at unit stakes.
Bet EV @ Unit Stakes = Pinnacle Advised Odds / True Pinnacle Closing Odds - 1
However, WinnerOdds is advising different stakes according to the size of the edge identify. To account for this, I calculate second series of Expecting Value:
Bet EV @ Advised Stakes = Bet EV @ Unit Stakes * Advised Stake
Finally, I plot the EV of the whole series at Unit Stakes and at Advised Stakes next to the actually achieved results.
What is the outcome?
The resulting graph (with a starting bank of 100 units) for the complete series looks like this:
- Number of picks = 3076
- Yield = 0.79%
- EV Yield = 0.78%
You can see the orange line staying in the +EV zone for most of the sample, meaning for the average advised bet in Pinnacle WinnerOdds beats the margin-adjusted Pinnacle Closing line.
Furthermore, you see the actual profit moves more or less in line with the grey line, which itself is above the orange one. That is the EV multiplied by the advised stakes, meaning WinnerOdds also correctly identifies the size of the edge it has over Pinnacle.
How does it look like in SBOBet?
WinnerOdds have delivered the same stats for SBOBet as well and they look… a bit stranger. In SBOBet the service fails to beat the margin adjusted closing line, but delivers a much higher yield! The results in SBOBet look as follows:
- Number of picks = 1460
- Yield = 3.98%
- EV Yield = -1.05%
We would expect the real yield to move more or less in line with the EV yield and they deviate quite a bit in the SBOBet sample. What could the reasons be?
For one, we are looking at much smaller sample (more than two times) as compared to Pinnacle, so luck has a larger role to play in this sample as compared to Pinnacle. Second, the margin in SBOBet is significantly higher compared to Pinnacle’s. The average overround in the Pinnacle sample is 4.6%. In SBOBet you are looking at 7.04%.
Of course, adding this higher overround to the closing odds makes them harder to beat,, but there is another side to that story. Bookmakers normally assign higher overround to events where they are less sure about the offered odds. Since they expect to lose more money due to the not so sharp odds, they compensate for that with a higher margin. Obviously, SBOBet have less confidence for their odds in those games than Pinnacle. In fact, if we compare the advised odds in WinnerOdds to the non-margin-adjusted closing prices in SBOBet we see a clear trend:
Of course, the orange line is no realistic yield expectation either, but I think it still has a story to tell. The fact that WinnerOdds consistently beats the non-margin-adjusted closing odds shows some inefficiency there. SBO compensates a part of it with the high margin, but not all. This is how I would explain the results of the service in SBOBet and the divergence between the actual yield and the EV yield there. If anyone has another ideas I am looking forward to hear them.
Can you really play those?
How accessible are the sharp picks analysed above?
You have the following summary stats in Pinnacle:
- Number of picks = 3076
- Number of picks / day= 3.87
- Number of picks / month = 116.22
And for SBO you have:
- Number of picks = 1460
- Number of picks / day = 1.84
- Number of picks / month = 55.16
- Yield = 3.98%
- EV Yield = -1.05%
A total of 5.71 picks per day with the two sharp bookmakers are not too many but not too few either.
Do the odds last?
How long do they last? Well, to be honest, even though I tested the service paper trading for the most time, once I found out it was beating the Pinnacle closing line I started playing with some real money. For the games I played, I always managed to find the advised odds and I only had a look at the service once or twice a day, so I don’t know how long the odds were available before I saw them.
Compared to tipster services I followed in the past, where the odds are dropping the second they were out, the sharp odds availability in WinnerOdds is way better than anything I have seen. Of course, the relatively low yield also has a part to play in this.
Can you pay your software fees with that?
If you only place the bets in Pinnacle and manage to place all of them, you would have paid around €2100 in software fees (considering you use no discount) for the time the record above has been collected. So to break even, you would need a bank of around €3500. Anything above that should deliver net profit for you.
Of course, that calculation does not include the bets in SBO, nor in all the soft books, the second of which are much more in count and yield. On the other hand, it relies on the pretty optimistic assumption that you will manage to place all bets. So there are a few things to take into account, but I overall if you have a decent bankroll you would certainly benefit from the Pinnacle & SBO picks that WinnerOdds is offering.
Speaking of software fees, let’s shortly have a look at the…
The product is currently priced at €79/month. With the biannual package you get a discount of around 20% and arrive at a price of €65/month.
TThe service only offers tennis and therefore the total number of picks is comparably lower than other services, while the price is within the range of the likes of BetBurger, RebelBetting or Trademate Sports. Furthermore, looking at past reviews one finds out there was the occasional price hike throughout the years. However, you should be careful to avoid…
Comparing apples to oranges
Once you find out a bit more about the service, you find out it’s approach to finding value bets differentiates it from other market players. You are getting original picks identified by an AI engine which, although shared with all the other customers of WinnerOdds, are not seen by customers of any other software out there. This certainly means the value lasts longer (as I have seen myself) and potentially, that your accounts placing those bets would last longer too (not so sure about that and haven’t tested it, it is just my assumption).
You can easily take a liquidity-following Technnical Value Betting software such as Trademate Sports, RebelBetting or BetBurger and combine it with WinnerOdds and you will get two completely different sets of picks, so I would argue WinnerOdds is a different category altogether.
Importantly, the picks at sharp books are a very valuable addition to the service. For some punters they alone would justify the service fee. In any case, they are a testimony for the work of the WinnerOdds team. It is extremely rare to find a service beating Pinnacle’s closing line, even by a small margin.
30-day profit guarantee
As an additional incentive, WinnerOdds offers a 30-day money back guarantee in case you do not make a profit with WinnerOdds within your first month of betting. The condition is that you place at least 200 bets within the first month (or 100 bets for subscriptions made during the months of November, December or January).
Additionally, if your initial bankroll was over € 1000 and the profit is lower than the cost of the subscription, WinnerOdds will reimburse the difference up to a maximum of the cost of the subscription.
You can get the full conditions of the money back guarantee here.
So, what’s the verdict?
- Original picks: The biggest advantage of WinnerOdds is that you get original picks identified with the help of a proprietary AI algorithm identifying value.
- Picks with the sharps: Furthermore, the service delivers picks with sharp bookmakers, which fare well both in real and EV terms (beating the Pinnacle Closing Line)
- Longevity: Both those factors contribute to longer lasting value bets on the one hand & (potentially) longer lasting accounts on the other.
- Finally, you get a very transparent track record and a neat user-friendly interface
- Premium Pricing: The service comes at a premium pricing, while only delivering picks for tennis, which results in a lower total volume of picks as compared to other Value Betting services out there
- Higher price per pick: As a direct result of the above you end up paying a higher price per pick.
To compensate for the disadvantages and make full use of the benefits, you must carefully consider if you would be able to make full use of the advised picks. The more you are available at the time the relevant tournaments are being played, the higher volume you will be able to place, but I guess that is no news for you. Furthermore, you must evaluate how many of the scanned bookmakers you can access.
If you can tick those boxes, the profit from your WinnerOdds‘ picks should more than offset the software fee and contribute to a good return on your investment. Combining WinnerOdds with another Technical Value Betting software might also be a good idea, since you would have two alternative streams of value bets with different selections, but is by no means necessary. Finally, the software is perfect for the high-roller, with sharp picks that beat the Pinnacle closing line.
Thanks for reading and if you have any questions feel free to drop a comment below. I would answer those if I can and if not, I would forward them to Miguel. I hope you would enjoy the review while I’m enjoying Rock im Park this weekend – if anyone is around and wants to meet for a beer just let me know. Stay tuned for more and see you around!