Soccer Predictions using Python (part 4) by stevencurrie

View this thread on: hive.blog | peakd.com | ecency.com

programming · @stevencurrie · Sep 23 '17

$0.02

Soccer Predictions using Python (part 4)

<html>
<p>Evening sports fans. &nbsp;Hope everyone's having a wonderful weekend.</p>
<p>Before we get to the code, I'm happy to say that 7 out of 10 predictions were correct and the 3 that were wrong were draws!</p>
<p>If we had put £1 single bets on each game, then for our £10 stake, we'd have had £12.86 back. &nbsp;Only time will tell if this 28.6% ROI will continue.</p>
<p>In <a href="https://steemit.com/programming/@stevencurrie/soccer-predictions-using-python-part-3">Part 3</a>, I spoke about limiting how far back the system would look when making it's predictions and chose 100 games as a default limit. &nbsp;I've now added a function to backtest different values for this.</p>
<p>The updated code is available on my <a href="https://github.com/steviecurrie/soccer-prediction">GitHub</a>.</p>
<p>What I've done is take 60 days of games from a year before the current date and backtest with values from 50 to 500 games, outputting the most successful value.</p>
<p>I've also added a cutoff value for the predicted probability to decide if the game is worth betting on. &nbsp;So the code also sweeps through values for this from 40 to 95.</p>
<p>There was a problem with this approach initially in that it would get to 100% accuracy but only suggest betting on 1 game out of 100. &nbsp;In other words only games that were pretty much foregone conclusions and therefore not worth betting on.</p>
<p>So I've now limited this to advise of at least 1 game out of 10. &nbsp;It reports somewhere in the region of 70-90% accuracy during the backtest.</p>
<p>Now this is a pretty naive form of machine learning, basically a brute force scan through what could be called our hyperparameters, so there's likely to be a danger of curve fitting. &nbsp;To rule this out, I also added a function to test the parameters found during the scan on the next 60 days of games. &nbsp;If the reported accuracy still looks good then we're golden.</p>
<p>New command line options are "-t" or "--test" to scan through the values, and "-b" or "--cutoff" to have the program print out predictions with predicted probabilities above that value.</p>
<p>Running the following command line will find the best values to use for the Scottish Premiership.</p>
<p><code>python3 soccerprediction.py -c Scotland -l Premiership -t</code></p>
<p>This returns with values of 450 for history and 70 for cutoff with 100% accuracy for 7 predictions out of 70 games. &nbsp;Sounds too good to be true, I know! &nbsp;However, it also returns 100% accuracy in the validation test.</p>
<p>Running the tests on the English Premier League returns 400 &amp; 70 with 71% accuracy for 7 games from 70. &nbsp;The validation test returns 93%.</p>
<p>I've only tested with the English Premier League, English Championship and the Scottish Premiership so far but as the predictions the code made in <a href="https://steemit.com/programming/@stevencurrie/soccer-predictions-using-python-part-3a">Part 3a</a> show, it appears to be working pretty well.</p>
<p>Hey, I know the code isn't pretty, efficient, elegant or any of the things it would be if a professional programmer had written it but who the hell cares if it works eh? &nbsp;I'll be continuing to test it and hope some of you guys give it a try too. &nbsp;Feel free to use or change the code in any way you want and if you've any ideas for improvements or fixes please share them here. &nbsp;</p>
<p>Maybe we can all stick it to the bookies. hehehehe.</p>
<p><br></p>
</html>

👍 digitalking, precise, stevencurrie, ubg, minnowpond1, dkabii

`author`	stevencurrie
`permlink`	soccer-predictions-using-python-part-4
`category`	programming
`json_metadata`	{"tags":["programming","python","soccer","predictions"],"links":["https://steemit.com/programming/@stevencurrie/soccer-predictions-using-python-part-3","https://github.com/steviecurrie/soccer-prediction","https://steemit.com/programming/@stevencurrie/soccer-predictions-using-python-part-3a"],"app":"steemit/0.1","format":"html"}
`created`	2017-09-23 20:33:09
`last_update`	2017-09-23 20:33:09
`depth`	0
`children`	13
`last_payout`	2017-09-30 20:33:09
`cashout_time`	1969-12-31 23:59:59
`total_payout_value`	0.020 HBD
`curator_payout_value`	0.001 HBD
`pending_payout_value`	0.000 HBD
`promoted`	0.000 HBD
`body_length`	3,488
`author_reputation`	7,132,661,654
`root_title`	"Soccer Predictions using Python (part 4)"
`beneficiaries`	`[]`
`max_accepted_payout`	1,000,000.000 HBD
`percent_hbd`	10,000
`post_id`	15,738,974
`net_rshares`	6,789,601,814
`author_curate_reward`	""

properties (23)vote details (6)

voter	rshares	pct
precise	2,074,846,048	100%
ubg	219,765,453	1%
digitalking	3,402,948,466	100%
minnowpond1	147,094,329	1%
stevencurrie	944,947,518	100%
dkabii	0	100%

@outcast81 · Sep 23 '17

Hi its me again :)
What data goes into the prediction? You said the system looks back 100 games to make the predictions, just the result? Bookie odds? what does it use? 
Thanks

👍 stevencurrie

`author`	outcast81
`permlink`	re-stevencurrie-soccer-predictions-using-python-part-4-20170923t212303486z
`category`	programming
`json_metadata`	{"tags":["programming"],"app":"steemit/0.1"}
`created`	2017-09-23 21:23:03
`last_update`	2017-09-23 21:23:03
`depth`	1
`children`	3
`last_payout`	2017-09-30 21:23:03
`cashout_time`	1969-12-31 23:59:59
`total_payout_value`	0.000 HBD
`curator_payout_value`	0.000 HBD
`pending_payout_value`	0.000 HBD
`promoted`	0.000 HBD
`body_length`	176
`author_reputation`	209,754,362,359
`root_title`	"Soccer Predictions using Python (part 4)"
`beneficiaries`	`[]`
`max_accepted_payout`	1,000,000.000 HBD
`percent_hbd`	10,000
`post_id`	15,741,630
`net_rshares`	944,947,518
`author_curate_reward`	""

properties (23)vote details (1)

voter	weight	wgt%	rshares	pct	time
stevencurrie	0 B		944,947,518	100%

@stevencurrie · Sep 23 '17 (edited)

Hey bud, the only data it uses is the scores from the previous games.  We take averages for home and away goals in the competition, then the averages for each team we want to predict, use these to calculate expected goals for each team.  These are then used to pick random samples from the poisson distribution which get averaged out to give us our prediction.  [Here](https://www.pinnacle.com/en/betting-articles/Soccer/how-to-calculate-poisson-distribution)'s an excellent description of the process - it's the one I followed to get started.

👍 outcast81, stevencurrie

`author`	stevencurrie
`permlink`	re-outcast81-re-stevencurrie-soccer-predictions-using-python-part-4-20170923t220838373z
`category`	programming
`json_metadata`	{"tags":["programming"],"links":["https://www.pinnacle.com/en/betting-articles/Soccer/how-to-calculate-poisson-distribution"],"app":"steemit/0.1"}
`created`	2017-09-23 22:08:39
`last_update`	2017-09-23 22:34:15
`depth`	2
`children`	2
`last_payout`	2017-09-30 22:08:39
`cashout_time`	1969-12-31 23:59:59
`total_payout_value`	0.000 HBD
`curator_payout_value`	0.000 HBD
`pending_payout_value`	0.000 HBD
`promoted`	0.000 HBD
`body_length`	543
`author_reputation`	7,132,661,654
`root_title`	"Soccer Predictions using Python (part 4)"
`beneficiaries`	`[]`
`max_accepted_payout`	1,000,000.000 HBD
`percent_hbd`	10,000
`post_id`	15,744,005
`net_rshares`	2,036,356,883
`author_curate_reward`	""

properties (23)vote details (2)

voter	weight	wgt%	rshares	pct	time
stevencurrie	0 B		888,250,667	100%
outcast81	0 B		1,148,106,216	100%

@outcast81 · Sep 23 '17 (edited)

You are guilty for making me lose a couple of hours already, boy I look dumb staring at all this stuff on my screen, we have a saying in portuguese, "like an ox  staring at a palace".
Wouldn't going so far back in games kind of ruin the prediction a bit, when calculating averages for each team? (Because you go into past seasons, diferent players, managers etc.)

Anyway I tried to run your code and got this error: ModuleNotFoundError: No module named 'pandas'
I need the database right? Can I use my own database?

The possibiltys with this are immense, unfortunatly my knowledge of this isn't.

👍 stevencurrie, outcast81

`author`	outcast81
`permlink`	re-stevencurrie-re-outcast81-re-stevencurrie-soccer-predictions-using-python-part-4-20170923t233320223z
`category`	programming
`json_metadata`	{"tags":["programming"],"app":"steemit/0.1"}
`created`	2017-09-23 23:33:21
`last_update`	2017-09-23 23:35:30
`depth`	3
`children`	1
`last_payout`	2017-09-30 23:33:21
`cashout_time`	1969-12-31 23:59:59
`total_payout_value`	0.000 HBD
`curator_payout_value`	0.000 HBD
`pending_payout_value`	0.000 HBD
`promoted`	0.000 HBD
`body_length`	597
`author_reputation`	209,754,362,359
`root_title`	"Soccer Predictions using Python (part 4)"
`beneficiaries`	`[]`
`max_accepted_payout`	1,000,000.000 HBD
`percent_hbd`	10,000
`post_id`	15,747,956
`net_rshares`	907,149,618
`author_curate_reward`	""

properties (23)vote details (2)

voter	weight	wgt%	rshares	pct	time
stevencurrie	0 B		907,149,618	100%
outcast81	0 B		0	0%

1 reply

@precise · Sep 30 '17

I like what you've done here. I was able to run your code. Do you mind if i 
 use and test it? I can see where improvements can be done. Great job! 

By the way, an easier to install dependencies is to store them in a file for example **requirement.txt** then place these lines inside that file: 

```
pandas
requests
beautifulsoup4
```

you can then install it by 
`
pip install -r requirements.txt
`

you can also specify module versions on **requirements.txt** ie. `pandas=0.20.0`

👍 stevencurrie

`author`	precise
`permlink`	re-stevencurrie-soccer-predictions-using-python-part-4-20170930t034843069z
`category`	programming
`json_metadata`	{"tags":["programming"],"app":"steemit/0.1"}
`created`	2017-09-30 03:48:45
`last_update`	2017-09-30 03:48:45
`depth`	1
`children`	1
`last_payout`	2017-10-07 03:48:45
`cashout_time`	1969-12-31 23:59:59
`total_payout_value`	0.000 HBD
`curator_payout_value`	0.000 HBD
`pending_payout_value`	0.000 HBD
`promoted`	0.000 HBD
`body_length`	483
`author_reputation`	4,300,224,767,048
`root_title`	"Soccer Predictions using Python (part 4)"
`beneficiaries`	`[]`
`max_accepted_payout`	1,000,000.000 HBD
`percent_hbd`	10,000
`post_id`	16,339,109
`net_rshares`	822,104,341
`author_curate_reward`	""

properties (23)vote details (1)

voter	weight	wgt%	rshares	pct	time
stevencurrie	0 B		822,104,341	100%

@stevencurrie · Sep 30 '17

Hi, I don't mind at all if you use the code, in fact I welcome it.  It would be great if you could improve it.

There's so much that could be added, making use of the over/under and BTTS predictions, adding over/under predictions for other scores, predictions for handicaps, correct scores, the list goes on.

I've also thought about scraping a list of available countries & competitions and implementing a GUI with some nice graphs comparing each team.

Anyway, feel free to do anything you like with the code.  And if you wanted to, you could post your improvements here on Steem.

Thanks as well for the advice on requirements.txt.  I'll just go and add that now.

Have fun mate.  :-)

properties (22)

`author`	stevencurrie
`permlink`	re-precise-re-stevencurrie-soccer-predictions-using-python-part-4-20170930t095231830z
`category`	programming
`json_metadata`	{"tags":["programming"],"app":"steemit/0.1"}
`created`	2017-09-30 09:52:30
`last_update`	2017-09-30 09:52:30
`depth`	2
`children`	0
`last_payout`	2017-10-07 09:52:30
`cashout_time`	1969-12-31 23:59:59
`total_payout_value`	0.000 HBD
`curator_payout_value`	0.000 HBD
`pending_payout_value`	0.000 HBD
`promoted`	0.000 HBD
`body_length`	687
`author_reputation`	7,132,661,654
`root_title`	"Soccer Predictions using Python (part 4)"
`beneficiaries`	`[]`
`max_accepted_payout`	1,000,000.000 HBD
`percent_hbd`	10,000
`post_id`	16,361,398
`net_rshares`	0

@stevencurrie · Sep 24 '17

$0.69

My program has only 2 predictions for today.
Match|Predicted|Prob%|My Odds|Betfair Odds
-|-|-|-|-
Motherwell v Aberdeen| Aberdeen Win|78.38|1.28|1.57
Verona v Lazio|Lazio Win|95.28|1.05|1.9

👍 thing-2, outcast81, stevencurrie

`author`	stevencurrie
`permlink`	re-stevencurrie-soccer-predictions-using-python-part-4-20170924t002231165z
`category`	programming
`json_metadata`	{"tags":["programming"],"app":"steemit/0.1"}
`created`	2017-09-24 00:22:30
`last_update`	2017-09-24 00:22:30
`depth`	1
`children`	6
`last_payout`	2017-10-01 00:22:30
`cashout_time`	1969-12-31 23:59:59
`total_payout_value`	0.517 HBD
`curator_payout_value`	0.169 HBD
`pending_payout_value`	0.000 HBD
`promoted`	0.000 HBD
`body_length`	189
`author_reputation`	7,132,661,654
`root_title`	"Soccer Predictions using Python (part 4)"
`beneficiaries`	`[]`
`max_accepted_payout`	1,000,000.000 HBD
`percent_hbd`	10,000
`post_id`	15,750,235
`net_rshares`	202,692,619,336
`author_curate_reward`	""

properties (23)vote details (3)

voter	rshares	pct
thing-2	200,585,584,338	100%
stevencurrie	935,498,043	100%
outcast81	1,171,536,955	100%

@outcast81 · Sep 24 '17

yup another two lmao fortunatly I saw this in time to bet on Lazio :)

👍 stevencurrie

`author`	outcast81
`permlink`	re-stevencurrie-re-stevencurrie-soccer-predictions-using-python-part-4-20170924t183722257z
`category`	programming
`json_metadata`	{"tags":["programming"],"app":"steemit/0.1"}
`created`	2017-09-24 18:37:24
`last_update`	2017-09-24 18:37:24
`depth`	2
`children`	4
`last_payout`	2017-10-01 18:37:24
`cashout_time`	1969-12-31 23:59:59
`total_payout_value`	0.000 HBD
`curator_payout_value`	0.000 HBD
`pending_payout_value`	0.000 HBD
`promoted`	0.000 HBD
`body_length`	69
`author_reputation`	209,754,362,359
`root_title`	"Soccer Predictions using Python (part 4)"
`beneficiaries`	`[]`
`max_accepted_payout`	1,000,000.000 HBD
`percent_hbd`	10,000
`post_id`	15,819,230
`net_rshares`	883,525,930
`author_curate_reward`	""

properties (23)vote details (1)

voter	weight	wgt%	rshares	pct	time
stevencurrie	0 B		883,525,930	100%

@stevencurrie · Sep 24 '17

Hey, that's great. I'm going to test a little longer before I start betting live.

Did you manage to get the code running btw?  (I'm looking into the best way to automate installing the prerequisite packages.)

👍 outcast81

`author`	stevencurrie
`permlink`	re-outcast81-re-stevencurrie-re-stevencurrie-soccer-predictions-using-python-part-4-20170924t190541754z
`category`	programming
`json_metadata`	{"tags":["programming"],"app":"steemit/0.1"}
`created`	2017-09-24 19:05:42
`last_update`	2017-09-24 19:05:42
`depth`	3
`children`	3
`last_payout`	2017-10-01 19:05:42
`cashout_time`	1969-12-31 23:59:59
`total_payout_value`	0.000 HBD
`curator_payout_value`	0.000 HBD
`pending_payout_value`	0.000 HBD
`promoted`	0.000 HBD
`body_length`	209
`author_reputation`	7,132,661,654
`root_title`	"Soccer Predictions using Python (part 4)"
`beneficiaries`	`[]`
`max_accepted_payout`	1,000,000.000 HBD
`percent_hbd`	10,000
`post_id`	15,821,114
`net_rshares`	1,165,679,270
`author_curate_reward`	""

properties (23)vote details (1)

voter	weight	wgt%	rshares	pct	time
outcast81	0 B		1,165,679,270	100%

3 replies

@stevencurrie · Sep 24 '17

Woohoo.  Both predictions correct!

properties (22)