I have been using tennis trader and I've noticed some quite alarming discrepancies between the predicted odds (for the ends of games) and the actual odds.
For example, I've just been looking at the Gavirilova v Goerges WTA Wuhan match. At 4-4 in the final set, the predicted odds, if Goerges held serve to go 5-4 up, were 2.07 - this seemed inaccurate as a forecast to me as it seemed likely that she would go odds-on. She won the game and indeed her actual price went to 1.78.
I had calibrated before the start of the service game and the match type was set correctly, so this discrepancy is huge in betting terms and could lead to big losses if people are taking such predictions as gospel. It has been happening frequently in the ATP as well so the erratic serve statistics in the WTA don't explain it. It occurs throughout matches for a variety of in-play scenarios.
This needs to be looked at - I am not convinced the predictive algorythms employed in Tennis Trader are still accurate. Perhaps the markets have changed dramatically since its inception (due to in play trading becoming so popular) and to such an extent that Tennis Trader is no longer reliable.
Tennis trader inaccurate
The model is as good as it gets without making it completely unusable. It simply takes the chance of winning a point on serve based on user input or the current market and projects it forward. The only flaw with a model like this is that it assumes that the chance remains constantly which it rarely does. So it can only be used as a guide. If you want it to be more accurate then you can seed the model with your own values so it derives odds from there.
https://www.betangel.com/user-guide/cal ... 3D&mw=MzIw
If you are looking at the final game in a match then I'd imagine the chance of closing out the game will be subject to many, many factors. So it's least likely to be accurate at the point unless you specifically take into account everything available at that critical point in the match.
https://www.betangel.com/user-guide/cal ... 3D&mw=MzIw
If you are looking at the final game in a match then I'd imagine the chance of closing out the game will be subject to many, many factors. So it's least likely to be accurate at the point unless you specifically take into account everything available at that critical point in the match.
Cheers Euler. I realised the liquidity wasn't great on that match but it also happens frequently on high liquidity matches, at numerous points through the matches. It's especially erratic in the womens game (understandably when you view the stats). I take your points though - but when I watched the guidance videos I got the impression is was far more accurate than it is - if you're only expecting a 2 tick movement and it moves over 10, you really could get your fingers burnt - maybe the guidance videos for beginners should highlight this.
Don't forget also that ticks don't accurately reflect % chance so a move can be quite shallow in % terms but wide in ticks.
Models are perfect, but the reality isn't unfortunately. There will always be a bit of judgement, especially in that last set.
Models are perfect, but the reality isn't unfortunately. There will always be a bit of judgement, especially in that last set.
- breakbacktennis
- Posts: 90
- Joined: Mon Aug 28, 2017 11:11 am
For what its worth I always have the tennis trader window up on one screen and simply allow a margin of error when I enter a trade. Its a fantastic tool to use as a guide.
-
- Posts: 165
- Joined: Wed Mar 01, 2017 2:06 pm
I've recently begun using the tennis trader, and I mainly use it for the Game Matrix, but it is nice to see "what if" scenarios in case of a break of serve leading to a set win.
I have noticed that the odds are sometimes drastically different.
I've watched the video on configuring it with a seed value, or giving individual players certain values. I'm just not sure on how to come by those values. My first guess is that it comes from years of experience with trading tennis? Or is there some sort of rule of thumb, ranking? Need for a win? etc..
I have noticed that the odds are sometimes drastically different.
I've watched the video on configuring it with a seed value, or giving individual players certain values. I'm just not sure on how to come by those values. My first guess is that it comes from years of experience with trading tennis? Or is there some sort of rule of thumb, ranking? Need for a win? etc..