Entries Tagged 'Footy' ↓

Footy tipping – round 22

And thus the footy year ends. Well, not really: we still have the finals. But the tipping competitions are over.

The final week brought no big surprises, except maybe for Adelaide beating West Coast (but that seemed somewhat likely given the results of the last few weeks). The lack of surprises brought a good number of hits (seven for money, six for ladder) but not good financial results (low odds on most matches).

The numbers for this week:

Round 22 Ladder Money Home Random
Correct tips 6 7 4 5
Accuracy 75% 87.5% 50% 62.5%
$ result $-1.80 $4.75 $-14.90 $-2.35
 
Cumulative Ladder Money Home Random
Correct tips 107 114 107 95
Accuracy 60.80% 64.77% 60.80% 53.98%
$ result $-7.25 $-52.10 $74.55 $19.35

So, the big question: can you make money with these strategies? The answer seems to be “yes” for the “home” strategy (that is, tipping always the home team), but let’s look at a graph first:

footy tipping strategies - performance

The graph shows the cumulative net result of each strategy over the year. “Ladder” did very well in the first seven weeks, then went straight downhill; remarkably, that’s exactly the opposite of what I’d expect. The reason for this might be that, after a few weeks, the teams at the top of the ladder starting being the favourites and paying less. This would explain the series of coincidences between “ladder” and “money” in the final weeks.

“Money”, on the other hand, went straight down and remained there. It did stabilise after a while, but never really recovered. Very strange, to be honest.

“Random” was very random. Except for a big jump upwards in rounds 6 to 9, the rest looks like Brownian motion to me, with a slight tendency downwards.

Finally, “home” is pretty much stable most of the time, with a few jumps (rounds 4, 17, 18 and 21) providing all of the net gain. It looks to me like it only made money because of a few unpredictable results here and there, but it could easily have gone the other way.

One important point, which I noticed during the year, is that you can tip with two different goals in mind, and your strategy will need to be different depending on your goal. If you are part of a tipping competition, you need to tip on every match and the only interesting data point is the number of results you get right. For this, the “money” strategy can be useful, and in fact it may be enough to get you to win in small competitions (in your office or school, for example), but not in the large, public ones (the winner of the Channel 10 competition finished with 127 points; “money” would have finished around position #700).

If, however, you are looking to make money, you don’t need to tip on every match (and, in fact, you probably shouldn’t): part of the strategy must be to select which results to tip and which to ignore. None of these strategies are useful for that, although I believe it is possible to develop an automated method that can reliably make money. This is something to think about over summer…

Links:

Footy tipping – round 21

Second-to-last round, and some interesting results: three matches were decided by less than a goal, and three others by more than 10 goals. A few upsets, as well, which made for mediocre results for “ladder” and “money” (which, once again, tipped identically): only four hits, which caused the cumulative financial results for “ladder” to become negative for the first time in the season. “Home” did very well with seven, and random was, well, random, and got three. So far, not once has any strategy got all the results right in a round.

The numbers for this week:

Round 21 Ladder Money Home Random
Correct tips 4 4 7 3
Accuracy 50% 50% 87.5% 37.5%
$ result $-10.35 $-10.35 $25.75 $-12.75
 
Cumulative Ladder Money Home Random
Correct tips 101 107 103 90
Accuracy 60.12% 63.69% 61.31% 53.57%
$ result $-5.45 $-56.85 $89.45 $21.70

Next week: the final results of the season and an analysis of the results of this experiment.

Links:

Footy tipping – round 20

We’re approaching the end of the season; only two more weeks to go. The results seem pretty much stable by now, and it’s clear that we’ll get a “hit rate” of approximately 60% for “ladder” and “home”, 65% for “money” and slightly over 50% for “random” (as predicted). That said, the results this week were pretty much average: 5 hits for “ladder” and “money”, only 3 for “home” (lower than usual) and 6 for “random”(higher than usual). This was also the third week in a row where “ladder” and “money” tipped exactly the same teams (and, of course, got exactly the same results).

The numbers for this week:

Round 20 Ladder Money Home Random
Correct tips 5 5 3 6
Accuracy 62.5% 62.5% 37.5% 75%
$ result $-3.80 $-3.80 $-11.20 $11.05
 
Cumulative Ladder Money Home Random
Correct tips 97 103 96 87
Accuracy 60.63% 64.38% 60.00% 54.38%
$ result $4.90 $-46.50 $63.70 $34.45

Links:

Footy tipping – round 19

Excelent results this week. Once more, “ladder” and “money” had exactly the same selections, and both did very well with seven correct tips; that was the best result ever for “ladder”. “Home” got 5 hits and “random” got only 4 (the average tipper this week got six correct results).

The numbers for this week:

Round 19 Ladder Money Home Random
Correct tips 7 7 5 4
Accuracy 87.5% 87.5% 62.5% 50%
$ result $7.80 $7.80 $-3.75 $-6.75
 
Cumulative Ladder Money Home Random
Correct tips 92 98 93 81
Accuracy 60.53% 64.47% 61.18% 53.29%
$ result $8.70 $-42.70 $74.90 $23.40

Links:

Footy tipping – round 18

The results were not so bad this week, but not so good either. An unusual event was that “ladder” and “money” had exactly the same selections, and “random” had almost exactly the reverse of those (except for the Bulldogs-Lions match, where it followed them in tipping the Lions; the Bulldogs won). In the end, “home” got 6 hits, “money” and “ladder” got 5 and “random” got only 2.

The numbers for this week:

Round 18 Ladder Money Home Random
Correct tips 5 5 6 2
Accuracy 62.5% 62.5% 75% 25%
$ result $-4.45 $-4.45 $22.60 $-11.00
 
Cumulative Ladder Money Home Random
Correct tips 85 91 88 77
Accuracy 59.03% 63.19% 61.11% 53.47%
$ result $0.90 $-50.50 $78.65 $30.15

Links:

Footy tipping – round 17

I almost forgot to put this up this week… and the results were not that bad. “Home” had its best week so far, with 7 hits, and “money” had 6; “ladder” got only 4, however.

With only five weeks left until the end of the “home and away” season, it is pretty clear that (a) no strategy is a clear money-winner, and (b) the performance of all strategies doesn’t seem to change that much as the season advances. I’ll do a more detailed analysis after the end of the season, though.

The numbers for this week:

Round 17 Ladder Money Home Random
Correct tips 4 6 7 5
Accuracy 50% 75% 87.5% 62.5%
$ result $-10.05 $3.20 $26.45 $11.70
 
Cumulative Ladder Money Home Random
Correct tips 80 86 82 75
Accuracy 58.82% 63.24% 60.29% 55.15%
$ result $5.35 $-46.05 $56.05 $41.15

Links:

Footy tipping – round 16

As the end of the season approaches, it becomes clearer and clearer that simple strategies such as the one used here are not good enough to make any serious money with (or to win any tipping competitions…).

This week brought average results: nothing particularly good (6 for “random” and “money”), nothing terribly bad (3 for “home”). It’s interesting to point out that no strategy has ever got all results for one week right; our best performance was 7 correct hits, which happened three times for “money” and once for “random”.

The numbers for this week:

Round 16 Ladder Money Home Random
Correct tips 5 6 3 6
Accuracy 62.5% 75% 37.5% 75%
$ result $-2.60 $4.90 $-16.60 $12.35
 
Cumulative Ladder Money Home Random
Correct tips 76 80 75 70
Accuracy 59.38% 62.50% 58.59% 54.69%
$ result $15.40 $-49.25 $29.60 $29.45

Links:

Footy tipping – round 15

All strategies did well this week, at least in correct guesses; not so much in financial results. “Money” repeated once again its best result so far, with 7 hits; “home” got 6 and “ladder”, once more, got 5.

Round 15 is an interesting one because is the last of the “first round” of matches, so to speak: all teams have already played each other once, and the next seven weeks will have “repeat” games. It will be interesting to look at how the teams go when repeating previous matches.

The numbers for this week:

Round 15 Ladder Money Home Random
Correct tips 5 7 6 5
Accuracy 62.5% 87.5% 75% 62.5%
$ result $-3.85 $11.20 $6.55 $-2.50
 
Cumulative Ladder Money Home Random
Correct tips 71 74 72 64
Accuracy 59.17% 61.67% 60.00% 53.33%
$ result $18.00 $-54.15 $46.20 $17.10

Links:

Footy tipping – round 14

Very similar results for all strategies this week: 5 hits for each of ladder, money and home. Random got only three. In fact, random performed so badly that it lost its lead and is no longer the best performing strategy, financially-speaking: home, the only one to record a net profit this week, has that honour now.

The numbers:

Round 14 Ladder Money Home Random
Correct tips 5 5 5 3
Accuracy 62.5% 62.5% 62.5% 37.5%
$ result $-4.15 $-5.75 $2.20 $-21.80
 
Cumulative Ladder Money Home Random
Correct tips 66 67 66 59
Accuracy 58.93% 59.82% 58.93% 52.68%
$ result $21.85 $-65.35 $39.65 $19.60

Links:

Footy tipping – round 13

Not really many upsets in this split round; things seem to be settling down. Other than Essendon beating St. Kilda and, arguably, Brisbane beating Geelong, every result was pretty much as expected. This meant reasonably good results for most strategies: 6 correct hits for “ladder”, 5 for “money” and “home”. The random pseudo-strategy did poorly, with only three hits.

Financially, “ladder” and “home” did well, with positive results, and the two other had losses. Remarkably, on aggregate, “random” is still ahead, but I don’t expect this to last (although the best possible strategy so far, financially speaking, would be to reverse the money strategy, as I mentioned before). Here are the numbers for this week:

Round 13 Ladder Money Home Random
Correct tips 6 5 5 3
Accuracy 75% 62.5% 62.5% 37.5%
$ result $12.95 $-3.30 $6.20 $-11.00
 
Cumulative Ladder Money Home Random
Correct tips 61 62 61 56
Accuracy 58.65% 59.62% 58.65% 53.85%
$ result $26.00 $-59.60 $37.45 $41.40

Links:

Footy tipping – round 12

After many, many upsets, this was definitely the most predictable week so far, and by far. The average tipper scored 5.9 correct hits, and my strategies didn’t score too badly either. “Money”, as could be expected, got 7 correct results, equaling its best week up to now (round 7); “home” got 6, “ladder”, disappointingly, got just 5, and “random”, true to its name, got 4.

Round 13 will be a split round, which means that the next update will be in two weeks. Here are the numbers for this week:

Round 12 Ladder Money Home Random
Correct tips 5 7 6 4
Accuracy 62.5% 87.5% 75% 50%
$ result $-5.95 $7.95 $5.50 $-11.50
 
Cumulative Ladder Money Home Random
Correct tips 55 57 56 53
Accuracy 57.29% 59.38% 58.33% 55.21%
$ result $13.05 $-56.30 $31.25 $52.40

Links:

Footy tipping – round 11

This week ended with slightly better results than the last two, and with a mix of results that caused all three major strategies to finish with the same number of correct hits, cumulatively: 50 out of 88. The random “strategy” has one less hit.

During this week, there was also an interesting report on news.com.au saying that the AFL season is so unpredictable that the experts are doing no better than the novices at tipping. That does make me feel a little better, but it makes me wonder whether the results of this little experiment mean anything. That also explains why the best possible strategy so far, financially speaking, would be to bet always on the team that pays the most (the one least expected to win); with my rules, this strategy would have netted $86.65 on 38 correct guesses (43.18%).

In any case, here are the numbers:

Round 11 Ladder Money Home Random
Correct tips 5 4 3 4
Accuracy 62.5% 50% 37.5% 50%
$ result $1.40 $-8.60 $-5.85 $8.65
 
Cumulative Ladder Money Home Random
Correct tips 50 50 50 49
Accuracy 56.82% 56.82% 56.82% 55.68%
$ result $19.00 $-64.25 $25.75 $63.90

Links:

Footy tipping – round 10

Well, it took me nine weeks, but I finally noticed that, if I wanted this little experiment to at least look scientific, there was something missing: a control group. Which, in this case, translates to a “pseudo-strategy” that would select random winners for all games. In this way, I’d be able to compare the results of my simple strategies to blind luck.

Fortunately, this little flaw is easy to fix. I wrote a program to randomly select eight results (“home” or “away”, with a 50-50 chance); ran it a few million times to ensure that I was indeed getting a random spread of results; and then ran it nine times and compared the results retroactively with the first nine rounds of matches. Then I ran it once more to generate results for this week.

The results were… interesting. When compared to the other three strategies, random tipping is achieving at least as many correct guesses as the worst-performing of the others, and is returning more money. Most of this is due to managing to get seven correct results in round 9, where most of those were upsets. So, it would seem that none of the other strategies is that much better, or even as good as, random guessing.

As for this week, we got mixed results: “money” and “home” got five correct guesses each (both had exact the same tips, by the way), and “ladder” and “random” got three each. This was the second week in a row when all three strategies, other than the control group, returned net financial losses.

And here are the numbers:

Round 10 Ladder Money Home Random
Correct tips 3 5 5 3
Accuracy 37.5% 62.5% 62.5% 37.5%
$ result $-5.70 $-2.50 $-2.50 $-15.45
 
Cumulative Ladder Money Home Random
Correct tips 45 46 47 45
Accuracy 56.25% 57.50% 58.75% 56.25%
$ result $17.60 $-55.65 $31.60 $55.25

Links:

Footy tipping – round 9

Another disapointing week: four correct guesses for “money”, three for each of the others. There were lots of upsets this week, including Collingwood, who was (and still is) at the bottom of the ladder, beating West Coast, who was (and still is!) at the top. So much so that the average tipper scored 3.5 correct guesses (that is, less than 50%).

Speaking of money, all three returned net losses this week, but “home” still leads in the cumulative results. As a side note, I did some calculations and found out that, by doing exactly the opposite of the “money” strategy (that is, by selecting for each game the team that pays the most) I would have a better financial result that with any of the other strategies. However, this return comes mostly from the last two rounds ($49 out of a total of $51), which were full of upsets; it doesn’t look like a good long-term strategy.

The numbers:

Round 9 Ladder Money Home
Correct tips 3 4 3
Accuracy 37.5% 50% 37.5%
$ result $-13.15 $-10.50 $-3.00
 
Cumulative Ladder Money Home
Correct tips 42 41 42
Accuracy 58.33% 56.94% 58.33%
$ result $23.30 $-53.15 $34.10

Links:

Footy tipping – round 8

Well, following a week with excelent results, this last weekend was terrible: four correct guesses for “ladder” and “home”, and three for “money”. The average tipper scored either 4.3 or 4.4 hits, depending on whom you ask, so my results were not exactly that bad, but they were disappointing nonetheless. There were so many “easy” games this week that I honestly expected better results on all strategies.

Interestingly, since “home” got right some rather unexpected results, it has actually made money this week, and cumulatively it is now returning more than “ladder”; “money” continues with a net loss, and sinking quickly. Always remembering that financial results are fictitious: no money is actually changing hands.

With 14 rounds left, these are the results:

Round 8 Ladder Money Home
Correct tips 4 3 4
Accuracy 50% 37.5% 50%
$ result $-10.10 $-20.10 $3.25
 
Cumulative Ladder Money Home
Correct tips 39 37 39
Accuracy 60.94% 57.81% 60.94%
$ result $36.45 $-42.65 $37.10

Links: